Experience in automated interpretation of aerospace photographs
Abstract
A software method for use in automated interpretation of aerospace video information, in particular, from a multizonal aerospace survey is described. The Aerospace Methods Laboratory at Moscow State University developed two algorithms for solving this problem in two stages. The first algorithm is for the preliminary classification of multizonal information and evaluating the uniformity of multizonal data. This algorithm for the rapid discrimination of classes (RDC) is used in the breakdown of initial data into a relatively small number of groups, after which each group is regarded as a single object. Use of this algorithm does not require a priori information on these objects other than a multizonal photograph. The second algorithm is a successive clusterization iteration algorithm (SCIA). In contrast to the RDC algorithm it requires a definite volume of a priori information on the considered objects. It is necessary to stipulate the standard brightness characteristics of these objects. The algorithms is used in identifying the classes formed by the RDC algorithm with standard objects (called clusters), in each iteration evaluating the results on the basis of experimentally determined criteria. The results of automated interpretation of agricultural crops are given as an illustration of the practical use of the method.
 Publication:

USSR Report Space
 Pub Date:
 February 1985
 Bibcode:
 1985RpSpR.......75L
 Keywords:

 Image Analysis;
 Photointerpretation;
 Spaceborne Photography;
 Algorithms;
 Brightness;
 Classifications;
 Computer Techniques;
 Farm Crops;
 Remote Sensing;
 Instrumentation and Photography