Tungsten Ore Primaries Algorithm Based On Machine Vision Research

With the machine vision inspection widly used in industry, its intelligence provides a powerful technical support for the industrial production of information. In light of China's tungsten mining process, ore beneficiation used the traditional manual method in the primary. Its drawbacks are obvious. The tungsten ore in the ore texture and unique color are characteristics of the distribution, for the machine vision inspection provides an objective basis. Machine vision technology has powerful advantages: fast, informative, multi-function, high efficiency. If those advantages were applied to primary of tungsten ore, manual method would have unmatched advantages.Using machine vision detected not only to exclude subjective factors, but also to quantitative description of these indicators. The results avoided and reduce the primary ore errors, improved production efficiency and sorting accuracy.Therefore, for the problem of mechanization in tungsten primary sorting, this paper demonstrated the machine vision technology to achieve viability in the tungsten primary detection system based on the software and hardware. Focusing on system sorting process of the core algorithm -- target mine edge detection, this paper launched a concrete analysis and research. This paper mainly some work to do the following:Firstly, according to the characteristics of the machine vision inspection, it studied the general structure and system functions of machine vision inspection system. At the same time the paper give the machine vision technology for tungsten ore primary process.Secondly, this paper reposed on specific hardware and software algorithms solutions of machine vision sorting technology, which is made in the study of tungsten ore on the basis of physical image characteristics and combining with tungsten primary process.And it summed concrete realization of the system about machine vision inspection technology in the primary tungsten.Moreover, the paper focus analysis and research of the system sorting process of the core algorithm - tungsten ore image edge detection. First, it compared efficiency of Roberts operator, Sobel operator, Prewitt operator, Laplacian-Gaussian operator, Canny operator, in the tungsten mine of the image feature extraction, operator of the anti-noise, stability and so on; and analysised characteristics of each operator, and shortcomings of each algorithm in the detection of tungsten picture. So this paper suggested a solution to the problem - SUSAN algorithm based on edge detection. Which focuses on the SUSAN algorithm is used in tungsten mine the desirability of image edge extraction.Finally, based on mathematical morphology is a nonlinear filtering method that can be used to solve the noise suppression, feature extraction, edge detection, image segmentation, texture analysis, image processing problems. To improve the image detection results, this paper presented an improveing based on combination mathematical morphology and SUSAN operator edge detection method. The experiments show that the method is effective and feasible.