Semantic labelling of point cloud into various landuse landcover categories is one of the pivotal steps to utilize LiDAR point cloud for creating measurable 3D virtual models. The computationally complex, voluminous and unstructured LiDAR point cloud makes this a daunting task. As with any high resolution dataset, heterogeneous appearance of the same object makes it even more complex. Hence, automatic extraction of various objects from these complex datasets demands advanced 3D processing techniques and has remained an open research problem.

Major research goal was to develop an efficient and reliable algorithmic framework for semantically labelling 3D coloured LiDAR point cloud (point cloud with spectral data integrated) acquired over an urban environment using computer vision techniques in an open source prototype system. Some of the highlights of this research are: A novel 3D object-based framework for semantically labelling the 3D coloured LiDAR point cloud obtained by integrating LiDAR point cloud and multispectral imagery was developed. To improve the efficiency of the algorithm while processing highly dense point cloud, a computationally efficient supervoxels-based LCCP (Local Cloud Connectivity Patches) segmentation approach has been adapted and extended for creating meaningful segments from the point cloud. The study also critically assess the role of spectral and geometrical information in the various stages of object- based point cloud labelling, namely, segmentation, feature extraction, and classification.

Publications: 

Ramiya, Anandakumar M., Rama Rao Nidamanuri, and Ramakrishnan Krishnan. 2016. Supervoxels Based Spectro Spatial Approach for 3D Urban Point Cloud Labelling. International Journal of Remote Sensing. 37:17, 4172-4200, DOI:10.1080/01431161.2016.1211348. Publisher:Taylor and Francis.

Ramiya, Anandakumar M., Rama Rao Nidamanuri, and Ramakrishnan Krishnan. 2018. Critical assessment of semantic object based point cloud labelling on urban LiDAR dataset. GeoCarto International. Publisher: Taylor and Francis.

Ramiya, Anandakumar M., Rama Rao Nidamanuri, and Ramakrishnan Krishnan. 2016. Object-Oriented Semantic Labelling of Spectral - Spatial LiDAR Point Cloud for Urban Land Cover Classification and Buildings Detection. Geocarto International. 31 (2). doi:10.1080/10106049.2015.1034195.Publisher: Taylor and Francis.

Ramiya, Anandakumar M., Rama Rao Nidamanuri, and Ramakrishnan Krishnan. 2017. Segmentation based building detection approach from LiDAR point cloud. Egypt. J. Remote Sensing Space Sci.20(1), 71- 77.Publisher : Elsevier.

Ramiya, Anandakumar M., Rama Rao Nidamanuri, and Ramakrishnan Krishnan. 2014. Semantic Labeling of urban point cloud data. ISPRS - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XL-8, 2014, pp.907-911 .DOI : 10.5194/isprsarchives-XL-8-907-2014.