An Improved Silicon Steel Detection And Location Algorithm Research And Implementation

In this thesis, taking edge detection and location for silicon steels as research background, we carefully studied and improved the Hough transform algorithm, which is a well-known edge detection algorithm in image processing. The main contents of this thesis include:edge detection, straight edge location, Implementations and improvements on edge detection algorithms about for round edge, arc edge, oval edge respectively. Based on above mentioned algorithms, we can locate the silicon steel edges quickly and accurately. The main research contents and achievements of this thesis are as follows:For the edge detection of the silicon steels, we applied the traditional edge detection algorithms to detect the silicon steel edge. We conclude that the detection result of Canny edge detection algorithm is the most ideal result on the experimental results. This algorithm is suitable for the edge detection of the silicon steels.For the edge detection and location on both sides of the silicon steels, based on the noise-robustness characteristics of the Hough transform algorithm, we applied the traditional Hough transform algorithm to detect and locate the silicon steel edges. This algorithm can accurately determine the line equation of straight edge of the silicon steels.For the detection and location of upper and lower arc edges of silicon steels, we first detect them according to the circle detection principles of traditional and random Hough transform. To overcome the problems these two algorithms exist, we proposed an improved Hough transform algorithm to detect circular arc. Furthermore, based on the geometric features of silicon steels, we applied our proposed improved Hough transform algorithm to detect and locate upper and lower arc edges of silicon steels. From the comparison of the three algorithms, we conclude that the proposed algorithms in this thesis have better performance. In detail, the proposed algorithm can ensure the location accuracy as traditional Hough transform algorithm meanwhile can promote the processing speed. In addition, to overcome the time-consuming and memory consumption drawback of the Hough transform algorithm when extracting ellipse edges, we propose an improved elliptical edge detection algorithm. Experiments validate the effectiveness and processing speed of the proposed algorithm.In this thesis, we designed and implemented an software system to detect and locate silicon steel edges. The system can extract straight edge and the upper and the lower circular edges accurately. It also can accurately give the equations of the circle and the straight line on which the circle edge and the straight edge located respectively. The software is useful to compare the performance of location algorithms.