Integrated UAV photogrammetry and automatic feature extraction for cadastral mapping

The principal objective of this research is to investigate the applicability of UAV photogrammetry integrated with automatic feature extraction for cadastral mapping. We present here the concluding part of the paper. The first part was published in the October 22 issue

Oluibukun Gbenga Ajayi Department of Land and Spatial Sciences, Namibia University of Science and Technology, Namibia

Emmanuel Oruma Federal University of Technology, Minna, Nigeria

Results and discussion

Figure 2a presents the orthomosaic generated from the acquired images which serves as the base map for the extraction of land boundaries, building footprints and other features while Figure 2b is a screenshot of a zoomed portion of the image showing a sample of the markers used for the GCPs. The land plots and building lines from the orthomosaic were obtained by two different techniques; which are manual digitization and automated feature extraction technique using MRS, while the ground survey parcel data was used as reference. Also, the manually digitized property boundary features were used as reference information for the automatically extracted features. Figure 3 depicts the AutoCAD drawing obtained from ground survey approach which was used as reference data for the UAV survey data.

Other results obtained are the result of the accuracy evaluation of the automatically extracted visible boundaries using the MRS algorithm at different SPs (see Tab. 4 and Tab. 5), and the result of the cost and time comparison of the UAV and GNSS survey (see Tab. 6 and Tab. 7).

Scale Parameters of MRS

The results obtained when the SP of the MRS was set at 150, 400 and 500 are shown in Figure 4a – Figure 4c with the blue line depicting the boundary lines of every segment. It was observed that the output polygon continues to decrease with an increase in the value of the SP. Also, the visual clarity of the output polygon improves with increase in the value of the SP. When compared to the output of the automatically extracted visible boundaries using MRS with SP values 150 and 400 (Fig. 4a and Fig. 4b), the pixel level in the output map was observed to be decreasing continuously with increase in SPs as observed in the result obtained when the SP value was set at 500 (Fig. 4c).

Figure 5 presents the automatically extracted visible boundaries when the SP value of the MRS was set at 700, while Figure 6 presents the segmentation result obtained when the SP was fixed at 1000. Analysis of these two results showed that the pixel level decreased with increased value of SP. The output map of SP = 1000 gives a more cartographically appealing result based on its stronger pixel level. The yellow line in Figure 5 shows the boundary lines of every segment when SP = 700 was used while the red line in Figure 6 shows the boundary lines of every segment when SP = 1000 was used.

Accuracy assessment for the generated orthomosaic and automatic feature extraction

Table 2 presents the planimetric coordinates and discrepancy between the GNSS acquired coordinates and the extracted coordinates of the CPs from the UAV generated orthomosaic. From Table 2, ΔN (m) and ΔE (m) represents the difference in planimetric (northing and easting) coordinates as obtained from GNSS acquired data and UAV generated orthomosaic.

The obtained horizontal RMSE (RMSEx, y) as computed using equation (7) is 0.3575. This is consistent with the result obtained by Karabin et al. (2021) and it affirms the applicability of UAV in cadastral or property mapping.

The result of the estimated completeness, correctness and overall accuracy of the automatically extracted building footprints at different SPs is presented in Table 3, while the results obtained from the estimated completeness, correctness and overall accuracy of the automatically extracted land parcels at different SPs is presented in Table 4.

The result (see Tab. 3) shows that a completeness, correctness and overall accuracy of 16%, 12% and 14%, respectively, was obtained when the MRS algorithm was deployed for the automatic extraction of the building lines or footprints at a SP = 150. When the SP was set at 700, 89% and 91% completeness and correctness, respectively were obtained with overall accuracy of 86% while an overall accuracy of 88% was obtained when the SP was set at 1000 with 92% and 95% completeness and correctness, respectively. Meanwhile, a completeness, correctness and overall accuracy of 25%, 18% and 19% was obtained when the SP was prefixed at 150 for the automatic extraction of land parcels (see Tab. 4), while 65%, 59% and 54% were obtained for the completeness, correctness and overall accuracy, respectively, when the SP was set at 1000 for the automatic extraction of the land parcels using the MRS algorithm. The poor completeness, correctness and overall accuracy obtained from the automatically extracted land parcels when compared to the result of the building footprints can be attributed to the presence of shadows, unclear delineation of the boundary lines of the land parcels in vegetated areas, and the presence of mixed pixels in the automatic extraction (Horkaew et al., 2015; Wassie, 2016). The findings show that increase in the SP of the MRS algorithm also leads to increase in the obtained completeness, correctness and overall accuracy for the extraction of the building footprints and the land parcels. Also, it was observed that optimal completeness, correctness and overall accuracy of the automatic feature extraction was obtained when the SP is set at 1000, while setting the SP at 150 will not yield a reliable result. The result of the accuracy assessment is consistent with the findings of Luo et al. (2017), Munyati (2018) and Chen et al. (2019).

Cost and time comparison

The results obtained from the time and cost analysis of the integrated UAV-photogrammetry approach and the GNSS survey methods used for the survey of 248 land parcels are presented in Table 5 and Table 6, respectively.

From the results presented in Table 5 and Table 6, it can be observed that the parcel boundary extracted using GNSS method requires more intense field observation, thus, it consumes more time and cost. However, cadastral boundary extractions from UAV generated orthomosaic involves less field work and more off-field processing, and it is also more economical when compared to the GNSS method. Based on the time analysis, it was observed that the total time taken to map the 248 properties using the UAV photogrammetry approach was just about one-third (1/3) of the total time expended when GNSS method was adopted. While the project was executed within just 12 days using the UAV approach, it took a total of 30 days for the project to be completed using the conventional GNSS approach which shows that the integrated UAV approach is 2.5 times faster than the conventional GNSS approach, even when the same manpower was deployed for the project.

It was also observed that the cadastral boundary obtained using GNSS method requires more personnel, equipment and resources for detail field observation and data processing. However, less human effort with very few equipment is required for UAV data capturing and image processing, and also in vectorizing the UAV generated orthomosaic, which is also consistent with the findings of Karabin et al. (2021). The results obtained from the cost comparison of these two approaches as presented in Table 6 shows that a total amount of N1 190 000.00 was expended for the mapping of 248 land parcels at the cost of less than N5 000.00 per parcel when the UAV approach integrated with the automatic feature extraction was used, while an approximate cost of N11 160 000.00 was expended when GNSS approach was used to survey the same 248 land parcels at an average cost of N45 000.00 per parcel. This implies that for large scale property mapping, the presented UAV approach integrated with automatic feature extraction is approximately nine (9) times cheaper or less expensive than the classical GNSS surveying approach without compromising the obtainable accuracy.

Conclusions

The principal objective of this research is to investigate the applicability of UAV photogrammetry integrated with automatic feature extraction for cadastral mapping. MRS algorithm with different SP was implemented for the automatic extraction of visible cadastral boundaries defined by linear features such as defined nodes and building footprints. The result obtained from the automatic feature extraction shows that the accuracy of the cadastral boundary line extraction depends majorly on the SP which is the key control of the MRS algorithm. For the experiments conducted using varying SPs and constant shape and compactness value, the result obtained shows that the pixel level in the output map decreases continuously with increase in SPs while the optimal result of the conducted experiment was obtained when the SP was set at 1000, while the shape and compactness values were set at 1.5 m and 0.8 m, respectively. The result of the evaluation of the reliability of the automatic extraction also shows that the completeness, correctness and overall accuracy or quality increases with increase in the value of the SPs. Also, the MRS algorithms proved to be more efficient in automatically extracting building footprints when compared to its performance in the extraction of land parcels.

Furthermore, the results of the accuracy assessment obtained from the integrated UAV approach when compared with conventional survey approach shows that 99% of automatically extracted property boundaries from the UAV survey falls within the minimum acceptable horizontal accuracy for cadastral and property mapping of third order (1: 5,000). Further analysis on the cost and time expended for the property mapping using the integrated approach shows that the approach is approximately 2.5 times faster and 9 times cheaper than the conventional ground surveying approach, especially when GNSS receivers are used for the spatial data acquisition. While MRS algorithm has proved to be a veritable model for automatic extraction of building footprints in cadastral mapping in this study, further research efforts shall seek to investigate the applicability of other segmentation algorithms in the automatic extraction of parcel boundaries for cadastral applications. Meanwhile, it should be noted that the automatic extraction of boundaries is only a step to the facilitation cadastral mapping, as mere detecting and extracting the boundaries alone is not sufficient for complete and correct cadastral mapping.

Author contributions

O.G.A: Conceptualization, research concept and design, article writing, critical revision of the article and final approval of the article. O. E: Collection and assembly of data, data analysis and interpretation, and writing of the article draft.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Acknowledgements

The authors would like to thank the anonymous reviewers and the editor for their constructive comments and contributions.

References

[1] Ajayi, O.G., Palmer, M., and Salubi, A.A. (2018). Modelling farmland topography for suitable site selection of dam construction using unmanned aerial vehicle (UAV) photogrammetry. Remote Sens. Appl: Soc. Environ., 11, 220-230. DOI: 10.1016/j.rsase.2018.07.007.

[2] Ali, Z., and Ahmed, S. (2013). Extracting parcel boundaries from satellite imagery for a Land Information System. In 6th International Conference on Recent Advances in Space Technologies (RAST), 12-14 June. pp.79-81. Istanbul, Turkey.

[3] Aplin P., Atkinson P., Curran P., (1999). Per-field classification of land use using the forthcoming very fine resolution satellite sensors: problems and potential solutions. In: Advances in Remote Sensing and GIS Analysis. Chichester: Wiley and Sons Inc., 219-239

[4] Aslam, A., Khan, E., and SufyanBeg, M.M. (2015). Improved edge detection algorithm for brain tumor segmentation. Procedia-Procedia Computer Science, 58, 430-437. DOI: 10.1016/J.PROCS.2015.08.057.

[5] Baatz, M., and Schape, A. (2000). Multiresolution segmentation: an optimization approach for high quality multi-scale image segmentation. Retrieved February, 2022, from http://www.ecognition.com/sites/ default/files/405baatzfp12.pdf.

[6] Babawuro, U., and Beiji, Z. (2012). Satellite imagery cadastral features extractions using image processing algorithms: A variable option for cadastral sciences. Int. J. Comput. Sci., 9(4), 30-34.

[7] Bandyopadyay, S.K. (2012). Edge detection from CT images of lung. Int. J. Eng. Sci. Adv. Tech., 2(1), 34-37.

[8] Barnes, T., and Engel, A. (2014). The photogrammetric potential of lowcost UAVs in forestry and agriculture. International Archives of the Photogrammetry, Remote Sensing and Spatial Information Science, XXXVI pp. 16-19. Retrieved February, 2022, from http://www.isprs.org/proceedings/ XXXVII/congress/1pdf/206.pdf.

[9] Bowyer, K.K., Kranenburg, C., and Dougherty, S. (2001). Edge detector evaluation using empirical ROC curves. Comput. Vis. Image Underst., 84(1), 77- 103. DOI: 10.1109/CVPR.1999.786963.

[10] Chai, H., Wee, L. K., & Supriyanto, E. (2012). Edge detection in ultrasound images using speckle reducing anisotropic diffusion in canny edge detector framework. WSEAS Trans. on Biology and Biomedicine, 8(2), 51–60. Retrieved from http://www. mendeley.com/research/edgedetectionultrasound-images-using-specklereducing-anisotropicdiffusioncanny-edge-detector-fr/

[11] Chen, Y., Qiang, C. and Changfeng, J. (2019). Multi-resolution segmentation parameters optimization and evaluation for VHR remote sensing image based on meanNSQI and discrepancy measure. J. Spat. Sci., 66(2), 253-278. DOI: 10.1080/14498596.2019.1615011.

[12] CMU (1997). Performance evaluation for feature extraction. Slides presented at the Terrain Week 1997. Retrieved from February 2022, http://www.cs.cmu. edu/afs/cs/usr/maps/www/rcvw/terrain week97/roads/tw97-roadeval.ROOT.html.

[13] Crommelinck, S., Bennett, R., Gerke, M. et al. (2016). Review of automatic feature extraction from high-resolution optical sensor data for UAV-Based cadastral mapping. Remote Sens., 8, 689. DOI: 10.3390/rs8080689.

[14] Crommelinck, S., Bennett, R., Gerke, M. et al. (2017). Contour detection for UAV-based cadastral mapping. Remote Sens., 9, 171. DOI: 10.3390/rs9020171.

[15] Crommelinck, S., Höfle, B., Koeva, M.N. et al. (2020). Interactive cadastral boundary delineation from UAV data. ISPRS Ann. Photogr. Remote Sens. Spat. Inf. Sci., IV-2, 81-88. DOI: 10.5194/isprs-annals-IV-2-81-2018.

[16] Drǎguţ¸ L., Tiedec, D., and Levick, S.R. (2010). ESP: a tool to estimate scale parameter for multiresolution image segmentation of remotely sensed data. Int. J. Geogr. Inf. Sci., 24(6), 859-871. DOI: 10.1080/13658810903174803.

[17] Drǎguţ, L., Csillik, O., Eisank, C. et al. (2014). Automated parameterization for multi -scale image segmentation on multiple layers. J. Photogr. Remote Sens., 88, 119-127. DOI: 10.1016/j.isprsjprs.2013.11.018.

[18] Eisenbeiss, H. (2009). UAV Photogrammetry. Retrieved from February 2022, http://www.igpdata.ethz.ch/ berichte/blaue_berichte_pdf/105.pdf.

[19] Fetai, B., Oštir, K., Fras, M., Lisec, A. (2019). Extraction of visible boundaries for cadastral mapping based on UAV imagery. Remote Sens., 11(13), 1510. DOI: 10.3390/rs11131510.

[20] Fetai, B., Račič, M., Tekavec, J. et al. (2020). Detection of visible boundaries from UAV images using U-Net. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, XLIII-B1-2020, 437-442.

[21] François, C. et al. (2015). Keras. Retrieved from March 9, 2021, https://github.com/fchollet/keras.

[22] Galarreta, J., Kerle, N., and Gerke, M. (2015). UAV-based urban structural damage assessment using objectbased image analysis and semantic reasoning. Nat. Hazards and Earth Syst. Sci., 15, 1087-1101. DOI: 10.5194/nhess-15-1087-2015.

[23] Giuliani, D. (2012). Edge detection from MRI, DTI images with an anisotropic vector field flow using a divergence map. Algorithms, 5(4), 636-653. DOI: 10.3390/a5040636.

[24] Gupta, R. (2016). Enhanced edge detection technique for satellite images. International Conference on Cloud Computing and Security. China: Springer, pp. 273-283.

[25] Heipke, C., Mayer, H., Wiedemann, C.O. et al. (1997). Evaluation of automatic road extraction. Int. Arch. Photogramm. Remote Sens., 47-56.

[26] Horkaew, P., Puttinaovarat, S., and Khaimook, K. (2015). River boundary delineation from remote sensed imagery based on SVM and relaxation labelling of water index and DSM. J. Theor. Appl. Inf. Technol., 71(3), 376-386.

[27] Jazayeri, I., Rajabifard, A., and Kalantari, M. (2014). A geometric and semantic evaluation of 3D data sourcing methods for land and property information. Land Use Policy, 36, 219-230. DOI: 10.1016/j.landusepol.2013.08.004.

[28] Jena, K.K., Mishra, S., and Mishra, S. (2015). Edge detection of satellite images: A comparative study. Int. J. Innov. Sci., Eng. Tech., 2(3), 75-79.

[29] Karabin, M., Bakula, K., and Luczynski, R. (2021). Verification of the geometrical representation of buildings in cadastre using UAV photogrammetry. Geomat. Environ. Eng., 15(4), 81-99. DOI: 10.7494/geom.2021.15.4.81.

[30] Kohli, D., Crommelinck, S., and Bennett, R. (2017). Object-based image analysis for cadastral mapping using satellite images. In Proceedings of the International Society for Optics and Photonics, Image Signal Processing Remote Sensing XXIII; SPIE. The International Society for Optical Engineering: Warsaw, Poland.

[31] Kohli, D., Unger, E.-M., Lemmen, C. et al. (2018). Validation of a cadastral map created using satellite imagery and automated feature extraction techniques: A case of Nepal. In Proceedings of the FIG Congress, 6-11 May, pp. 17, Istanbul, Turkey.

[32] Lakhani, K., Minocha, B., and Gugnani, N. (2016). Analyzing edge detection techniques for feature extraction in dental radiographs. Perspectives in Science, 8(4), 395-398. DOI: 10.1016/j.pisc.2016.04.087.

[33] Luo, X., Bennett, R.M., Koeva, M. et al. (2017). Investigating semi-automated cadastral boundary extraction from airborne laser scanned data. Land, 6(3), 60. DOI: 10.3390/land6030060.

[34] Manyoky, M., Theiler, P., Steudler, D. et al. (2011). Unmanned aerial vehicle in cadastral applications. International Archives of the Photogrammetry, Remote Sensing and Spatial Information Science, XXXVIII-1/(September), pp. 57-62. DOI: 10.5194/isprsarchivesXXXVIII-1-C2257-2011.

[35] Martha, T.R., Kerle, N., van Westen et al. (2011). Segment optimization and data-driven thresholding for knowledgebased landslide detection by object-based image analysis. IEEE Trans. Geosci. Remote Sens., 49(12), 4928-4943. DOI: 10.1109/TGRS.2011.2151866.

[36] McGlone, J.C., and Shufelt, J.A. (1994). Projective and object space geometry for monocular building extraction. IEEE Conf. Comput. Vis. Pattern Recogn., 54-61

[37] Munyati, C. (2018). Optimizing multiresolution segmentation: delineating savannah vegetation boundaries in the Kruger National Park, South Africa, using Sentinel 2 MSI imagery. International J. Remote Sens., 39, 997-6019. DOI: 10.1080/01431161.2018.1508922.

[38] Neubert, M., Herold, H., and Meinel, G., (2007). Assessing image segmentation quality – concepts, methods and application. Lecture Notes in Geoinformation & Cartography, 769-784. DOI: 10.1007/978-3-540-77058-9_42.

[39] Nyandwi, E. (2018). The battle of cadastral intelligence: measuring the results of competition between people and machine in creation of cadastral boundaries. MSc thesis submitted to Enschede: University of Twente, Faculty of GeoInformation and Earth Observation (ITC). Retrieved from https://webapps. itc.utwente.nl/librarywww/ papers_2018/msc/la/nyandwi.pdf.

[40] Pal, N.R., and Pal, S.K. (1993). A review on image segmentation techniques. Pattern Recognition, 26(9), 1277-1294. DOI: 10.1016/0031-3203(93)90135-J.

[41] Parida, P.K., Sanabada, M.K., and Tripathi, S. (2014). Cadastral resurvey using high resolution satellite ortho image-challenges: A case study in Odisha, India. Int. Arch. Photogram., Remote Sens. Spat. Inf. Sci., 40(8), 1165-1170. DOI: 10.5194/ isprsarchives-XL-8-1165-2014.

[42] Punarselvam, E., and Suresh, P. (2011). Edge detection of CT scan spine disc image using canny edge detection algorithm based on magnitude and edge length. In 3rd International Conference on Trendz in Information Sciences & Computing (TISC2011). Chennai, India, pp.136-140

[43] Puniach, E., Bieda, A., Cwiakala, P., et al. (2018). Use of Unmanned Aerial Vehicles (UAVs) for Updating Farmland Cadastral Data in Areas Subject to Landslides. ISPRS Int. J. Geoinf., 7(8):331. DOI: 10.3390/ijgi7080331.

[44] Qui, K., Sun, K., Ding, K. et al. (2016). A fast and robust algorithm for road edges extraction from Lidar data. Int. Arch. Photogram., Remote Sens. Spat. Inf. Sci., 693-698. DOI: 10.5194/ isprs-archives-XLI-B5-693-2016.

[45] Rijsdijk, M., van Hinsbergh, W.H.M., Witteveen, W. et al. (2013). Unmanned Aerial Systems in the process of Juridical verification of Cadastral border. ISPRS Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci., XL-1/W2, 325-331. DOI: 10.5194/ isprsarchives-XL-1-W2-325-2013.

[46] Selvakumar, P., and Hariganesh, S. (2016). The performance analysis of edge detection algorithms for image processing. International Conference on Computing Technologies and Intelligent Data Engineering, 7-9 January Kovilpatti, India, pp. 1-5

[47] Sirmacek, B., and Unsalan, C. (2010). Road network extraction using edge detection and spatial voting. International Conference on Pattern Recognition, pp. 3117-3120. Istanbul, Turkey.

[48] Taha, L.G.E., and Ibrahim, R.E. (2021). Assessment of Approaches for the Extraction of Building Footprints from Pléiades Images. Geomat. Environ. Eng., 15(4), 101-116. DOI: 10.7494/geom.2021.15.4.101

[49] Wassie, Y.A. (2016). Toward automatic cadastral boundary mapping from satellite imagery. Thesis submitted to the Faculty of Geo-Information Science and Earth Observation of the University of Twente in partial fulfilment of the requirements for the degree of Master of Science in Geoinformation Science and Earth Observation.

[50] Wassie, Y.A., Koeva, M.N., Bennett, R.M. et al. (2018). A procedure for semiautomated cadastral boundary feature extraction from high-resolution satellite imagery. J. Spat. Sci., 63(1), 75-92. DOI: 10.1080/14498596.2017.1345667.

[51] Witharana, C., and Civco, D.L. (2014). Optimizing multi-resolution segmentation scale using empirical methods: exploring the sensitivity of the supervised discrepancy measure Euclidean distance 2 (ED2). ISPRS J. Photogram. Remote Sens., 87(1), 108-121. DOI: 10.1016/j. isprsjprs.2013.11.006.

[52] Zevenbergen, J., and Bennett, R. (2015). The visible boundary: More than just a line between coordinates. In Proceedings of the GeoTech Rwanda, International Conference on Geospatial Technologies for Sustainable Urban and Rural Development, Kigali, Rwanda, 18-20 November, pp. 1-4.

[53] Zhang, X., Xiao, P., Song, X. et al. (2013). Boundary-constrained multiscale segmentation method for remote images. ISPRS J. Photogram. Remote Sens., 78, 15-25. DOI: 10.1016/j.isprsjprs.2013.01.002

The paper was first published in Advances in Geodesy and Geoinformation ,Vol. 71, no. 1, article no. e19, 2022 and is republished with the authors’ permission. Copyright The Author(s). 2022 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/ by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.