基于RBCT算法的复杂背景下铁路接触网电力线自动提取研究

Research on Automatic Extraction of Railway Catenary PowerLines Under Complex Background Based on RBCT Algorithm

针对视频监测得到铁路接触网图像背景复杂、目标细弱而使电力线提取困难、效率降低的问题,提出一种结合背景特征和改进Ratio算子的区域链码(Ratio-based background features and chain code tracking,RBCT)电力线提取方法。首先,分析接触网电力线图像背景特征,对不同类别背景图像进行灰度值分析、预处理加强来消除背景噪声并增强电力线目标。然后,利用分析得到的线特征检测阈值进行边缘检测,对电力线目标边缘做Ratio算子处理。最后,对检测得到目标边缘进行四方向链码聚类分析,识别电力线目标。结果表明,相比传统Canny、Ratio边缘提取方法,所提方法具有良好的抗噪能力和更高的电力线识别精确度。算法能够消除边缘检测出现的噪声,解决电力线检测中出现的断股、分裂问题,可准确完整地提取不同种类复杂背景下的电力线目标,具有较高工程应用价值。

Aiming at the problems of complex background and weak target in video monitoring of electrified railway catenary image, which lead to the difficulty and low efficiency of power line extraction, a power line extraction method combining Ratio-based background features and chain code tracking(RBCT) is proposed. Firstly, the background features of catenary power line images are analyzed, and the gray value of different kinds of background images is analyzed and preprocessed to eliminate the background noise and enhance the power line target. Then, the edge of the power line target is detected by using the line feature detection threshold obtained from the analysis, and the Ratio operator is used to process the edge of the power line target. Finally, the edge of the detected target is analyzed by cluster analysis of four direction chain codes to identify the power line target. The results show that, compared with the traditional Canny and Ratio edge extraction methods, the proposed method has good anti-noise ability and higher power line recognition accuracy.The algorithm can eliminate the noise of edge detection, solve the problem of broken strands and splits in power line detection, and can accurately and completely extract power line targets under different types of complex backgrounds,which has high engineering application value.