An Edge Extraction Method Based on Gray Lever-Gradient Two-Dimensional Maximum Entropy Threshold Method for Footprint Inspection

摘要

The extraction of the edge of the footprint image is the basis of measuring various features, which is used to identify various shapes and analyze various features. In this paper, a gray-scale-gradient two-dimensional threshold method is applied to extract the footprint profile. Firstly, the gray-gradient co-occurrence matrix model is established, and the two-dimensional threshold of gray level and gradient is obtained with the principle of maximum entropy. Then the footprint image is binarized by the two-dimensional threshold vector to accurately segment the blurred edge pixels of the image and extract the edges. Experimental results show that this edge extraction method presented has strong anti-noise ability, which can deal with the edge fuzzy pixels accurately, improve the quality of image segmentation, and further improve the precision of edge extraction.