Use of BRDF for slope estimation in machine vision

Abstract

Imaging cameras are used routinely in remote sensing to provide the first detailed look at the remote surface that cannot be examined by a human explorer. Images in the visible spectrum provide useful data for human analysis, because the human is most likely to interpret the images. Recently, there has been a renewed interest in human planetary exploration. As a precursor to establishing a human presence on other bodies in the solar system (Mars or the moon), a detailed survey of the landing site is planned from the orbit. Such a survey will include the classification of the objects on the preselected site by their shape and slope. There are two types of object that are difficult to characterize using the shade techniques. Both of these surfaces are of great interest to the landing vehicle or to the roving vehicle that will precede the human exploration. The first is a steep hole, and the second is the slope of the same angle as the angle of illumination. A model of the bidirectional reflectance function of the Martian surface is employed to devise imaging strategy for terrain surveying and classification from orbit. It is used in slope estimation algorithms to determine rock and hole shapes in conjunction with illumination parameters.