基于改进像素聚类算法的输电线路冰区识别方法
2017-10-17杨恒虢韬沈平杨渊杨洋
杨恒 虢韬 沈平 杨渊 杨洋
摘 要: 为了提高输电线路的防灾减灾能力,特别是防治冰雪灾害的能力,需要进行输电线路冰区自动识别方法和系统设计,提出基于改进像素聚类算法的输电线路冰区识别方法。对气象卫星遥感监测采集的输电线路覆盖区域的遥感图像进行图像处理,提取图像中的冰覆盖像素特征点,结合边缘轮廓检测方法进行输电线路覆盖区域的冰区边缘轮廓检测,对冰区覆盖区域的像素特征进行模糊C均值聚类处理,实现区域分割识别。仿真结果表明,采用该方法进行输电线路冰区识别的准确率较高,视觉直观效果较好,便于指导施工抢修人员进行现场排除和巡线检修。
关键词: 像素; 聚类; 图像处理; 输电线路; 冰雪灾害
中图分类号: TN911.73?34; TP399 文献标识码: A 文章编号: 1004?373X(2017)20?0181?03
Abstract: In order to improve the disaster prevention and mitigation capability of power transmission lines, especially ice and snow disaster prevention capability, automatic identification method and system of power transmission line ice coverage areas need to be designed. Therefore, a method of power transmission line ice coverage area identification based on the improved pixel clustering algorithm is proposed. The remote sensing images of power transmission line coverage areas from remote sensing monitoring device in meteorological satellites are processed to extract the pixel features of ice coverage areas in images. Combined with the edge contour detection method, the edge contours of power transmission line ice coverage areas are detected, and the pixel features of ice coverage areas are processed with fuzzy C?means clustering to realize region segmentation and recognition. The simulation results show that this method has high accuracy and perfect visual effect for recognition of power transmission line ice coverage areas, and is appropriate to guide the repair personnel to do on?site troubleshooting and maintenance.
Keywords: pixel; clustering; image processing; transmission line; snow disaster
发生冰雪自然灾害时,随着冰雪积雪厚度的增加,将会导致输电线路出现坍塌,造成严重的经济损失。由于输电线路大多分布在野外,线路分布广阔,长度较大,采用人工巡线检测方法进行冰雪灾害预测和防灾减灾的收效较小,需要研究一种可行的智能的输电线路冰区识别方法,针对性地对重点区域进行除冰作业,保障线路正常运行,研究输电线路冰区智能识别方法具有重要的现实意义[1]。随着图像处理技术的发展,图像处理广泛应用在模式识别、智能监控等领域,对输电线路采用图像处理和信息分析方法进行冰区识别具有可行性[2]。为了实现对输电线路的全天候冰区识别和监控,提出基于改进像素聚类算法的输电线路冰区识别方法。……
