"图像处理中Hough变换原理要点及预处理方法研究"

Hough transform is an effective method for extracting target shape features and has been widely used in image processing and pattern recognition. However, the Hough transform algorithm is mainly applied to binary images (i.e., edge images), so preprocessing of grayscale images (including image filtering and edge detection) is necessary before applying the Hough transform. Image preprocessing is an important preliminary work in the target detection process of the Hough transform, and its results directly affect the quality of the detection results. This study introduces two common types of noise in images and conducts research on spatial domain noise filtering of images. The limitations of Gaussian filtering and median filtering are analyzed, and based on bilateral filtering, a bilateral filtering method based on multiple median extractions is proposed. This method takes into account the spatial neighborhood correlation and pixel intensity similarity, selects the reference pixel value of the intensity filter through pseudomedian filtering, and better preserves edges while smoothing Gaussian noise and salt-and-pepper noise. Compared with traditional filtering methods designed for specific types of noise, this method can handle images with mixed noise (Gaussian noise and salt-and-pepper noise) and achieve a good filtering effect. Furthermore, this study analyzes several classic edge detection operators, replaces the Gaussian filtering process or adaptive filtering process in the Canny operator with an iterative bilateral filtering method, and to some extent avoids the blurriness caused by the filtering process to edges, resulting in better edges.