An Edge Detection Algorithm of Anti-Harmonic Image Based on Tensor Form

Edge detection of image is used to improve the visual recognition ability of the pattern by identifying points with obvious variation of brightness in a digital image. In the case of uneven conversion of color information of antiharmonic image pixel, edge detection becomes difficult. Traditional edge detection algorithm of anti-harmonic image adopts tensor model, as due to relatively simple structure of the tensor vector and operation, the performance of edge detection is poor. Therefore, An edge detection algorithm of anti-harmonic image based on colored tensor form is proposed. It starts from building the tensor model of colored and anti-harmonic image, accurately measuring the relationship between pixels by means of rich computing of tensor, finding the maximum and minimum values of the first derivative of the image to detect the boundary in the space of tensor form, and finally improve the algorithm based on the color tensor morphological operators of referenced total order. Simulation results show that the anti-harmonic mean of the proposed algorithm is higher than the conventional algorithms because it not only takes into account the correlation between the color components, but also considers the conversion feature of color information, which has superior performance of detection.