Image Measurement System Performance Optimization Research

Vision measurement is an emerging technique in measurement field in recent years. Due to its uncompetitive advantages:stable, uncontact, good environment-adaptive, it will bring a new reform in measurement method. Based on optics, It combines optoelectronics technology, computer technology, laser technology, image process technology and other related modern technology together to constitute a compound and complicated system that include light,mechanical, electric, algorism and so on.To improve the performance of measuring machine, This paper mainly focuses on the research of motion control and light of system, namely analyzes the motion error of system and the affect of light for image and purposes some measurements based on theoretical analysis. For motion control, on one hand, to ensure the motion stability of measuring machine, motion strategy and velocity are optimizated; on the other hand, considering the convenience and the efficiency of operation, tutorial function is added to help machine to find the last position rapidly. Meanwhile, by laser interferometer, the error analysis the compensation of measuring machine are also performed. Experiment shows that the accuracy after compensation can satisfy the accuracy requirement of image mosaic based on hardware. Take the large dimension part as experimental object, experimental result indicates that the mosaic based on measuring machine reduces 50%～75% of mosaic time compared with mosaic based on algorithm.About light control, this paper elaborates how to evaluate the image quality, compares and analyzes some evaluation function, selects gray level co-occurrence matrix to evaluate the image and adjusts the light according to evaluation result. Due to the variation of natural daylight and workpiece material and color, the proper brightness of light for different workpiece is variable.Besides, because the light has four channel, it needs a lot of time to complete the adjustment, namely it is impracticable. To realize the adjustment, we use the experience value which concluded by a lot of experimental data, while we use average gradient algorithm to make sure the picture we get is the best for the back light. While measuring machine captures the image of new workpiece, machine finds the reference part which is most like the current and uses its light data to complete adjustment, and therefore can save a lot of time.At last, this paper introduce interference that System will be interrupted,analyze the causes of interference, and mainly analyzes the interference caused by motion blur. Based on the theory of image restoration and related algorithm, the noised image can be recovered, and finally be evaluated by gray level co-occurrence matrix, the experiment data shows that effect is significant.