Detection of Features by Trinarization of Space Images

Abstract

It is proposed that the presence of a shadow of an object be used to ensure the required reliability of automated detection of observed features on space images. The detection procedure is performed using the results of an analysis of the first derivative of brightness change on the image. New concepts are introduced when examining this approach: trinarization and a ternary image. The concept of a uniform image background is formalized. New methods for automated processing of images of terrain features are proposed and are tested in practical work.