基于图像处理的接触网定位线夹目标检测

摘要

随着接触网故障检测技术的发展，接触网悬挂状态检测监测装置可得到接触网定位装置的海量高清图像用于线下故障排查。为提高故障检测的效率和准确性，基于图像处理技术，根据定位线夹区域的结构特征设计定位线夹目标区域的检测算法，并通过实际图像数据验证算法的有效性和通用性。该算法对于背景复杂、光照不均、含噪声等现场实际图像均具有良好的适应性，算法检测准确率达90％以上，极大提高了检测效率并为进一步的缺陷检测奠定了基础，具有重要的实际工程意义。%With the development of catenary fault detection technology,catenary suspended state detection and monitoring device can obtain massive high -definition images for offline troubleshooting.Based on image pro-cessing technology,this paper proposed an algorithm of steady clamp detection method according to the structure features toimprove the detection efficiency and accuracy.By applying the method to practical image databases, the results show that the method could achieve good performance and detection accuracy rate could reach 90%. Even for poor quality image such as background complex,noisy,underexposure or overexposure,the effective-ness and generality of the method are still reliable.The algorithm greatly improved the detection efficiency and lay the foundation for further defect detection,of great practical engineering significance.