TABLE OF CONTENTS Radarclinometry is the second technique used to build an elevation chart based on SAR imaging [TOU 00]. In contrast to the other two methods [1], radargrammetry (see Chapter 10) and interferometry (see Chapter 12), radarclinometry uses a single image. It is an easy method to implement, and working conditions, although restrictive, lead to reduced precision results. Radarclinometry can be considered a qualitative relief reconstruction method, insofar as the reconstructed reliefs are globally and morphologically similar to the real reliefs, but with elevations marred by large errors.
Radarclinometry is based on the shape from shading principle; this is a monoscopic image reconstruction method largely studied in computer vision domain [HOR 89]. It was first proposed for radar by Wildey, who applied it successfully to Venus landscapes captured by the Magellan space probe [WIL 86b, WIL 86a]. However, it has not been possible to validate these results by cartography in the field. While few researchers have been interested in this method since, we will mainly present the works of Frankot and Chellappa [FRA 87, FRA 90] and those of Paquerault [PAQ 97, PAQ 98] (see also [GUI 90, THO 91]).
[1]
[1]In section 11.5 a fourth method will be mentioned: radarpolarimetry, but this method is not well developed.
[1]Chapter written by Henri MA TRE
The problems of shape from shading [HOR 89] are expressed by an equation with partial derivatives connecting reflectance R of an object to lighting function
