Analysis of Image Processing Algorithms in Technical Vision Systems Based on a Cropped Video Sequence

The paper provides an analysis of classical image processing methods, such as filtering methods, contour selection methods, and segmentation methods. Linear and median filtering, as well as adaptive filtering, are described. Among the many contour extraction methods, the methods of Sobel, Pruitt, Roberts, Laplacian, Gaussian, and Canny are considered. Two approaches to image segmentation are distinguished and such methods as the global threshold processing method, the Bayesian method, and the centroid binding method are considered. The main advantages and disadvantages of the considered methods are highlighted, depending on different initial conditions.