Research On Carrying Current Safety Of Catenary Based On Thermal Imaging Technology

Catenary is one of the key components in the electrical railway, which provides the sole source of current for traction power of the electrical locomotive without spare devices, its failures can affect the normal operation of the electrical railway directly. However, the most directly manifestation of carring current safty failures is overheating, which is generated by overcurrent or poor contact, and the carring current safty failures are found difficultly, so it leads to current receiving of electrical locomotive improperly. Above all, to detect the heating circumstance of the contact lines and ancillary equipments on line can find out risks of carring current safty for catenary, and ensure the normal operation of the electrical railway. The infrared thermal imaging technology is used to detect the temperature changing of current, which is a powerful method. Making sure that the method can detect the carring current safty for catenary, so it has practical significance. Thesis has completed the following.In order to eliminate the influence of noises, which is produced in the process of the acquisition for infrared images, in the basis of analysis for image pre-processing and feature of infrared images, the specific image pre-processing of infrared images for catenary was studied. The specific image pre-processing of infrared images for catenary includes image gray, image enhancement and image filtering method, the combined image filtering algorithms which the median filter and a high pass filter are combined was presented for eliminating the salt&pepper noises of catenary, and the filtering effect of the neighborhood average method, median filter method, selective masking filtering method were compared. The results show that the effect of the combined filtering algorithm is best.The principle of traditiongal threshold segmentation algorithms was introduced deeply for extracting the heating areas of infrared image for catenary. The global bimodal segmentation algorithm based on temperature threshold of infrared image for catenary was presented. The image was segmentated with traditiongal threshold segmentation algorithms and the global bimodal segmentation algorithm based on temperature threshold. The results of image segmentation show that extracted the heating areas of the global bimodal segmentation algorithm based on temperature threshold are most accurate.The matching algorithm based on mutual information of infrared image was used to judge position for the heating areas in the infrared image, which have obvious shape feature. To find the original point and original direction in the image of the heating areas, and compute the maximum mutual information in the correspondended matching image, so the images achieved image matching and the position of the heating areas in the infrared thermal image was judged. Non-Parametric(NP) window in the image matching exsits the problem of long interpolation compution time, Modified Non-Parametric(MNP) window method was introduced into bilinear interpolation, the results of image matching show that interpolation compution time is short, and MNP window is an effective method of interpolation compution.The image fusion algorithm of infrared image and visible image of catenary based on Contourlet transform was used to judge position for the heating areas, which have not obvious shape feature. The infrared image and visible image of catenary was decomposited by multi-scale, on the basis, the high-frequency and sub-band fusion were integrated with criteria D and the fused image was reconstructed. The results of evaluation parameters of image fusion show that the image quality of fused image is better and is determined to achieve position judgement for the heating areas.