Research On Dynamic Parameters Detection For High-speed Railway Catenary Based On Image Processing

With the rapid development of high-speed railway in China,the safety guarantee during train operation is extremely important.Catenary in electrified railway is for pantograph flow of high voltage transmission line,catenary equipment in good running,is related to the safety of the train.To detect the running state of catenary,China's railway sector has formulated the general technical specification for power supply safety detection and monitoring system(6C system)of high-speed railway.The 6C system collect catenary images by a camera,then analyze and process these images to achieve catenary detection.The vehicle-mounted catenary operation state detection device(3C)is partly installed on the roof of the train in this system,so it will shoot a large amount of video data when the high-speed train is running.The amount of image data is huge,which brings great difficulties to the staff.The article aims at the problem that too many alarm pictures are taken by 3C,which makes it difficult for maintenance personnel to analyze,research on intelligent detection of dynamic parameters of catenary identified by 3C image using image processing technology,the main research contents are as follows:Firstly,the image acquisition system is introduced.According to the principle of camera calibration and the calibration result of field personnel,the distortion of camera lens is calculated,and the coordinate transformation formula is obtained;by comparing the existing methods for detecting the dynamic parameters of catenary,the methods for detecting the dynamic parameters of catenary are selected.Then,in the detection of contact wire height and pull-out value,aiming at the low accuracy of traditional Canny operator edge detection,this article selects an improved Canny edge detection operator.In the process of image smoothing,introducing adaptive median filter algorithm instead of Gaussian filter,which can eliminate the noise and retain the edge details of the image.The iterative improvement method of introducing adaptive threshold processing.According to the characteristics of the catenary image,to adaptively determine the threshold,which can avoid false edge when the edge detection.According to the characteristics of the pantograph,to draw a pantograph profile and determine by the pantograph on the edge.The contact wire is identified,through the contact wire recognition strategy,and substituted into the coordinate conversion formula to calculate the contact wire height and pull-out value.The simulation results show that the improved Canny operator and the method based on Opencv contour extraction can accurately detect the contact line height and draw value.Finally,in the arcing detection of pantograph and catenary,using a two-dimensional Gabor filter to filter for image.Then judging the light received condition of the filtered imageand choosing an improved method of measure of the homogeneity: after the extraction of the amount of light in the unevenly received images,the correction of the light component is based on the adaptive correction principle of the two-dimensional gamma function,the ideal segmentation threshold can be determined by the histogram statistical method,which is to traverse the image to find the optimal segmentation threshold;however,there is no need for two-dimensional gamma function correction for light-receiving uniform images.After extracting the arc burning area through improving the uniformity measurement method,the arc burning recognition is performed by the centroid method.Simulation results show that this method can effectively complete arc detection and meet the detection requirements.