工业检测中图像采集标定与预处理技术研究

Research on Calibrating and Preprocessing Technology of Image Acquisition in Industrial Detection

为了较好地解决工业现场劳动强度大、效率低下的不足,彻底消除工业现场中可能存在的威胁和危害操作人员人身安全的因素,针对工业图像检测技术中的校正、识别问题,提出了摄像机的快速标定方法及视频信号的有效提取算法,并给出了工业现场应用的案例。测试数据及现场运行情况表明,对有一定变形的检测对象,该算法仍可以准确地提取其几何特征,达到了系统要求。经测试,定位精度为±3 mm。较好地解决了工业现场摄像机参数的快速标定及噪声条件下图像信号的有效提取问题。

Aiming at resolving the problem of calibration and recognition in industrial image detection technology, fast calibration approach for camera and efficient extracting approach for video signal are presented, and an application case in the industrial environment is given. The test data and performance indicate that the proposed algorithm is able to extract geometrical features with some distorted detecting objects accurately, and its location precision is ± 3 mm, which reaches the requirement of the system. The algorithm and technology proposed resolved fast camera parameters demarcate and efficient image signal extraction with noisy conditions.