Edge detection of printed matter based on improved Canny operator

In order to solve the problem of irregular edge and more noise in printed products, this paper proposes an improved method of Canny operator to detect edges. Firstly, the filter is changed from gaussian filter of traditional Canny operator to bilateral filter, so as to improve the noise reduction of gaussian filtering, there may be a problem of unclear noise detection; secondly, the gradient amplitude and direction of the printed product image are calculated based on the 3 × 3 gradient template of sobel operator, and the final edge detection image is obtained. The final experiment also verifies that the image edge obtained by the improved algorithm is relatively complete, and the effect is slightly improved compared with the traditional Canny operator.