Abstract
This work addresses the generation of high quality digital surface models by fusing multiple depths maps calculated with the dense image matching method. The algorithm is adapted to very high resolution multi-view satellite images, and the main contributions of this work are in the multi-view fusion. The algorithm is insensitive to outliers, takes into account the matching quality indicators, handles non-correlated zones (e.g. occlusions), and is solved with a multi-directional dynamic programming approach. No geometric constraints (e.g. surface planarity) or auxiliary data in form of ground control points are required for its operation. Prior to the fusion procedures, the RPC geolocation parameters of all images are improved in a bundle block adjustment routine. The performance of the algorithm is evaluated on two VHR (Very High Resolution)-satellite image datasets (Pléiades, WorldView-3) revealing its good performance in reconstructing non-textured areas, repetitive patterns, and surface discontinuities.
Highlights
The modern high resolution satellites are capable of frequent revisit times thanks to their pointing agility
The implemented methods are tested on numerous multi-view satellite image configurations using two datasets – a rural zone captured in a single epoch and an urban zone captured at various epochs across the year
At this stage the algorithm has at the disposal several depth maps registered in some absolute reference frame (RF), with their respective normalized cross-correlation (NCC) maps and 2D occlusion masks
Summary
The modern high resolution satellites are capable of frequent revisit times thanks to their pointing agility. Depending on the adopted acquisition mode, a satellite can view an area covering up to 1000 Â 1000 km, all in a single pass.1 This necessitates large viewing angles (up to % 30 in the standard mode, 45 in the extended mode and 60 and more if desired) and manifests in both, less controlled base to height ratio (B=H), varying ground sampling distance (GSD), as well as possible multi-view acquisitions. E. Rupnik et al / ISPRS Journal of Photogrammetry and Remote Sensing 139 (2018) 201–211 while being computationally and memory-use efficient (Section 3.3.3); – no geometric constraints (e.g. surface planarity) are assumed and no auxiliary data (e.g. ground control points) are required. The implemented methods are tested on numerous multi-view satellite image configurations using two datasets – a rural zone captured in a single epoch and an urban zone captured at various epochs across the year.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have