Computer Vision Based High-Precision Multi-size Measurement Method And Application For Valve

Automobile industry as a higher degree of automation of the high-tech industry, the demand of the high-precision testing of the product is increasing, and a lot of advanced automation technology has been successfully applied to the various production processes in the industry. Machine vision as one of the emerging technologies of the21st century has been widely used in all aspects of automotive manufacturing, such as dimensions and appearance quality detection of auto parts, auto-Assembly of correctness testing. Previously traditional methods of detection are labor-intensive, and vulnerable to the impact of workers’ subjective emotional and their own skill level of workers, a high test pass rate is not guaranteed. Application of machine vision systems can significantly reduce testing costs and improve test accuracy, increase production speed and efficiency.The object of this paper is engine valve size measurement, the detection algorithm of variety sizes of valves based on computer vision, improve detection precision and efficiency. In this paper, detection system using high-speed CCD extract valves image, and median filtering the image, and then locate the rough edge of the target area, use quadratic curve fitting algorithm to get precise sub-pixel coordinates. Multiply the difference between the left and right edge of the coordinates by pixel equivalent and equivalent to error compensation to get precise diameter, and finally achieve by computer programming, and actual application and apply to practical production. Main research contents are as follows:1. According to the valve size detection requirements and the images collected by the system, research the detection algorithm of dimensions. In the algorithms of Pixel flat-field correction, standard measurement calibration, multiple algorithm based on edge detection of sub pixel positioning, for further study, a new higher precision, more stable, and higher speed test methods to achieve valve image edge such as horizontal, vertical, conical and an circle high-precision detection. Precise extraction the edge above, and establish its describing function, using describing function intersects the edge position to determines the focus of features. 2. Defined the scope of testing range as far as possible in the pure cone conic surface and setting of the cone location detection search with the same diameter as the diameter of the gauge. Within the scope of testing range, fitting two cones line and get the angle between two lines.3. Extended detection data to20dimensions of data, and research the measure algorithms. Expand the database records, based on convenience to modify the order of items in the database. According to the software engineering thought, to optimize the entire software architecture, modular design of share functions. Software programming, to achieve the above algorithm, applied to the accurate detection of various sizes actually.