Corner Detection using Image Processing
Corner detection is an important step in the image processing of machine vision.
An improved algorithm is proposed in this project following the analysis on the existing corner detection algorithms and on the localization precision and computation efficiency in the Harris corner detection algorithm.
In this algorithm, a large number of irrelevant points are rejected by statistical analyzing the pixel gray level difference around the target pixel, and then the response function of residual points is calculated and compared with the set threshold value to certify the real corner.
Finally computation program is programmed, using this program, the synthesize images of five types of corners are analyzed and calculated, which shows that the improved algorithm acquires better efficiency and accuracy in corner detection.