在线视觉测量系统的刀具图像去抖动方法研究

Research on tool image deblurring method of online vision measurement system

研究视觉采集图像时由于机床电动机振动导致的刀具图像抖动问题。对比研究3种经典图像去抖动方法,在自适应L-R算法的基础上,针对自适应L-R算法的耗时性,结合金字塔的结构,研究一种自适应L-R算法—金字塔相结合的去抖动方法,以图像复原质量的主观评价方法和客观评价参量为标准,对该算法的实现进行定性认识和定量分析。同时以实验室自主研发的在线视觉刀具测量系统为对象,对算法实现进行实验验证。此算法具有较高的PSNR值和MSE值,而且计算时间骤减。通过在线刀具图像的采集和算法的实现,验证此算法具有去抖效果。

To study the problem of tool image dithering is studied as a result of motor vibration of machine tools when visual images are collected. Aiming at the problem of image jitter,comparison of three kinds of classic image dithering method,on the basis of adaptive L-R algorithm,the time consuming problem of adaptive L-R algorithm is discussed,combined with the structure of Pyramid,an adaptive L-R algorithm of Pyramid combined to dither method,as to subjective evaluation and objective evaluation of the restoration quality parameters standard,qualitative analysis and quantitative understanding of the implementation of the algorithm. This algorithm has higher PSNR value and MSE value,and the computing time is greatly reduced. Through the online tool image acquisition and algorithm implementation,it is proved that the algorithm has the effect to reduce shake.