Grading system of pear's appearance quality based on computer vision

Abstract

In order to realize the automatic grading system of pear's appearance quality based on computer vision, a grading system is constructed which includes hardware and software environment. It includes transport unit, control module, image acquisition module, and image processing and recognition module. It can realize automatic feature extraction. The features include shape, color and defect. According to the influence of spots on pear's surface on the defect detection, a spot removal method based on V component's dynamic threshold is put forward. According to the national standards, grading rules of fruit type and defect levels based on above features are constructed. Furthermore grading model of pear based on artificial neural network is also constructed. The recognition rate of 630 images of pear reaches up to 90.3%. The system, equipment and method in this paper have positive significance to on-line grading of pear's quality of appearance.