Machine Vision Based Anti-engine Piston Assembly Mistake Detect Technology Research

Machine vision research using the computer to simulate explicit biological or the function of the macro perspective, and it is an integrated technology, which includes the digital image processing technology, mechanical engineering technology, control technology, lighting light source technology, optical imaging technology, sensor technology, analog and digital video technology, computer hardware and software technology, human-machine interface technology, etc. Machine vision system acquires the target image by the machine vision products, and then it sends to the exclusive processing systems, which take characteristics of the target image according to the distribution of pixel and brightness, color and other information into the target image and give the electrical signal to control the operation of the equipment. In the field of industry, Machine vision has been successfully used to identify the parts and positioning, product quality inspection, tool wear monitoring, precision measurement and mobile robot navigation and other fields.Motor vehicle is a key component, along with the automobile industry development in our country, and the engine assembly line is leading to the flexible assembly line direction. Flexible assembly line is characterized by the assembly can beat a certain freedom within the framework of adjustment, can be realized more productive varieties. Automatic engine assembly line at the error proofing installations and the production line engine testing technology are an important direction of motor vehicle equipment line, to ensure the engine assembly line realization of multi-species flow production line, is an important aspect of increasing the degree of automation. Fault found sooner, it caused less damage, on-line technology could be used to test for manufacturing defects to achieve the maximum capacity of the retrospective. Therefore, if we can take the lead in this research results and applications, it will certainly promote the machine vision technology, achieving rapid detection products, at the same time improves the level of industrial technology in automotive engines and the efficiency of production, enhances enterprise to adapt to rapid changes in the market's ability ,to generate huge social benefits.This article first about the machine vision as well as a number of theories related to the application, then introduced the topic of the meaning and purpose of the research study of the status quo at home and abroad, put forward the main elemerits in this paper. This article related to the use of machine vision theory, in light of the actual subject of testing and inspection requirements, take into account the choice hardware. This article focuses on the selection of the CCD camera, video capture card, camera and light source. In addition to carrying out a systematic framework for software design, through which the image of the template matching technology and template matching algorithm of choice to achieve the extraction of character image; the adoption of image enhancement, threshold segmentation, morphological transform and character segmentation and a series of image processing methods to achieve the image pretreatment; then by pattern recognition theory, select the pattern recognition method to identify and determine the steps; finally, the realization of the classifier of the design, we will use the BP algorithm neural network classifiers to achieve the ultimate goal of the classification.In this paper, a combination of digital image processing, character recognition , pattern recognition, classifier design and relevant technologies. The use of digital computer through the Visual C + + programming to realize algorithm theory procedures, has designed a Piston assembly quality of visual inspection of the software testing environment. The results show that the research of the piston assembly quality of the visual detection system is relatively satisfactory.