基于显微视觉的微小型零件跟踪系统

Microminiature parts tracking system based on micro-vision

微装配过程中的运动目标跟踪是一个新兴的研究方向。构建了一个由CCD相机、显微镜头、电控云台和图像处理模块组建的、针对微小型零件的显微视觉跟踪系统。为克服显微视场范围小的局限性,提出一种基于SIFT特征点的模板匹配和Kalman预测相结合的跟踪算法,通过Kalman预测实现在局部范围内的模板匹配,利用SIFT特征对模板匹配的结果进行校正和更新。实验结果表明:提出的跟踪算法能得到稳定的目标局部特征,并准确地跟踪到目标,对亮度变化、成像模糊等影响因素有较强的适应能力。

Moving target tracking in micro-assembly process is a new research direction. Construct a microscopic visual tracking system for microminiature parts,including CCD camera,microscope lens,electric control Yuntai and image processing module. In order to overcome limitations of small microscopic field of view,propose a tracking algorithm combining template matching based on SIFT feature point and Kalman prediction,which realizes template matching in local range by using Kalman prediction and correction and update result of template matching is carried out by using SIFT features. The experimental results show that the proposed tracking algorithm can obtain stable local features of targets and track target accurately,and have a good adaptability to illumination change and imaging blur.