The paper provides some operative replies to evaluate the effectiveness and the critical issues of the simultaneous localisation and mapping (SLAM)-based mobile mapping system (MMS) called ZEB by GeoSLAM™ https://geoslam.com/technology/.In these last years, this type of handheld 3D mapping technology has increasingly developed the framework of portable solutions for close- range mapping systems that have mainly been devoted to mapping the indoor building spaces of enclosed or underground environments, such as forestry applications and tunnels or mines. The research introduces a set of test datasets related to the documentation of landscape contexts or the 3D modelling of architectural complexes. These datasets are used to validate the accuracy and informative content richness about ZEB point clouds in stand-alone solutions and in cases of combined applications of this technology with multisensor survey approaches. In detail, the proposed validation method follows the fulfilment of the endorsed approach by use of root mean square error (RMSE) evaluation and deviation analysis assessment of point clouds between SLAM-based data and 3D point cloud surfaces computed by more precise measurement methods to evaluate the accuracy of the proposed approach. Furthermore, this study specifies the suitable scale for possible handlings about these peculiar point clouds and uses the profile extraction method in addition to feature analyses such as corner and plane deviation analysis of architectural elements. Finally, because of the experiences reported in the literature and performed in this work, a possible reversal is suggested. If in the 2000s, most studies focused on intelligently reducing the light detection and ranging (LiDAR) point clouds where they presented redundant and not useful information, contrariwise, in this sense, the use of MMS methods is proposed to be firstly considered and then to increase the information only wherever needed with more accurate high-scale methods. . . . . . . Keywords 3D mapping Cultural heritage Sensor integration SLAM ZEB Mobile mapping systems Landscape Introduction landscape scenarios. Similarly, the methodological approaches and technological solutions offered by geomatics research The concepts corresponding to the pursuit of rapid mapping should comply with time-cost ratios in the overall resource con- solutions for the cultural heritage (CH) domain in supporting sumption within a context that is as prone to the underinvest- multiscale documentation of indoor and/or outdoor built heri- ment of funds as the CH one. These solutions are essential tage should be more tailored on several needs that may arise in whenever monitoring documentation operations are frequently individual operative contexts, particularly, historic structures and required in cases of reduced accessibility to spaces or wherever the possibility of implementing the consolidated 3D survey pro- cedures would be limited or not enough for wider areas. Thus, the effectiveness of imaging or ranging measurement * Giulia Sammartano systems for 3D data acquisition can be evaluated in their email@example.com adaptability to complex settings and based on the achievable Antonia Spanò descriptive capabilities in richly geometrically featured envi- firstname.lastname@example.org ronments or enclosed spaces for a suitable surface reconstruc- tion responsive to the expected purposes. In this regard, the Dipartimento di Architettura e Design (DAD), Politecnico di Torino, proposed research tests a new rapid mapping technology: a 3D Viale Mattioli 39, 10129 Turin, Italy Appl Geomat handheld mobile system based on the simultaneous the trajectories are evaluated based on their stand-alone use localisation and mapping (SLAM) algorithm applied to vari- and on their abilities to support and integrate the geometric ous profile measurements, in which the new system is called description provided by other methodological image-based or the ZEB scanner by GeoSLAM (marketed in Italy by Me.s.a range-based approaches. srl). In fact, beyond the close-range photogrammetry (CRP) imaging approach, new worthy solutions have overlooked the panorama of ranging systems to produce dense and detailed Mobile mapping systems in CH domain point clouds. complexity The currently pervasive image-matching photogrammetric methods provide even more effective densification algorithms The LiDAR-based MMSs offer a very wide range of compet- and flexible and user-friendly solutions in 3D reconstruction itive solutions for the 3D mapping of extensive and/or com- both in terrestrial and in unmanned aerial vehicle (UAV) aerial plex spaces. They can be generally defined as combinations of datasets (Lingua et al. 2017; Murtiyoso et al. 2017; the following sensors for point cloud acquisition and position- Remondino et al. 2014). ing (or geo-positioning): a mapping sensor (i.e., LiDAR, ac- Also, the range-based techniques being currently imple- tive 3D imaging systems), an inertial measurement unit, a mented are many types of light detection and ranging global navigation satellite system (GNSS) receiver, and a time (LiDAR)-based mobile mapping systems (MMSs) (Puente referencing unit (Puente et al. 2013). et al. 2013; Rodríguez-Gonzálvez et al. 2017), and they are According to the acquisition mode movement, platforms flexible and competitive solutions for the 3D mapping of wide can be divided on those movements that equip a vehicle (by and complex spaces (Farella et al. 2016; Tucci et al. 2017) ground, on air, or on water) or portable movements (by towing compared with the close-/medium-/long-range laser devices trolley, man-portable backpacks, or portable handheld de- used for terrestrial laser scanning (TLS). Nevertheless, these vices), with or without GNSS positioning solutions MMSs remain a pricey and complex technological solution of (Nocerino et al. 2017). fusion-based sensors with different levels of equipment and The LiDAR-based embedded ranging measurements inte- manageability. The research and development on these inte- grated with the different navigation system solutions in the grated systems have helped address movability convenience MMS equipping cars or mobile/self-propelled platforms allow issues that better meet the needs of applicability in CH sites the helpful exploitation of derived point clouds for their anal- according to their potential needs related to their complexities. ysis and management by specific multidisciplinary activities In the recent research, the developments in the framework regarding heritage structures. of portability and compactness features for MMS solutions in Especially recently, the vehicle-equipping mobile naviga- close-range mapping systems have been mainly devoted to tion systems seem to be an effective answer to cover sizable transportable solutions. Handheld or by otherwise portable distances and obtain metrically dense point cloud surfaces devices can be thus deployed to survey indoor volumes and with geometric and radiometric data. Their use in dense enclosed or narrow environments, such as multilevel build- urban centres of historical cities, featured by high structures ings, industrial spaces, forestry, tunnels, mines, caves, and and narrow streets, has been studied for Bergamo city in all related environmental and geospatial applications. Toschi et al. (2017) and Toschi et al. (2015), to actually dem- Several studies have also investigated these portable technol- onstrate the point cloud ability to reach the detail for high- ogies in the CH domain. scale accuracy; the fruitful integration with aerial datasets Thus, the portable handheld MMS, the ZEB SLAM-based can provide a dense 3D reconstruction on which the 3D city scanner, in the ZEB1 and ZEB-REVO configurations provided modelling approaches can obtain a fundamental metric source by GeoSLAM™, has been tested in distinctive configurations in of geometric data. In larger outdoor complexes, the use of an the proposed datasets and has been studied in this work. The first MMS can also be a valid complementary integration or alter- dataset is from the San Silvestro archaeomining park in Livorno, native for 3D mapping purposes and to document a very ex- Tuscany, Italy (Brocchini et al. 2017), with its ancient under- tensively built heritage complex in advantageous times, as ground mining network and ruined masonry fortress, while the proposed in the impressive fortified walls of Avila Alcázar second dataset is from the Valperga castle in Torino, Piedmont, in Rodríguez-Gonzálvez et al. (2017). Similarly, the integra- Italy (Chiabrando et al. 2017a, b), with the system of spaces bility aspects of dense surfaces, retrieved by different MMSs, articulated in above- and underground volumes. The historical are being studied for their validation and point cloud optimi- complexes are featured by peculiar architectural geometries and sation for several uses in CH documentation and management. material surfaces. In fact, the ability of mobile ranging systems to cover great Indoor and outdoor scenes have been made available to complexes at different scales within the integration of other validate point cloud accuracy in surface reconstruction, com- different static and dynamic measurement techniques and re- puted by the mapping system along different trajectories: first, lated instrumental solutions can be a significant contribution Appl Geomat in the multiscale documentation of richly architectonic fea- tured sites. This aptitude is investigated in the multiscale com- bined workflow in Tucci et al. (2017), where, in parallel to imaging and ranging approaches (CRP, TLS, UAV photo- grammetry), a multisource range-based 3D mapping method was implemented by fourfold MMS solutions, a mobile sys- tem equipping a car, a portable ranging measurement trolley, a portable backpack, and a handheld portable solution that al- lows accurate 3D mapping of an articulated system of under- ground passageways. Fig. 1 The ZEB-REVO system with the rotating ranging head, tested by Geomatics Lab for Cultural Heritage, PoliTo, in 2017 in San Severino Handheld MMS for SLAM-based 3D mapping Marche (MC) Within the framework of portable scanners, this last type of handheld 3D mapping solution, quite unique of its kind, is the Many other contexts belonging to the built urban heritage, ZEB by GeoSLAM. The solution of the indoor positioning in the outdoor scenario but mostly in the indoor settings sim- problem is based on implementing a SLAM-based algorithm ilar to underground environments such as tunnels and caves, applied on range-based profiles progressively extracted by the have offered challenging validation purposes, such as the two- head of the continuously moving device in particularly fold technique proposed for this validation research in enclosed and richly geometrically featured spaces (Bosse BExperimental section: validation strategy^. and Zlot 2009; Bosse et al. 2012), as explained in BZEB sys- The use of the ZEB1 system inside the Pausilypon site, in tem operational behaviour^. Farella et al. (2016), returned a flexible solution for The main application fields in which the portable systems documenting the enclosed passageways with topographic var- firstly proved their effectiveness have been those ones in iations along lengthy trajectories in viable operation time. which the use of their operating principle could be better However, this system emphasises the problem in the final exploited: forestry (Bauwens et al. 2016;Rydingetal. accuracy validation based on both the risk related to the sub- 2015); civil applications and geology, such as underground division of an elongated path and the problem of regular and tunnels (Dewez et al. 2016;Eyre et al. 2016); and open pit low geometrically featured surfaces that do not allow the mines (Vanneschi et al. 2017). SLAM-based algorithm to optimally operate for the progres- Recently, this handheld SLAM-based MMS has also been sive profile alignment. The use of the ZEB1 mapping device specifically evaluated and metrically tested with other similar in similar indoor enclosed and mazy spaces is validated as a technological solutions for indoor/outdoor structure mobile winning approach in terms of time-consuming performances, mapping, as proposed in Thomson et al. (2013), in Sirmacek manoeuvrability, and continuity in surface reconstruction, et al. (2016) and Díaz-Vilariño et al. (2017), and in Nocerino et al. 2017. Table 1 Main specs of the ZEB devices by GeoSLAM, from https:// Parallelly, the crucial ability of the ZEB system to recover geoslam.com/technology/,Cadge 2016,Eyre et al. 2016, and Nocerino also large outdoor manufactured spaces and articulated struc- et al. 2017 tures has been examined and is still being studied. The use of ZEB was first proposed in Zlot et al. (2014), towards the ZEB1 ZEB-REVO scattered historical structures distributed in the extensive nat- Wavelength 905 nm 905 nm ural environment in the Peel Island Lazaret despite the state of Eye-safe laser Class1 Class1 conservation of the ruined artefacts in such a scenario. In Laser speed 40 Hz 100 Hz works by Sammartano (2017) and Sammartano (2018), the Laser lines 40 lines/s 100 lines/s ZEB 3D mapping approach is proposed and validated rather Scan speed × 1 × 2.5 with the use of UAV photogrammetric survey as a valuable Maximum range 15–30 m 15–30 m combination for the damage documentation of collapsed Points density ~ 43,200 pps ~ 43,200 pps structures by rapid mapping in risk areas and post-disaster 3D measurement ± 0.1% ± 0.1% scenarios. The details achievable by this ranging terrestrial declared accuracy approach overtake in most cases the photogrammetric ap- FoV – 270° H /100° V FOV FOV proach, despite the lack in the reconstruction of the upper Weight/portability ~1.5 kg ~2 kg parts: the integration of the essential aerial approach is needed (head + data logger) to be geometrically completed. Appl Geomat Fig. 2 The output data of ZEB acquisition: the time-marked 3D model, with related range colour scale (a) and the quality-marked trajectory along the castle courtyard (b). Shaded view of the 3D model (c) such as in roof garrets resulting from the extrados of the vaults & Application complexity of the transept in Milan’scathedral, in Mandelli et al. (2017). The proficiency of this technological solution to accom- – Rapid performance, avoiding time-consuming operations plish many kinds of complexity requisites in these kinds of – Possibly replacing consolidated systems in indoor CH framework can be briefly summarised in terms of com- configurations plexity of applicative contexts and feasible applications: & Physical complexity of contexts Table 2 Framework of the selected dataset subject to validation – Articulated landscape complexes or particularly wide Indoor Outdoor sites to be travelled through several paths configurations, Valperga castle context Cylindrical tower (A) Courtyard (C) such as variables regarding linear or circular trajectories Ancient ice house (B) with uphill/downhill and roundtrip configurations San Silvestro Medieval mining Fortified village (E) – Structured built heritage in indoor and outdoor develop- archaeominig park cave (D) ments, underground environments, ramblings, and nar- Stand-alone validation Multisensory validation row surface reconstructions Appl Geomat – Need to obtain essential volumetric features for planning In the ZEB1 device, the sensor head is mounted on a the design of in-depth specific elements and/or insight- spring that freely and passively swings during operator focused analyses in single parts and vehicle movement (Bosse et al. 2012). In the ZEB- – Capability to deepen and specify the geometric content of REVO-implemented solution, the head is regularly rotat- digital surface models (DSMs) defined by other measure- ing automatically during the operator motion (Table 1). ment approaches (i.e., UAV photogrammetry documen- In the last update of ZEB-REVO, the aspects related to tation), targeting the scale of surveying included among the acquisition and interactivity of the operator with the architectural and urban scales (between 1:100 and 1:200) data collection have been enhanced as the in-time visi- bility of profile SLAM-based registration on a handheld device coupled with the scanning body during the ac- quisition. The RGB GoPro action cam is now equipped ZEB system operational behaviour to the scanning body for the post-texturing of point clouds, as already proposed for ZEB1 by Nocerino The ZEB system by GeoSLAM (Fig. 1) is essentially based on et al. (2017) and Zlot et al. (2014). a moving head equipped with a ranging measurement laser The ranging sensor is a 2D lightweight pulse (Hokuyo capturing 2D point profiles, without a GNSS receiver or direct UTM-30LX scanner (Nikoohemat et al. 2017)) continuously RGB data. The system also comprises an inertial measurement emitted in the form of pulsed light beams in the near infrared unit with triaxial gyros, accelerometers, and three-axis at 905-nm wavelength. The travelling time of flight of the magnetometers. pulse from the sensor to the object and back provides the range Fig. 3 Valperga castle tower: a slam-based point cloud and coloured trajectory, b segmented CRP cloud, and c derived 2D section drawing Appl Geomat measurement collecting profiles of about 43,200 pps at 40 Hz deviations from the measured inertial measurement unit acceler- for ZEB1 and 100 Hz for ZEB-REVO (Cadge 2016; Dewez ations and rotational velocities and guaranteeing the precision of et al. 2016). They are simultaneously aligned during the tra- the global registration. jectory by the registration algorithm implemented in the sys- The resulting data processed either by GeoSLAM proprietary tem and based on 3D SLAM robotics technology (Riisgaard software (SW) application running on a local machine or by a 2005). pay-as-you-go process via cloud processing is a double kind of The specific SLAM algorithm first developed by the CSIRO point cloud comprising the 3D reconstruction and the reference ICT Centre in Brisbane (Australia) (Bosse et al. 2012; trajectory from which it is generated. Both these point-based data GeoSLAM 2016) is based on the exploitation of geometrical are potentially time-marked, as shown in Fig. 2a, and it is possi- attributes featuring the surveyed environment, and the algorithm ble to show in range colours the trajectory of the estimated qual- works for both the trajectory incremental motion estimation and ity of the SLAM-based profile registration, as shown in Fig. 2b. global PC registration along the trajectory. The trajectory [T(τ)] is Problems related to registration drift errors can be corrected by estimated by the algorithm by sequences of translation t(τ)and point cloud reprocess and merging in the last 2017 release of the rotation r(τ) functions of time (τ) according to six DOF (degrees GeoSLAM Hub SW tools. of freedom). The raw laser profiles continuously captured into The fundamental topics considered to avoid bad quality in time-windowed segments are progressively re-projected in the registering point cloud profiles are mainly related to the oper- 3D reconstruction according to the best correspondence to sur- ative fieldwork: face characterisation, indicating that profile matching is an ICP (iterative closest point)-like approach (Bellekens et al. 2014). – Initialisation procedure is mandatory on a planar surface, These iterative process conditions ensure during the ZEB acqui- and the return trip arrival point must be in the same posi- sition the continuity of profile registration with the previous time tion as the starting point. This is fundamental for the point segment, minimising errors between matching surfaces and cloud alignment and closure. Fig. 4 Castle courtyard and overlooking buildings: a alignment of the ZEB points with LiDAR model (in violet, the underground ice cellar position); b vertical view of the starting portion of ice cellar point cloud with red trajectory (red box) Appl Geomat Fig. 5 a Excerpt of the north side of the courtyard TLS 3D model with the integration of the roofing elements from UAV. b Dense 3D model derived from oblique and nadiral UAV images Fig. 6 Medieval mine Buca della Faina in San Silvestro archaeomining park (Brocchini et al. 2017) Appl Geomat (a) (b) (a) (c) Fig. 7 The mapping of Rocca San Silvestro (a)by ZEB-REVO (b)in 2016, and the 2017 integration in rex box (c)(Brocchini etal. 2017) – Trajectory and loops should be planned according to the local environment configuration. They are fundamental be- cause the SLAM-based system is originated on the iterative alignment of extracted profiles that are based on that featur- ing attributes of the space. Indoor spaces or outdoor (b) enclosed environments are favourable for better perfor- Fig. 8 Deviation errors in range colour map from the comparison mance of the ZEB system. Roundtrips are preferred to avoid between the full raw ZEB point cloud and the CRP reference model (a) and zoomed excerpt (b) drift error propagation in Bswing^ effects or linear deviation. – Time of acquisition is considered the sum of the few mi- Considering an on-foot walking operator with a handheld nutes required for initialisation and closure with an auto- device with a speed of almost 1–1.5 m/s, an itinerary of matic in loco pre-processing and storage of data in the utmost 20–30 min is recommended for the best precision embedded memory and the trajectory execution. of outcome data. Table 3 RMSE on the control points in the CRP model computed inside the tower of Valperga castle Table 4 Statistical Outward and return comparison on ZEB values of the deviation Metric control residuals on the CRP model surface analysis of the separated O&R paths for the (m) XY Z Tot Raw Optimised cylindrical tower GCP 0.004 0.003 0.004 0.007 Mean 0.091 0.021 CP 0.005 0.003 0.013 0.014 St.dev 0.131 0.028 Appl Geomat Experimental section: validation strategy Although the relationships between the future uses of point clouds/3D model and the validation strategy may not be The validation of the systems is always essential to determine straightforward, it is crucial to adopt validation rules involving how to use it in different application fields and determine the metric quality and other parameters concerning the usability effectiveness of the systems based on the different contexts. of datasets and to consider the application in a significant As is well known and has been already stated, the com- sampling that can cope with the uniqueness that the cultural plexity of the documentation and modelling of CH, as well as complexes present. being motivated by physical reasons, is reflected in the uses In other words, the validation concept that verifies the and in the extreme variety of interdisciplinary acknowledge- specified requirements was assumed to be sufficient for the ments and comparisons that the products must satisfy. intended use (ISO/IEC Guide 99:2007 & JCGM 200 2007). To control the overall metric quality of the ZEB clouds, their reliability was firstly considered based on their single use. The first statistical parameter that contributes to the eval- uation of the overall reliability was the accuracy, so some surfaces or clouds derived from more precise measurement systems were considered as ground truth to evaluate the devi- ation of the studied clouds by means of root mean square error (RMSE). It is also necessary to consider the precision of the clouds, because the precision is linked to the concept of re- peatability of the measurements and is normally described by the standard deviation (St.dev) parameter. Thus, for the ZEB clouds, the intrinsic precision that is linked to the acquisition mode has always been considered. With reference to the previous paragraph, the SLAM sys- tem, after calculating the raw trajectories, uses an iterative ICP-like process of automatic cloud-to-cloud profile registra- tion to generate the 3D cloud, and this process has always been controlled using loop paths. In addition, the correspon- dence of the detected surfaces during the time acquisition has been carefully optimised in a non-automatic way by segmenting the clouds and supplementary cloud matching and optimisation operations. Moreover, even if the ZEB scan- ner system is provided for the indoor and outdoor environ- ments, which has been studied in the aforementioned literature (a) (Díaz-Vilariño et al. 2017; Nocerino et al. 2017; Thomson et al. 2013;Zlot et al. 2014), the level of accuracy and detail and the noisiness of the clouds are quite different. Therefore, the validation strategy in this work includes datasets acquired in different outdoor and indoor configurations to account for these differences. Table 5 Statistical values of the deviation analysis of the O&R paths with the reference CRP model for the cylindrical tower of the Valperga castle Comparison of ZEB 3D data on ground truth Full raw Raw Optimised Roundtrip Outward Return Outward Return (b) Mean 0.025 0.026 0.024 0.025 0.022 Fig. 9 Deviation error representation in range colours for outward and St.dev 0.034 0.034 0.032 0.030 0.025 return ZEB point comparison (a) and zoomed excerpt (b) Appl Geomat Table 6 Statistical Outward and return comparison on ZEB values of the deviation surface analysis of the separated paths O&R for the Raw Optimised ancient ice house of Valperga castle Mean 0.078 0.047 St.dev 0.114 0.063 Lastly, since the need to obtain multiscale and multicontent heritage models has been ascertained and shared, the possibility of using the ZEB scanner in multisensor surveying configura- tions has also been investigated. In the validation strategy for this study, the abilities of integration relating to the ZEB clouds were compared with other cloud surfaces acquired with systems that offer different resolutions and accuracies and provide therefore different scales of surveys. In these cases, the dense clouds to which the ZEB data was (a) integrated were mainly derived from UAV photogrammetry, and the DSM and other photogrammetric products were generated to document the overall set of sites or clouds derived from TLS or CRP. To obtain the reference clouds (using UAV photogramme- try, TLS, or CRP techniques), the usual criteria and pipelines were adopted, and they are not described in this paper. For the generation of photogrammetric clouds, the orientation of the (b) blocks of images and the control of the results occurred through Fig. 11 Deviation error representation (top and lateral views) between using GCPs (Ground Control Points) and CPs (Check Points), or optimised ZEB point clouds of O&R in the ice house (scale ranges in in the LiDAR terrestrial applications, cloud recording with inte- Table 7). Top and side views (a) and zoomed excerpts (b) grated cloud-to-cloud alignment techniques and references (again GCPs and CPs). BTest dataset presentation^ provides the frame- work of the test datasets based on the selections related to the The first site corresponds to the castle of Valperga (Turin) validation strategy. built in a strategic defensive position on the top of a hill and forming a system of palaces and gardens Bat the French mode^ between the 17th and 18th centuries (see also Chiabrando Test dataset presentation et al. 2017a, b). The second site is the San Silvestro fortress located in the homonymous archaeological park that includes The datasets being validated belong to two projects of metric a territory rich in mines frequented from the Etruscan age documentation of two vast cultural complexes belonging to (VIII-I cent. BC) until the XX century. Both sites were sub- different construction periods that can be ascribed to architec- jected to a multisensor survey to obtain multiscale models tural heritage and the archaeological site. Fig. 10 Deviation error representation in range colours between raw ZEB point clouds of outward and return in the ice house (scale ranges in Table 7) Appl Geomat Table 7 Statistical results segmented in range error values from the been detected by means of a CRP survey and 3D modelling deviation analysis of O&R techniques from the structure-from-motion (SfM) algorithms (Chiabrando et al. 2017a, b). This 1-cm accuracy point cloud Comparison errors in ZEB surface outward and return has been the reference for evaluating the SLAM-based Raw (%) Optimised (%) dataset. 0.00 < error < 0.02 m Blue 78.5 81 0.02 < error < 0.05 m Cyan 17 16 Ice cellar (B) 0.05 < error < 0.10 m Green 2.8 2.2 0.10 < error < 1.00 m Red 1.7 0.8 Historical buildings often contain surprises; from an entrance of one of the buildings surveyed at the architectural scale in the Valperga complex, a long and dark corridor starts and goes derived from terrestrial and aerial techniques and formed by under two blocks of the castle, gradually descending and lead- the integration of datasets at different resolutions and scales. ing to the ice cellar. This path and this buried space seemed The sites can be read in the first column of Table 2, while the excellent to challenge the potential of the SLAM-based ZEB second and third columns state whether the recorded environ- system. Also, in this case, the acquisition using the handheld ments were indoor or outdoor. The (A), (B), and (D) test datasets ZEB scanner was performed by completing the roundtrip, regarding the inside of a cylindrical tower, an underground ice starting from the courtyard (Fig. 4b; ZEB1 dataset house, and a mining cave were evaluated using the stand-alone 13,900,000 pts./6 min). validation. The ZEB point clouds recording the courtyard in Valperga castle (C) and the dataset covering the inside paths of Courtyard (C) the Rocca of San Silvestro (E) were validated instead in their integration of the whole 3D models of the sites involving UAV, The integrated image and range-based survey at Valperga castle TLS, and CRP clouds. was planned with the aim of merging the DSM derived from UAV photogrammetry computed by nadir and oblique images Tower (A) (Fig. 5b), with the dense and very accurate models of the TLS technique by the FARO Focus 3D X120 scanner (Fig. 5a). The The first dataset considered was the cloud acquired along the use of the MMS ZEB scanner provided the opportunity to eval- narrow and restricted spiral staircase that runs inside a cylindrical uate the use in such cases of buildings so densely packed with tower of the castle of Valperga (Fig. 3a). The ZEB1 acquisition narrow courtyards that are not suitable for photogrammetric was performed starting from the outside and covering the stair- surveys of facades. In addition, the use of the scanner could way up to the dovecote and back to the entrance on the ground avoid the heaviness and density in the usage of terrestrial floor. (ZEB1 dataset 19,000,000 pts./10-min time acquisition). LiDAR clouds. This was an opportunity then to evaluate the The structural and material degradation of the staircase and use of ZEB clouds in a relevant multisensor survey context. the helical vault of marked constructive interest had already (Fig. 2; ZEB1 dataset 8,200,000 pts./4 min). Fig. 12 Range colours representing elevation on the surface topography by ZEB- REVO, with human size, from the street level (right), up to the entrance (black arrow, + 5 m, red), down to the lower level (blue, almost − 17 m) Appl Geomat Fig. 13 O&R raw point clouds refer to the Buca della Faina medieval mine and stress, in the excerpt, the closure error. The entrance location is represented by the black arrow Mining cave (D) necessarily heavy and time-consuming and was therefore an ide- al site to test the validity of the clouds acquired by the ZEB Having ascertained that the handheld ZEB system is profitable system. for the modelling of underground environments such as quarries The loop acquisition was laid along the entire visit path of the and mines, the scanner was tested in the medieval mine called fortress, starting from the entrance to the east and following Buca della Faina (Fig. 6; ZEB-REVO dataset 41,700,000 pts./ exactly the ancient ascent to the culminating part of the fortified 25 min). village and rearing from the other side according to the ancient This mine is strange because it can be traversed in many road that embraces the cone-shaped Rocca. places only on all fours, and the cloud was only collected with In 2016, a ZEB-REVO dataset was collected along the the help of speleologists. The slowness of acquisition explains whole circular pathway around the Rocca (33,900,000 pts./ the point density of this cloud, which counts 42 million points 23 min). In 2017, a second dataset (7,200,000 pts./8 min) compared to 33 million of the cloud acquired in the fortress (E). was conceived as an integration of the area and is shown in Fig. 7 with an outward and return track. Fortified village (E) Metric validation in stand-alone solution Even the multistratified site of San Silvestro, with its A stand-alone use of point clouds derived by this SLAM-based safeguarding landscape and archaeological heritage, was subject mobile mapping is conceivable due to the intrinsic metric values to UAV and TLS acquisition and modelling. The densely built of the endorsed concept of the so-called 1:1 scale of ranging area needed detailed 3D LiDAR survey and models because it is measurements returning, and this method is helpful in mapping subject to consolidation and restoration works since it lies on a indoor volumes wherever georeferencing issues are not required. slope of the hill presenting landslides. Once known and admitted as acceptable, the reliability verified The TLS survey, operated with a FARO Focus 3D X120 on the reference model for (A) (the designated CRP model for scanner, for the study of construction systems of masonries was the indoor space of the tower), the confidence level of the ZEB Table 8 Statistical values of the O&R comparison subjected to the Table 9 Statistical values of deviation errors grouped into ranges optimisation process related to the main optimisation steps Outward and return comparison on ZEB surface Outward and return comparison errors in ZEB surface Raw Optimising steps Raw (%) Filtering (%) In solo (%) Cleaning in/out Filtering in/out In only 0.00 < error < 0.10 m 69 87.4 88.5 Mean 0.214 0.175 0.103 0.055 0.10 < error < 0.20 m 22.7 4.5 4.3 St.dev 0.313 0.183 0.131 0.070 0.20 < error < 0.6 m 8.3 8.1 7.2 Appl Geomat Fig. 14 Range colour representation of deviation analysis errors for the only internal point cloud (scale ranges in Table 9) reconstruction in such indoor scenarios such as the ancient ice common noise factor in the ZEB surface. For these reasons, the house (B) and the archaeomining cave (D) can be reasonably alignment was evaluated without and after a process of circumscribed and validated for a related scale use. optimisation (segmentation, outliers cleaning, noise filtering). The problems about the establishment of the relationship with another reference surface are those problems that are primarily necessary to face because of the already clarified non-existence The cylindrical tower (A) of positioning data and the lack of direct radiometric content in the raw data. Some referencing issues emerge in the operative The CRP reconstruction inside the tower presented in Fig. 3b fulfilment of this purpose. A first solution can be the matching of is considered the ground-truth surface to validate the ZEB tie points that are detectable targets on both the 3D point cloud mobile mapping, and this surface has been computed with a with x, y,and z coordinates measured as references. controlled error propagation of less than 1 cm on GCPs and Complications deriving from this discrete method include the about 1.5 cm in the CPs (Table 3). difficulty in recognising and choosing the exact point; however, First, the cloud-to-cloud best fitting alignment on the CRP a statistical evaluation by means of RMSE on many matched model of the full raw ZEB surface returned a first mean value of point distances can grant the accuracy assessment (Farella et al. distance deviation of 0.025 m and a St.dev of 0.034 m (reported 2016) of the action. A more effective solution of point cloud in the first column in Table 5), showing 67% of the points that alignment is commonly offered by the control of deviation errors actually deviated from the reference model of a value error < on the performance of an ICP-like algorithm, a so-called cloud- 0.02 m and the 26% between 0.02 < error < 0.05 m (Fig. 8a). to-cloud, between the ZEB surface on the other point clouds (TLS; CRP). This method, more continuous along the whole Table 11 Metric control for the Valperga castle scans: the accuracy considered surface, undergoes possible different precision of validation on the LiDAR point cloud registration shows a mean error of about 1 cm on target check points and a mean value of 4 mm on the clouds the surface characterisation in local details and suffers the comparison Table 10 RMSE on control points in the UAV photogrammetric Metric control on the scans registration reconstruction of Valperga castle (mm) Cloud-to-cloud alignment Residual error on Metric control on the UAV photogrammetric DSM target points (m) XY Z Plan Tot Mean dev. Dev. error Mean Max error <4 mm GCP 0.009 0.012 0.021 0.015 0.026 CP 0.012 0.009 0.025 0.014 0.029 Courtyard 3.35 55% 12 20 Appl Geomat Fig. 15 The courtyard wall surface in a ZEB profile reconstruction and b LiDAR DSM. c Deviation error representation in coloured range values, as stated in Table 12 The most significant discrepancies > 5 cm are recognisable roundtrip, which is helped by the time-marked trajectory, two- on the treads of the steps and in horizontal surfaces as windows fold clouds are obtained: the first trajectory, outward (O), and the and doors along the rising trajectory, as shown in Fig. 8b. In second trajectory, the return (R), take 6 min plus 4 min, particular, the intermediate values of deviation, 2–5cmingreen colour, are spread in the spiral gradient of the stair ceiling and in Table 12 Statistical Comparison errors between ZEB and values divided into errors the north side of the upper volume, as shown in Fig. 8a. The LiDAR surfaces ranges corresponding to figure also shows the arrival of the outward trajectory as the Fig. 15c farthest from the starting point of the scanning. Mean 0.017 m For these reasons, some crucial issues occur when the se- St.dev 0.023 m quent validation is performed, separating the roundtrip and go- 0.00 < error < 0.02 m 88.8% ing into the ZEB cloud about its own precision evaluation, and 0.10 < error < 0.05 m 10.6% these issues are strictly related to the precision of the operating 0.05 < error < 0.20 m 0.63% principle on which the system is based. By separating the raw Appl Geomat respectively (11,300,000 + 7,700,000 points). The relative align- ment comparison, as listed in Table 4, retrieves critical values if the validation is performed without the cited optimisation of the surface points, denouncing a moderate problem of the SLAM- based alignment between outward and return (O&R). For this type of path, Fig. 3a shows the coupling of the alignment with the noisy effect of the raw point cloud in outlier errors, affecting higher values. An effective optimisation operation, as previously cited, on both O&R provided the more suitable values in Table 4,and these values are provided in Fig. 9 and ensure greater reliability of the SLAM-based profile alignment solution implemented in the ZEB1. The reliability can be strengthened in the observation along the whole tower elevation in Fig. 9a and in the zoomed excerpt in b, where the figure shows the reduced deviation be- tween the O&R mainly bordered in the steps and in the roofing intrados. Deviation distance errors between O&R are statistical- ly represented with a mean value of 2 cm and a St.dev of ap- proximately 3 cm. Specifically, 93% of the points appear with deviation error values under 2 cm (Fig. 9a); however, the 99.5% is under 5 cm. This precision proof cannot be admitted in a 1:50/ 1:100 scale of representation. Furthermore, a second validation with the reference CRP model is thus consequently proposed: after both the separation values in O&R and before and after optimisation execution values are presented in Table 5. Flanking is presented in the first column, and the initial values relate to the introduced full raw comparison. The Table 5 values of comparison with the reference show the enhancing obtained in the discrepancies by optimisation Fig. 16 The compared metric surfaces: ZEB1, blue, and TLS, gold (a). operation applied on the O&R ZEB surfaces, in which the The extracted profile comparison on the courtyard pavement in the railing discrepancies are affected by outliers and noise errors. The step (b). Thick stroke for the TLS profile, thinner for ZEB, and deviation values in Table 5 display, however, a not very remarkable error in range colours (c) improvement in deviation values from the CRP reference Fig. 17 The complete registration of several ZEB point clouds on the reference LiDAR DSM. The ZEB initialisation point is identified as a red star. The other surface integrations of sensors (ZEB and UAV) are indicated in A-A′ and B-B′ Appl Geomat Table 13 Statistical values of the deviation between the ZEB and the Table 14 RMSE on control points in the UAV photogrammetric TLS reference in the same courtyard reconstruction of the Rocca ZEB surface comparison with the LiDAR DSM Metric control on the UAV photogrammetric DSM Courtyard Ancient ice Basement Mezzanine (m) XY Z Plan Tot house floor floor GCP 0.012 0.017 0.024 0.021 0.032 Mean 0.065 0.057 0.052 0.053 CP 0.012 0.009 0.018 0.015 0.023 St.dev 0.090 0.080 0.074 0.076 underground path that leads to the ice house and to the mapping model, proving the intrinsic accuracy. This improvement is of the whole hypogeum volume. limited to a mean of 2.2–2.5 cm with a St.dev of about 2.5– The comparison between O&R from the raw surface recon- 3 cm for both the O&R. struction that had the courtyard as a starting point (necessary to In particular, the outward values can be assimilated to those connect from a common area the whole 3D ZEB reconstruction, values representing the full raw comparison. Therefore, the return as the following presented) returned the subsequent values in point cloud reconstruction based on the SLAM profile alignment Table 6. The deviation distances, as introduced, suffer from noise approach actually benefits from the enclosed environment, and and outlier errors, and the optimised process leads to a reduction its reconstruction, as shown in the last column in Table 5, shows in discrepancies of approximately 50%. a better accuracy validation with a 2.2-cm mean value from the If deviation maps are examined, as shown in Fig. 10,tocom- reference photogrammetric model. pare the raw O&R point clouds, and in Fig. 11 for the optimised and localised surfaces, the establishment of range values of de- The ancient ice house (B) viation errors distances, as reported in Table 7, allow the inter- pretation of some key aspects. According to the same principle of validation, in the (B) dataset, In Fig. 10, while in the arrival to the ice house, in the farthest the verification of deviation error values has been compared point from the initialisation, the ZEB SLAM system aligned both before and after the optimisation approach, i.e. with the whole the raw O&R profiles with deviation errors referable to the blue point cloud and surrounding, and then restricted to the colour for most of the hypogeum area, and the comparison of starting and arriving profiles in the courtyard significantly affect- ed the statistical values, as shown in Tables 6 and 7. The optimi- sation operation, whose comparison errors are represented in Fig. 11 and listed in Table 7, replaced more significant values supporting the precision aptitude of the system in the 3D map- ping of such narrow passages, but some critical areas remain in specific details, in particular, the values in green/red colours. These decreases of precision are identifiable, for example, in the ceiling of the starting portion of the passage towards the ice house, at the beginning of the scanning process; in some parts of the volume shown in Fig. 11a, in which these decreases are due to the space occupied by unrelated masses differently responding to the lightwave signal; in the ventilating outlet elements; and in some vertical planes (Fig. 11a) during the descending passage, which are differently modelled by the outward and return profiles Table 15 Metric control on LiDAR scans in the Rocca Metric control on the scan registration (mm) Cloud-to-cloud alignment Residual error on target points Mean dev. Dev. error Mean Max error <4 mm Fig. 18 Extracted sections, A-A′ crossing the basement and mezzanine floors and B-B′ crossing the entrance to the tunnel towards the ice house, Village 4.58 50.8 18 35 connected with the volume in the mezzanine floor Appl Geomat Fig. 19 Alignment strategies: a (I) ICP-like deviation error range values, explicated in Table 16; b (II) matching points, whose RMSE is also explicated in Table 16 possibly due to the topographic variation influence in the SLAM- the very impressive number of points during the almost 26 min 3D reconstruction and due to the uniformity of some tunnel (nearly 1.7 million/min) of the considerably limited trajectory areas. length (only ~ 100-m roundtrip, almost 0.1 m/s) executed in a It is necessary to considerate, however, the blue-cyan colour very intricate cave shown in Fig. 12, which featured limited profusion that is equivalent to a deviation error between O&R < accessibility for irregular topography, impervious spaces, signif- 2cmand2cm<err<5cm.Forthe optimisedsurfaces,the devi- icantly harsh surfaces, and reduced light conditions. ation error corresponds to 81% of the points for the former and The verification of the O&R deviation error on the ZEB 16% for the latter, comprising an entirety of 97% of points that points to validate and support the intrinsic precision of the deviate in the global SLAM-based alignment of lower than 5 cm. SLAM-based operating system that underwent restrictive condi- tions was performed in two phases, before and after the optimi- The mining cave (D) sation. In this case, with the massive amount of irregular vegeta- tion, it caused a noise error affecting the entrance zone and from To validate the ZEB 3D mapping system in such a peculiar the starting point to the arrival closure on the street level. framework as the Buca della Faina sample shown in Fig. 6, The first segmentation in O&R parts consists of 14-min plus the reasoning about the results should consider the local condi- 12-min paths (22,170,642 + 19,499,539 points), from whose tion under which the acquisition has been performed, which is comparison a problem of loop closure occurred, as in Fig. 13, Appl Geomat Table 16 Statistical values for strategies I, absolute and divided in of almost 40 cm. This loop closure distance result is within 8.3% percentage of points per error ranges, and II, with residual deviation of the point value, as shown in Table 9, meaning a mean devia- error on matching points tion error between 0.20 and 0.60 m. The deviation distribution Comparison errors between ZEB and UAV surfaces along the whole path is, as shown in the first column of Table 8, defined into 0.2 m with a St.dev of approximately 0.30 m. IMean 0.531m The high values are most likely conceivable because of the St.dev 0.715 m diffuse noise errors and of the weight in the statistical evaluation 0.00 < error < 0.10 m 43% of the densely vegetated area at the entrance. 0.10 < error < 0.50 m 53% Based on these considerations, the validation approach tack- 0.50 < error < 1 m 4% led the possibility of properly segmenting the cloud and proceed- II RMSE CPs 0.047 ing to the optimisation steps, dividing the data densification in- side the mining cave at the entrance from the rest of the cloud outside. The two steps of cleaning and filtering are presented in Two samples of 12 × 7 m segments have been selected and the central columns of Table 8 and have been conducted optimised, as shown in Fig. 15a, b: 450,000 pts. for the ZEB parallelly to both segments, differentiating the parameter values segment and 5,600,000 for the LiDAR DSM, which is more than and tailoring them according to the area (in, out). In the last step, 12 times denser than the ZEB one. From the comparison analysis the validation considers the deviation results only related to the showninFig. 15c, the detected deviation errors with mean values O&R for the surface into the cave, and the deviation is presented ofalmost 2cm(2.3cmSt.dev), asreportedinTable 12, confirm in the final column of Table 8. the possibility to control the accuracy of the SLAM-based system Statistical results on deviation errors from the explained com- in an outdoor scenario between a 1:100 and 1:200 scale of rep- parison have been grouped into simpler range values, in which resentation, in which 99% of points deviate from LiDAR less the results are shown in Table 9, allowing better appreciation of than5cm. the optimisation result improvement in the ZEB 3D reconstruc- TheproblemshowninFig. 15c indicates the deviation (green) tion. Figure 14 presents the graphic representation of the in solo increasing with the height, which is due to a kind of systematic deviation analysis, whose values are reported in Table 9.The error related to the mechanical system of distribution mode of challenging 3D survey into the Buca della Faina reported a final laserraysinZEB1,as confirmedinCadge (2016) (this has been evaluation on its own accuracy and established its confidence improved in profile uniformity coverture by the ZEB-REVO level for metric purposes: the obtained value is a mean error on system). However, the tangible limited descriptive capabilities O&R based on the only cave of about 5.5 cm with a St.dev of intrinsic in the ZEB system, as expected, are localised in archi- 7cm. tectural details and edges, as shown in Fig. 16, and they become rounded and approximated. Analytical range values display Metric validation in the multisensor survey The challenging aspects related to the use of this kind of spatial data, whose own precision has been investigated and accuracy evaluated in the previous section, are now faced based on its compatibly and integrability with other multisensory data on a multiscale survey organisation. The use of DSM coming from TLS with higher accuracy and UAV photogrammetry with sig- nificant 3D completeness and continuity is used in this phase and integrated to the ZEB reconstruction to evaluate their comple- mentarity or exchangeability. The Valperga castle courtyard (C) The integration of the Valperga castle surveyed volumes is performed with LiDAR scans and UAV photogrammetric DSM. The confidence level is based on the reliability of their Church metric reconstruction declared in Table 10 for the UAV pho- togrammetric reconstruction and Table 11 for the TLS and Fig. 20 UAV orthoimage of the San Silvestro Rocca with indications of corresponded to an admitted scale of representation between the sample area in the lower village, the church, and the climbing path to the upper area 1:50 and 1:100. Appl Geomat Table 17 Church: statistical values for strategies I and II Table 18 Village: Comparison between ZEB and reference statistical values for surfaces Comparison errors between ZEB and UAV surfaces surface comparison UAV Mean 0.057 I Mean 0.578 m St.dev 0.077 St.dev 0.771 m TLS Mean 0.072 II RMSE CPs 0.126 St.dev 0.091 errors of ± 1 cm (green) for 72.7%, almost 20.7% for + 1 cm < error < + 2 cm and 6.3% for −1<error< −5cm. has been conducted (~ 6000 m ), and in the loop mode (~ In the proposed validation, the TLS DSM served as a basis for 660 m), i.e., the circular closed path, avoids the roundtrip the configuration and referencing of a system of roundtrip ZEB intended because the return path is along the same outward path. scans starting from initialisationinthecourtyardsettinganddi- Due to the critical issues clarified, a second dataset in 2017 rected to the several castle volumes in Fig. 17. was collected in an ~ 2000-m area and in the roundtrip loop All the SLAM-based mappings belong to the LiDAR DSM as mode (~ 450 m) Fig. 7. reference and deviation error values are presented in Table 13. The reference surfaces employed in this phase are the UAV These error values are related to the raw alignment via the cloud- photogrammetric DSM, whose planimetric and vertical reli- to-cloud method and initial control of deviation distances. With able accuracy are reported in Table 14, and the LiDAR scans better control of error propagation in such a kind of articulated are restricted to the rooms at the entrance area of the village, 3D mapping via optimisation procedures of the single clouds, as whose metric control is reported in Table 15. cited, the accuracy related to the 1:200 scale of documentation is The problem of referencing and validating the ZEB scan a more than achievable and a very reliable goal for the ZEB laid on the entire area of the village was performed in a circu- system. lar loop. First, Fig. 7 shows the quality-marked trajectory Starting from the integration of 3D data in this multiscale superimposed to the width of the SLAM-based point cloud. and multisource model, some schematic cut sections, as The best working areas are generally those areas that have shown in Fig. 18, prove the possible integration with the aerial been travelled twice. DSM by UAV photogrammetry. If a deviation analysis is performed on the comprehensive These sections can show the added value of this measurement highly detailed UAV DSM, as shown in Fig. 19,tovalidatethe technology to effectively support the spatial analysis of architec- behaviour of the ZEB surface on such a kind of reliable re- tural environments based on its verified reliability and scale ac- construction, a crucial problem emerged, as reported in the curacy and in relation to the more consolidated TLS approach. form of statistical analysis in Table 16.It is reasonablyattrib- utable to the drift error affecting the circular path mode in such an irregular topographic setting, where open spaces and nar- The fortified village (E) row passages alternate, straining the SLAM-based alignment algorithm operation. In fact, the analysis of the whole fitting The validation performed in the second outdoor architectural complex adds more critical issues to the testing of the ZEB verification retrieved higher values distributed in the path. By choosing the correspondences of some recognisable matching system in 2016 because the wide area on which the trajectory Fig. 21 Alignment strategies: (left) I, ICP-like with the entire ZEB point cloud; (right) II, local alignment of segmentation by matching points Appl Geomat Table 19 Path: statistical Comparison between ZEB and UAV values for surface surfaces comparison Raw Mean 0.077 St.dev 0.113 Opt Mean 0.060 St.dev 0.093 points (i.e., corners) as the example of Fig. 19, the residual errors are locally reduced, as shown in Table 16,confirming the strong need of the system to base its operational effective- ness of featuring geometries in the environment, especially if Fig. 23 Return of the achievable level of details in a sample triangulated an outdoor scenario is considered. mesh computed from the optimised ZEB surface According to a profitable optimisation, the approach that can be deployed is based on filtering, segmenting, and separating alignment underwent a linear deviation along the path in those non-improved parts, noising errors, and encumbering elements, cases where the trajectory quality denounced criticalities. such as rich vegetation inside the village. For example, in the higher area of the church, as identified in The slight resulting variation obtained by this intermediate Fig. 20, the greater discrepancy is pinpointed with values that phase confirmed the intrinsic precision problem in the trajectory locally reach a mean value of almost 0.6 m with a St.dev of deviation, and a focused segmentation approach was performed 0.7 m, as shown in Table 17. If the locally segmented surface to verify the drift areas and localise where the SLAM-based is aligned by means of matching points in Fig. 21,the controlon Fig. 22 Comparison analysis between UAV DSM (blue) and ZEB DSM (red): a mapping of the discrepancies in two ranges 0– 5 cm and 5–10 cm; b extracted section profiles describing steps Appl Geomat residual errors on distances provides much better values, with an clouds, as seen when the marked time trajectory must be order of magnitude of 10 cm, as shown in Table 17. carefully controlled manually and when the essential aid The segmented sample area in the lower village, as shown in of the colour ramp quality must be optimised). Fig. 20, should be measured, and in this case, the 2017 ZEB 3D & Compactness and handiness for unfriendly environments. reconstruction was validated due to the beneficial O&R trajecto- & Problems related to the lack of high-quality radiometric ry mode employed for updating that area. By flanking the UAV data. (The update started in 2017 with the equipment of DSM surface comparison, the deviation analysis with a LiDAR a GoPro camera. The images can be projected on the sur- point cloud was computed with the TLS approach and added as a face points in post-processing, facilitating the point reference surface for the validation. matching for scan registration, segmentation, and other The residual values, which are summarised in Table 18,for analyses of the clouds). this sample confirm the improved SLAM-based registration on & Demand of alignment verifications for loop configuration the second dataset (2017), reasonably deviating more from the and closure. highly detailed TLS surface than the UAV model. This increase & Need of validation in very high scale and small detail. At can be corroborated by the analysis and validation of the ZEB present, the scale of ZEB-derived models can be consid- (2017) surface on the climbing path to the church. Table 19 ered to correspond to what the tradition calls the architec- reports the comparison values between ZEB and UAV DSM tural scale, that is, 1:100 or 1:200. For a full availability for the segmented narrow footpath made by steep steps, as shown and handling of architectural details, a scale 1:50 or above in Fig. 22a, with lateral surfaces in a parallel analysis, pre-/post- is required, and more accurate clouds derived from the optimisation. The analysis is finalised to evaluate, through their TLS technique continue to be needed. precision, the descriptive capabilities of both these rapid mapping approaches to replace a demanding TLS deployment. By changing the perspective, this last point can show a new If the segments of the stair samples are considered and both vision. The clouds derived from the roundtrip ZEB portable the contributions of the techniques are evaluated and compared, scans are lighter and easier to handle than those of the LiDAR the ZEB reconstruction, as shown in Fig. 22b, potentially sat- technique, and in fact, this study shows how other authors pre- isfies the scale detail of the mapping purposes from the ground, figure a promising development perspective in the construction supporting and even perfecting the aerial photogrammetric pur- of 3D city models of historic cities and the possibility of facing poses. Referring to the comparison of points shown in Fig. 22a, big data problems. 0.00 m < 91.7% pts.<0.05mand0.05m< 8.3% pts. < 0.10 m In addition, the possibility of reducing the use of the time- (higher values are related, as visible, to the railing modelled by consuming LiDAR technique to only indispensable occasions theZEB andnot reachedbyUAV). has deep implications in the field of CH. As also emerged in discussions at conferences such as at the The optimised ZEB surface can also profitably support the surface triangulation, as shown in Fig. 23, returning a Geores2017 (Ottonietal. 2017), the meaning of Bmonumental remarkable level of detail, i.e., in step edges. complexes^ aims to identify a very wide range of characterisa- tion of assets and actions of documentation projects. The characterisations can range from the physical complexity of the Conclusion and future perspectives contexts, the extension and articulation of study objects, and immediately connect the concept of Bcomplex sites^. The cloud tests and analyses performed confirm an extremely Therefore, the need for multiscale studies, approaches that positive judgement on the increased use of MMS GeoSLAM foresee the need for multidisciplinary interaction in studies ZEB in the documentation of CH in the literature. Currently, and analyses, and outcomes that lead to the planning of con- the main problem in applying this technology is the lack of servation and restoration interventions articulated according to radiometric information that would seem to be the next update, the various static-structural aspects, chemical materials, and starting with the recent introduction of a commercial off-the-shelf innovations of materials must be considered. Particularly sig- (COTS) camera whose images are not yet fully fused with the nificant are the number of dedicated contributions, the themes range data. of the use of HBIM platforms, and the contributions focusing The actual benefits of the handheld system, flanking by the derivation from reality-based models of other models suit- critic points, can be summarised, and they are the following: able for structural analyses in the static and dynamic field (Ottoni et al. 2017). & Efficiency in indoor spaces and articulated architectural, The request for multilevel modelling is an increasing issue; landscape, and archaeological massings where GNSS po- therefore, the studied technology that effectively fills the sitioning is not available. scales between the architectural detail and the building/urban & Rapidity in both acquisition and processing (this consid- scale enables intense and increasingly specialised uses to be prefigured. eration is mainly addressed to the automatic pre-process of Appl Geomat Farella E, Menna F, Nocerino E, Morabito D, Remondino F, Campi M Compliance with ethical standards (2016). Knowledge and valorization of historical sites through 3d documentation and modeling. ISPRS—International Archives of the Competing interests The authors declare that they have no competing Photogrammetry, Remote Sensing and Spatial Information interests. Sciences, XLI-B5,255–262. https://doi.org/10.5194/isprsarchives- XLI-B5-255-2016 Open Access This article is distributed under the terms of the Creative GeoSLAM CSIRO research (2016) http://geoslam.com/slam/csiro- Commons Attribution 4.0 International License (http:// research/ creativecommons.org/licenses/by/4.0/), which permits unrestricted use, ISO/IEC Guide 99:2007, JCGM 200 (2007) International vocabulary of distribution, and reproduction in any medium, provided you give appro- metrology—basic and general concepts and associated terms (VIM). priate credit to the original author(s) and the source, provide a link to the Available at: http://www.iso.org/sites/JCGM/VIMintroduction.htm Creative Commons license, and indicate if changes were made. Lingua A, Noardo F, Spanò AT, Sanna S, Matrone F (2017). 3D model generation using oblique images acquired by UAV. ISPRS— International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, XLII-4/W2,107–115. https://doi. org/10.5194/isprs-archives-XLII-4-W2-107-2017 References Mandelli A, Fassi F, Perfetti L, Polari C (2017). Testing different survey techniques to model architectonic narrow spaces. ISPRS— Bauwens S, Bartholomeus H, Calders K, Lejeune P (2016) Forest inven- International Archives of the Photogrammetry, Remote Sensing tory with terrestrial LiDAR: a comparison of static and hand-held and Spatial Information Sciences, XLII-2/W5,505–511. https://doi. mobile laser scanning. Forests 7(6):127. https://doi.org/10.3390/ org/10.5194/isprs-archives-XLII-2-W5-505-2017 f7060127 Murtiyoso A, Koehl M, Grussenmeyer P, Freville T (2017). Acquisition Bellekens B, Spruyt V, Weyn M (2014). A survey of rigid 3D pointcloud and processing protocols for UAV IMAGES: 3D modeling of his- registration algorithms. AMBIENT 2014, The Fourth International torical buildings using photogrammetry. ISPRS Annals of Conference on Ambient Computing, Applications, Services and Photogrammetry, Remote Sensing and Spatial Information Technologies. 2014.,(c),8–13 Sciences, IV-2/W2,163–170. https://doi.org/10.5194/isprs-annals- Bosse M, Zlot R (2009). Continuous 3D scan-matching with a spinning IV-2-W2-163-2017 2D laser. In 2009 IEEE International Conference on Robotics and Nikoohemat S, Peter M, Oude Elberink S, Vosselman G (2017). Automation (pp. 4312–4319). IEEE. https://doi.org/10.1109/ Exploiting indoor mobile laser scanner trajectories for semantic in- ROBOT.2009.5152851 terpretation of point clouds. ISPRS Annals of Photogrammetry, Bosse M, Zlot R, Flick P (2012) Zebedee: design of a spring-mounted 3- Remote Sensing and Spatial Information Sciences, IV-2/W4,355– D range sensor with application to mobile mapping. IEEE Trans 362. https://doi.org/10.5194/isprs-annals-IV-2-W4-355-2017 Robot 28(5):1104–1119. https://doi.org/10.1109/TRO.2012. Nocerino E, Menna F, Remondino F, Toschi I, Rodríguez-Gonzálvez P (2017). Investigation of indoor and outdoor performance of two Brocchini D, Chiabrando F, Colucci E, Sammartano G, Spanò A, Losè portable mobile mapping systems. In F. Remondino & M. R. LT, Villa A (2017). The geomatics contribution for the valorisation Shortis (Eds.), Videometrics, Range Imaging, and Applications project in the Rocca of San Silvestro landscape site. In International XIV, edited, SPIE 2017 (p. 103320I). https://doi.org/10.1117/12. Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences—ISPRS Archives (Vol. 42). https://doi.org/ Ottoni F, Rodriguez A, Spanò AT (2017). La documentazione dei 10.5194/isprs-Archives-XLII-5-W1-495-2017 complessi monumentali. In ANANKE Speciale GEORES 2017 (pp. Cadge S (2016). Welcome to the ZEB revolution. GEOmedia (3)22–26 13–14) Chiabrando F, Naretto M, Sammartano G, Sambuelli L, Spanò AT, Puente I, González-Jorge H, Martínez-Sánchez J, Arias P (2013) Review Teppati Losè L (2017a). The 4Dilan project (4Th Dimension in of mobile mapping and surveying technologies. Measurement Landscape and Artifacts Analyses). ISPRS—International Archives 46(7):2127–2145. https://doi.org/10.1016/j.measurement.2013.03. of the Photogrammetry, Remote Sensing and Spatial Information 006 Sciences, XLII-5/W1, 227–234. https://doi.org/10.5194/isprs- Remondino F, Spera MG, Nocerino E, Menna F, Nex F (2014) State of archives-XLII-5-W1-227-2017 the art in high density image matching. Photogramm Rec 29(146): Chiabrando F, Spanò AT, Naretto M, Tocci C (2017b). Strati dei palinsesti 144–166. https://doi.org/10.1111/phor.12063 e modellazione spaziale multi-scala: analisi integrate per la Riisgaard S (2005). SLAM for dummies: a tutorial approach to simulta- conservazione di sistemi culturali. In: ANANKE Speciale neous localization and mapping. Citeulike.org. ht tps://doi.org/10. GEORES 2017, Altralinea, pp. 135–140 1017/S0025315400002526 Dewez TJB, Plat E, Degas M, Richard T, Pannet P, Thuon Y, … Lucas J Rodríguez-Gonzálvez P, Jiménez Fernández-Palacios B, Muñoz-Nieto Á, (2016). Handheld mobile laser scanners Zeb-1 and Zeb-Revo to map Arias-Sanchez P, Gonzalez-Aguilera D (2017) Mobile LiDAR sys- an underground quarry and its above-ground surroundings. In nd tem: new possibilities for the documentation and dissemination of Virtual Geosciences Conference: VGC 2016.Bergen, Norway. large cultural heritage sites. Remote Sens 9(3):189. https://doi.org/ Retrieved from https://hal-brgm.archives-ouvertes.fr/hal-01348956 10.3390/rs9030189 Díaz-Vilariño L, Verbree E, Zlatanova S, Diakité A (2017). Indoor model- Ryding J, Williams E, Smith M, Eichhorn M (2015) Assessing handheld ling from SLAM-based laser scanner: door detection to envelope mobile laser scanners for forest surveys. Remote Sens 7(1):1095– reconstruction. ISPRS—International Archives of the 1111. https://doi.org/10.3390/rs70101095 Photogrammetry, Remote Sensing and Spatial Information Sammartano G (2017). Optimization of 3D multi-sensor models for dam- Sciences, XLII-2/W7, 345–352. https://doi.org/10.5194/isprs- age assessment in emergency context: first tests on rapid mapping in archives-XLII-2-W7-345-2017 the 2016 Italian earthquake. In F. Remondino, A. Georgopoulos, D. Eyre M, Wetherelt A, Coggan J (2016) Evaluation of automated under- Gonzalez-Aguilera, & P. Agrafiotis (Eds.), Latest Developments in Reality-Based 3D Surveying and Modeling. MDPI ground mapping solutions for mining and civil engineering applica- tions. J Appl Remote Sens 10(4):46011. https://doi.org/10.1117/1. Sammartano G (2018) Suitability of 3D dense models for rapid mapping JRS.10.046011 strategies on cultural heritage documentation and conservation. Appl Geomat Validation of metric and non-metric information extraction from with advanced statistical methods. ISPRS—International Archives of the Photogrammetry, Remote Sensing and Spatial Information integrated solutions. PhD thesis, Politecnico di Torino Sirmacek B, Shen Y, Lindenbergh R, Zlatanova S, Diakite A (2016). Sciences, XL-5/W4,245–253. https://doi.org/10.5194/ Comparison of ZEB1 and LEICA C10 indoor laser scanning point isprsarchives-XL-5-W4-245-2015 clouds. ISPRS Annals of Photogrammetry, Remote Sensing and Tucci G, Bonora V, Conti A, Fiorini L (2017). Digital workflow for the Spatial Information Sciences, III-1, 143–149. https://doi.org/10. acquisition and elaboration of 3D data in a monumental complex: 5194/isprsannals-III-1-143-2016 the fortress of Saint John the Baptist in Florence. ISPRS— Thomson C, Apostolopoulos G, Backes D, Boehm J (2013). Mobile laser International Archives of the Photogrammetry, Remote Sensing scanning for indoor modelling. ISPRS Annals of Photogrammetry, and Spatial Information Sciences, XLII-2/W5, 679–686. https:// Remote Sensing and Spatial Information Sciences, II-5/W2,289– doi.org/10.5194/isprs-archives-XLII-2-W5-679-2017 293. https://doi.org/10.5194/isprsannals-II-5-W2-289-2013 Vanneschi C, Eyre M, Francioni M, Coggan J (2017) The use of remote Toschi I, Ramos MM, Nocerino E, Menna F, Remondino F, Moe K, … sensing techniques for monitoring and characterization of slope in- Fassi F (2017). Oblique photogrammetry supporting 3D urban re- stability. Procedia Engineering 191:150–157. https://doi.org/10. construction of complex scenarios. ISPRS—International Archives 1016/j.proeng.2017.05.166 of the Photogrammetry, Remote Sensing and Spatial Information Zlot R, Bosse M, Greenop K, Jarzab Z, Juckes E, Roberts J (2014) Sciences, XLII-1/W1,519–526. https://doi.org/10.5194/isprs- Efficiently capturing large, complex cultural heritage sites with a archives-XLII-1-W1-519-2017 handheld mobile 3D laser mapping system. J Cult Herit 15(6): Toschi I, Rodríguez-Gonzálvez P, Remondino F, Minto S, Orlandini S, 670–678. https://doi.org/10.1016/j.culher.2013.11.009 Fuller A (2015). Accuracy evaluation of a mobile mapping system
Applied Geomatics – Springer Journals
Published: Jun 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera