Study of Sub-pixel Level Edge Extraction for Graphics Under Different Environmental Conditions

Edge features are an important feature in our study of graphics, so accurate edge detection is of great importance for image recognition, processing of images and applications in areas such as video surveillance. While the recognition accuracy of our research images often only stays at the pixel level, there is still a lot of room for exploring the sub-pixel level edge extraction of images. In this paper, we compare the extraction principles of various mainstream operators and finally apply the Devernay algorithm based on the Canny operator. Using Gaussian filtering and Gaussian bilateral filtering for image pre-processing, combined with a reasonable threshold setting for the distance between the line segment set and the next detected line segment set and the double thresholding method, completes the sub-pixel level edge extraction for images in various environments, which improves the extraction clarity by more than 40%.