Computer aided signature analysis using tone and texture

Abstract

The degree of utility of texture as a supplement to tone in signature analysis of selected agricultural crop types has been investigated using computer aided automated classification methods. Out of the many parameters suggested to quantify texture, the “standarddeviation of repeated measurements of tone in a single field” was selected as the only measure of texture for this study. Five crop types (uncut corn, cut corn, wheat fields, alfalfa and idle fields) were selected in an area west of Ann Arbor, Michigan, U.S.A. Thirteen replicates per crop type were sampled on a single. aerial photograph, flown in September. Photographic prints at the three scales of 1:14,700, 1:7350, 1:3675 from the same aerial negative were used for measuring tonal reflectance using a densitometer with a fixed aperture size of 1.5 mm. Mean tone and standard deviation of tonal measurements for each field were computed Statistical analyses were carried out to determine the role of tone and texture, individually and jointly. The values of mean tone and texture, for all the three scales of photography were used to determine the effect of scale on discrimination capability.

It was found that texture as defined, significantly improved identification and classification for the five selected known crops as they appeared on a single aerial photograph. It had both a complementary and a supplementary role to that of tone. The success of the discrimination depended both on the photographic scale and the relative size of the densitometer aperture.