Application of Image Weight Models to Increase Canny Contour Detector Resilience to Interference

Article discusses the issues of creating models and methods for detecting contours in images in control systems and automation of technological processes and objects based on computer vision technologies. An approach based on the construction and analysis of image weight models is proposed. This approach involves performing a discrete wavelet transform of the image and calculating weight values based on the detailing coefficients to assess the significance of changes in the pixels of copies of various levels of decomposition. The set of weight values is a weight model. Since significant changes are characteristic of contour points, the analysis of weight models makes it possible to detect them in images. It is also shown that the application of the Canny detector to the image weight model reduces the sensitivity to noise and the complexity of selecting the lower and upper thresholds.