Analysis of Image Edge Checking Algorithms for the Estimation of Pear Size

The lower-layer image processing algorithm is the basis of pear on-line automatic grading with machine vision. The edge detection algorithms and edge thinning algorithms of the lower-layer image processing were analyzed and the lower-layer processing algorithm suitable to fruit grading was obtained in the work. The edge detection technique is one of the technologies to display the image contour distinctly with edge enhancement. The edge thinning is to process the contour lines in different sizes to lines with the width of only one pixel. Several classical difference operators for edge detection, such as Robert edge operator, Sobel edge operator, Prewitt edge operator and Kirsch edge operator were introduced. The edge detected with the classical operators was very thick, and even with complex form or not smooth, so the template detection algorithm was put forward to measure the pear´s edge and its steps were shown detailedly. The Hilditch thinning method was adopted to deal with the edge image in the work and the image got with the edge thinning were fine.