300MN液压机活动横梁横向偏移视觉检测图像分割方法研究及应用  

Research and Application of Image Segmentation Method for Vision Inspection System of Activity Heam's Lateral Excursion of 300MN Hydraulic Press

作　　者：彭玉凤[1] 谭建平[1] 陈晖[1] 全凌云[1] 

机构地区：[1]中南大学机电工程学院,湖南长沙410083

出　　处：《仪表技术与传感器》2010年第5期80-82,85,共4页Instrument Technique and Sensor

基　　金：国家高技术研究发展计划(863计划)资助项目(2007AA04Z417)

摘　　要：图像分割是视觉检测的关键技术。针对300 MN模锻液压机活动横梁横向偏移视觉检测系统现场光强不稳定的难题,在灰度拉伸Otsu法基础上提出了一种改进的阈值自适应新方法——灰度增强法,并结合其目标物特征,提出了基于圆斑特征值的分割评价准则。介绍了其基本原理,并进行了实验测试。结果表明,根据该方法确定的阈值对图像进行二值化处理,能适应更宽的光强范围,得到理想的图像分割效果,且在此基础上进行的二维位置检测实时性好、检测精度高、稳定性好,能够适应液压机现场复杂的使用环境,实现活动横梁横向偏移检测功能。Image segmentation is the key technology of vision inspection.An improved new threshold adaptive method,grayscale strengthen method was proposed,which was based on grayscale stretching method,and direct to the difficult problem of unstable intensity of illumination in the vision detection system of activity beam＇s lateral excursion of 300 MN hydraulic press.Binding to the features of its target,a segmentation evaluate criteria due to the features of circular spot was also presented.Introduced its basic principles,and carried out the experimental tests.The result shows that using the threshold,which is obtained from this method to operate image binarization can adapt to wider scope of intensity of illumination,getting ideal images segmentation.And the two-dimension position detection which based on it has good real-time and stability,high inspection accuracy,it can accommodate the locality complex using condition of the hydraulic press,and achieve the detect function of the activity beam＇s lateral excursion.

关 键 词：液压机 机器视觉 阈值自适应 活动横梁 

分 类 号：TG315.4[金属学及工艺—金属压力加工]