Status Detection Of Railway Catenary Based On Image Processing

The running state of electrified railway’s catenary is closely related to the operational safety of the railway, trains gain electricity power through the catenary, so it is essential to ensure that the catenary works under normal condition. Furthermore, with the railway speed-up and electrification reform, traditional detection methods have become harder to meet the real-time and reliability requirements of the high-speed railway transport. This paper analyzes the development status and practical engineering applications of catenary dynamic detection technologies, and then studies real-time status detection technique based on image process. A integrated and complementary strategy is adopted to get high sampling speed and high coverage by combine1-D CCD and2D CCD camera together, and to respond this video acquisition structure a real-time image processing platform based on TMS320DM642is designed. In order to overcome the spurious background noise, the illumination difference and the camera shaking in the catenary images, rapid detection algorithms for catenary feature extraction, arcing/offline detection and catenary abrasion are considered.This paper firstly analyzes and summarizes the recent developments of catenary detection technologies in engineering application, and then studies the visual features and basic knowledge of real-time image processing. At the same time, relevant algorithms are optimized and transplanted for embedded system, such as noise filtering, linear feature extraction and so on. Secondly, this paper presents the DSP hardware circuits and software design in detail. In this part, circuits of DSP are designed in Altium Design6.0and the softwares for image processing are designed in CCS3.3compilation environment. Finally, design and analysis of algorithms are presented. On the one hand, a rapid line detection method is designed for target identification of catenary. On the other hand, adaptive thresholding algorithm is used for image segmentation, so that clearly distinguish between target and background, which improve the robustness of this system.Base on image process technology, real-time detection on stagers, offline status and abrasion of the catenary is realized in DSP system. Data analysis shows that offline status and abrasion of the catenary can be well assessed, which provides some references for detecting stagers, offline status and abrasion of the catenary.