An adaptive edge detection method for extracting images of defect steel plates under complex background

Aiming at the problem that the traditional single algorithm is poor for image edge detection and target region extraction in complex backgrounds, an improved Canny operator is proposed to realize defective steel plate image edge detection and target region extraction. The bilateral filtering and the improved Otsu method are introduced to realize image denoising and double threshold adaptive adjustment of Canny detector respectively, and the feature index with prior knowledge is constructed to filter the image background and extract the target region. The experimental results show that the proposed method outperforms the traditional edge detection method in terms of accuracy and robustness.