APP下载

基于图像识别的无人机输电线路断股检测系统设计

2017-10-17高晓东郑连勇王蔚曹飞

现代电子技术 2017年20期

高晓东 郑连勇 王蔚 曹飞

摘 要: 针对传统的输电线路人工沿线巡检方式成本高、效率和安全性低等缺陷,基于图像识别技术,设计并开发了一套实时的无人机输电线路断股检测系统。该系统利用输电线路直线特征,使用Laplacian算子对输电线路进行边缘检测,经闭运算、二值化运算后得到光滑边缘;使用改进后的霍夫变换与区域种子点获取和生长提取无人机实时拍摄的输电线路图像;最终,根据输电线路的宽度变化判断是否存在断股缺陷。该系统实际使用效果良好,能够明确显示断股缺陷信息,并能为相关输电线路的断股缺陷检测技术提供参考。

关键词: 线路断股检测; 边缘检测; 霍夫变换; 区域种子点; 区域生长

中图分类号: TN913?34; TP393 文献标识码: A 文章编号: 1004?373X(2017)20?0162?03

Abstract: Aiming at the problems of high cost, low efficiency and little safety of traditional artificial patrol, a real?time broken strand defects detection system for transmission line of unmanned aerial vehicle was designed and developed on the basis of image recognition technology. The straight line feature of transmission circuit is used in the system to perform the edge detection and Laplacian operator, and then the closed and binarization operations are adopted in the system to obtain smooth edge of transmission lines images. The improved Hough transforma regional seed acquisition and region growing are employed to extract the transmission line real?time images taken by UAV, and then judge whether broken strand defects exist according to width variation of the transmission line. This system works well in its actual running, can display the information of broken strand defects clearly, and can provide a reference for detection of broken strand defects in the related transmission lines.

Keywords: broken strand line detection; edge detection; Hough transform; regional seed point; region growing

传统的输电线路人工沿线巡检方式存在成本高、效率与安全系数低等缺点,而新兴的无人机巡检方式却能有效改善上述缺陷,为高压输电线路定期巡检提供新的方案[1]。断股缺陷作为高压输电线路常见且危害性高的重要缺陷,在无人机巡检方式的研究和分析中,存在与输电线路类似的线性特征区域干扰较难排除、输电线路环境和背景复杂引起的算法不可靠等问题[2?6]。因此,基于图像识别技术,本文引入改进的霍夫变换,选用区域生长种子点处的平均像素作为阈值参考,设计并开发了一套无人机输电线路断股检测系统。该系统能有效解决上述干扰难以排除、算法不可靠等问题,从而对输电线路进行准确检测及判断。……

登录APP查看全文