圆形工件的弹簧装配检测方法  

Spring-assembly detection method of round workpiece

作　　者：邱潇 季启政 周尚波[1] 范继云 

机构地区：[1]重庆大学计算机学院,重庆400030 [2]航天514所北京东方计量测试研究所,北京100086

出　　处：《计算机工程与设计》2018年第1期218-223,共6页Computer Engineering and Design

基　　金：重庆市基础科学与前沿技术研究(重点)基金项目(cstc2015jcyj BX0124)

摘　　要：为改善当前金属工件装配检测存在的成本高、效率低等问题,提出一种对复杂圆形金属工件适应性更高的零件检测方法。在图像增强阶段,引入自适应分数阶微分增强,结合图像梯度特征,突出工件的细节和纹理信息,提升后续检测的准确度。对于圆形轮廓的金属工件,引入极坐标转换,与传统的直角坐标处理方法相比,其具有效率高、处理过程清晰、不易出错的特点。在阈值检测阶段,通过长度阈值和灰度相关属性阈值对零件数量和类型进行判断。实验结果表明,该方法能准确地定位到圆形工件的弹簧区域,快速检测出弹簧零件区域的装配情况,大幅减少无效定位和不必要的计算,具有时效性高、准确性强的优点,可应用于该金属组件的自动化装配检测。To solve the problem currently exists in workpiece parts assembly detection of high cost and low efficiency,a method with high adaptability in detection of internal parts of complex round metal workpieces was proposed.Adaptive fractional-order differential enhancement was introduced to improve the detail and texture information of workpieces,which further developed the followup detection accuracy.For workpiece with circular contour,log-polar transformation was used instead of Cartesian coordinate system for its advantages of higher time efficiency,clearer process and lower error rate.In the stage of threshold detection,length threshold and grey-level relative threshold were used to detect the specified amount of springs.Experimental results verify that the proposed method can correctly locate the spring area and detect the assembly condition in a short time,while reduce invalid positioning and unnecessary computation time.The method has high efficiency and accuracy,which is good for the application of automatic assembly system.

关 键 词：图像处理 分数阶微分增强 圆形检测 极坐标转换 自动化检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]