ABSTRACT
Multiview image matching is an active topic in photogrammetry, with many achievements and applications; however, it is still abstruse due to its sophisticated mathematics, its complicated matching techniques, and the technical terminology used. Based on up-to-date literature, this chapter summarizes the fundamental techniques of multiview image matching and 3D surface reconstruction. The chapter defines multiview image matching using a series of equations, with steps from image orientation, matching candidate searching, and similarity calculation to 3D surface reconstruction. Advanced techniques, including matching parameters, matching primitives, global optimizations, and occlusion processing, are further discussed. Moreover, the broader issues related to matching are discussed; namely, the impact of sensor scanning modes, divergence in photogrammetry and computer vision, and matching efficiency, as well as the available multiview image matching software. Finally, multiview image matching accuracy and related challenges, limitations, and opportunities are discussed in order to suggest the research direction for this subfield.
