Role of Ground Control Points (GCPs) in Integration of Terrestrial Laser Scanner (TLS) and Close-range Photogrammetry (CRP)

The need for GCPs is increasingly more important with the increase in higher accuracy requirements and increase in user expectations. GCPs (Ground control points) are necessary for orientation and placement of photographs and 3D models in the spatial coordinate system, and they play a key role in co-registration of two point clouds. This paper deals with the assessment of the role of Ground Control Points in Co-registration of CRP and TLS point clouds by point-pair selection methodology and Automatic co-registration algorithm. In this work, the point cloud is generated from multiple overlapping sequences of images using Close-range Photogrammetry (CRP) and Terrestrial Laser Scanning (TLS) for a building over the planar surface. GCPs were collected by the total station to register the TLS and CRP point cloud. Overlapping photographs are processed in Agisoft PhotoScan software. TLS point cloud was generated from Riegl VZ 400 and GCPs were used to geo-reference it in Cloud Compare software. Various subsets of both point clouds are co-registered by the point-pair co-registration method and by Automatic point detection fine co-registration. Two subsets for each of CRP and TLS point cloud are considered in such a way that one is having some common overlap and other is having no common feature. GCPs registered point clouds integrate precisely as compared to that of the non-registered point cloud. The RMS error achieved in case of geo-referenced point cloud co-registration was 0.0091645433 m and in non-geo-referenced co-registration was found to be 0.03327466 m. In this study, it was also found that error is significantly higher in Automatic point detection method as compared to that of conventional point-pair selection co-registration. From this study, it is observed that higher accuracy of co-registration is achieved in the case of geo-referenced models by point-pair selection method. So, GCPs are the prerequisite for the effective and precise co-registration of 3D point clouds.