Elevation Measurements

This chapter presents different elevation reconstruction approaches using satellite images. Stereoscopic vision is based on the difference perceived between two images taken from two different points of view. The epipolar geometry is a property verified by the image couples acquired with a matrix sensor, which ensures that the exposure parameters, whether they are intrinsic or extrinsic, are the same for all the points of an image. The chapter explains tools that allow one to obtain a disparity map, which preserves the correspondences found in the two images, and a cloud of 3D points, which are the intersections of light rays coming from homologous pixels. A strong constraint connected to the fusion of data is the requirement to have these data in the same referential. The chapter explains what consequences this constraint has on the methods that can be implemented. It illustrates radargrammetry in the simple case of parallel trajectories.