Saturation High Light Removal of Pressure Plate Image Based on Feature Matching
Peng Li, Yongxin Chen, Hengxuan Li, Jingyan Yao, Shu Zhengyu, Di Wu · 2021 6th Asia Conference on Power and Electrical Engineering (ACPEE) · 2021
When the artificial intelligence inspection robot collects the image of the relay protection pressure plate, the texture of the pressure plate image will be missed due to the reflection of the glass cabinet door. In order to solve this problem, a saturation high light removal method of the pressure plate image based on feature matching is adopted. The pressure plate image with saturation high light should be analyzed first, and two photos with different positions of saturation high light are taken from different angles of the same scene at the same time. Considering the different shooting angles, the SIFT algorithm is used to match the features and perspective transformation is used to make the angle of view consistent. Then the Canny operator is used in the edge detection of the saturation high light region, and the corresponding position of the auxiliary image can repair the texture of the reference image. The experimental verification example is proposed to prove the rationality and effectiveness of this method.