基于图像处理的目标检测跟踪系统设计与实现.docx

基于图像处理的目标检测跟踪系统设计与实现 摘 要 香蕉型汽车悬架弹簧因其存在弧度，在机械手搬运过程中，会因为弹簧位姿的不同导致弹簧内外圈中心坐标的变化，若不进行弹簧位姿矫正，会导致机械手无法正确地将弹簧放到下一工位进行加工。本文提出了通过CCD视觉检测系统来检测香蕉型弹簧内外圈中心坐标的算法，并应用于昆山孚思格机电有限公司的弹簧认向系统中。 本文在HALCON12的编程环境下，提出了基于特征提取、圆拟合、模板匹配的弹簧内外圈中心检测算法，并进行了对比分析，最终选择了通过特征提取来检测弹簧内圈中心。本文运用单相机标定将相机坐标系转换成世界坐标系。在VS2010的编程环境下，设计了弹簧内外圈检测软件。该软件通过UDP通讯协议与机械手通讯，将测得的弹簧内外圈中心坐标反馈给机械手；该软件通过串口通讯与串口IO控制板卡通讯，当遇到非正常情况触发报警；该软件实时储存检测数据，并记录非正常情况下拍摄得到的照片。 在经过大量实验，进行数据分析之后，得出该软件的检测误差在5%范围以内，检测时间在1秒以内，满足昆山孚思格机电有限公司的应用要求。 关键词：香蕉型悬架弹簧；图像处理；图像特征提取；圆拟合；模板匹配；网口通讯；串口通讯 Abstract During the handing of the robot, curvature of the banana-type car suspension spring, the central coordinate of the inner and outer rings of the spring will change. If the spring posture correction is not performed, the robot will not be able to correctly place the spring to the next station for processing. This paper proposes an algorithm for detecting the center coordinates of the inner and outer rings of banana springs by CCD visual inspection system, and applies it to the system of Kunshan Fusige Electromechanical Company. In the programming environment of HALCON12, this paper proposes a spring inner and outer circle center detection algorithm based on feature extraction, circle fitting and template matching algorithm. With the comparative analysis, this paper use the feature extraction to detect the center of the inner ring of the spring, and use the round fitting to detect the center of the outer ring of the spring. This article uses a single camera calibration on convert the camera coordinate system into a world coordinate system. The spring inner and outer ring detection software design on the programming environment of VS2010. This software communicates with the robot through the UDP communication protocol, and feedback the masured center coordinates of the inner and outer rings of the spring to the manipulator. This software communicate with serial IO controller though serial communication, and it will trigger an alarm when an abnormal con