Modification of the Canny Edge Detector for Recognition of Image Markers of Architectural Plans

The article describes an algorithm used for recognition of image markers of design and estimate documents of the developer's architectural plans using computer vision technology. The main goal of the study is to propose a pre-processing method capable of improving and revealing the features of the given images and solving the problem of noise reduction. The article describes the analysis of the Canny operator and considers its modification, where the Chambollele algorithm-based Rudin-Osher-Fatemi model is used for image pre-processing. The method described is effective in detection of edges on images of architectural plans. An approbation of the developed software modules and a comparative analysis of the results of the algorithms were carried out.