Research And Error Analysis On The Vision Measures Algorithms Of High-accuracy NC Tools

In the process of production, tool parameter has direct influence on the precision of processing. Therefore, in order to ensure the quality of processing and the qualification rate, the tool parameter must be measured in the process of production. At present, tool testing equipment used in industrial production is mainly mechanical testing equipment, projection tool presetting device, and the digital tool presetting device. Due to these detection equipments have their own advantages and disadvantages, these equipments can no longer meet the growing needs of industrial production. So developing high-performance numerical control cutter test equipment is very imperative.The visual tool detector which has been studied on by this thesis can automatically measure the angle of the tool, the diameter of the tool and other parameters. The image processing algorithm has been researched on, the main content include edge detection algorithm, boundary tracking algorithm, straight-line fitting algorithm, arc fitting algorithm and error analysis. In order to improve the precision of measuring, the cubic polynomial fitting algorithm is adopted first for sub-pixel edge detection in the process of the edge detection. In the course of experiment this algorithm can't measure out the edge accurately when the edge measured is very precipitous. Gauss of contours fit of the gradient direction and measure algorithm has been adopted by this paper instead of the cubic polynomial fitting algorithm, this algorithm obtains the position on the edge through calculating. In order to improve the efficiency of detection, the side of the border has been used while simultaneously a straight line tracking has been fitted during the tool angle. The angle of the tool is the angle between two lines obtained by the slope of two lines. The same side of the border boundary tracking is adopted while at the same time fitting the arc to obtain the arc diameter. Model-based tracking algorithm is proposed during boundary tracking, and finally the detection algorithm of tool parameters is verified by experiments, and the measurement errors that mainly exist in the test system are analyzed. The sub-pixel edge detection algorithm used in this thesis firstly improves the detection accuracy. In addition, since the sub-pixel edge detection does not adopt the traditional test procedures but utilize the boundary tracking, linear fitting and arc fitting simultaneously in the testing process. Thus the detection efficiency has been greatly improved.