Research On Wear Detection Technique Of Rollre Based On Machine Vision

Roller is a crucial part of rolling mill, the wearing condition of roller can directly influence the quality of production and the service life of the roller. In recent years, the machine vision technology has been used in wear detection widely. It has improved the traditional detection method, enhanced the work efficiency immensely and solved many difficult problems of wear detection. Therefore, how to make better use of machine vision technology to detect the surface wear of roller is a topic which is worth researching.In this paper, roller wearing is the object of machine vision detection. The main purpose of the study is to provide advanced detection means and equipment support for the non-contact detection of roller wear. Therefore, the main content of this paper is listed as below:1. This paper designs a non-contact detection system for roller wear based the structure and principle of machine vision detection system. The structure and parameter of some crucial parts, such as light source, optical lens, CCD camera and image gathering card, are discussed in detail, as well as the types and technical parameters of the productions selected.2. The commonly used image processing techniques in machine vision detection are summarized in this paper, such as histogram processing of gray image, image smoothing and enhancement, image segmentation, image edge detection and connectivity. It also gives the comparison and analysis of some examples.3. The experiment of the system is carried out by the image processing function of MATLAB7.0. During the experiment, auto-adaptive median filter algorithm is used for pre-processing of image, sub-pixel positioning of edge points is accepted to achieve positioning and processing of the edge of image after the edge extraction of image using Sobel operator, what's more, the experiment results and the sources of error are analyzed.Based on the characteristics of roller, all aspects of the detection of roller wear using machine vision are analyzed and discussed separately, with the result that, it supplies the wearing detection using machine vision, especially using image processing technology, with certain theoretical basis.