Preliminary radar feature extraction and recognition using texture measurement

Preliminary results are presented for the automated extraction and classification of a selected set of radar imagery containing city, field, water, and forest images. A sensing array minicomputer system with image texture processing algorithm was employed for the scanning and conversion of radar images into digital signals and also for the extraction of a feature vector from the images. A sequential template-matching classifier was used for the classification of the set of radar imagery into preassigned categories. Slightly better than 90 percent classification accuracy was obtained.