APPLICATION OF COMPUTER VISION ALGORITHMS FOR EVALUATING THE PERFORMANCE OF COTTON PICKERS

Abstract

Collecting data on cotton field's agricultural background and quality of the cotton pickers' work is one of the most important works in preparation for the cotton picking by machine. The use of advanced information technologies for the collection and processing of these data allows to create automated software tools that significantly reduce the time and cost. Creating the algorithms that can be applied in the development of such programs is one of the urgent problems of computer vision and agricultural sciences. Algorithms for determining the number of green bolls based on the analysis of images of a section of cotton field from different angles are proposed and experimentally tested. Keywords: image analysis, Canny's edge detection method, Hough transform, decision trees, algorithm C4.5.