Analysis And Research Of Camera Calibration In Contact Rail Measurement

With the high speed development of subway construction, the power supply mode of contact rail with low energy consumption, installation easy and corrosion resistance has extensive application prospect in subway power supply system. The power supply mode of contact rail with low energy consumption, installation easy and corrosion resistance has extensive application prospect in subway power supply system. It’s a key issue to security and reliable operation of electric locomotives by maintaining the stability and safety of traction power supply system for subway. Computer vision measurement has gradually become the important means of contact rail detection with non-contact, high speed, and high precision. In the field of computer vision, Camera calibration is an essential step, and is the premises and foundation of computer vision recovering3D space information from2D image. In this context, this paper studies the technique of camera calibration in depth, and according to the test environment of contact rail.The basic principles of camera calibration start, a brief introduce to several commonly used calibration method, based on contact with the rail testing environment, the application of laser triangulation and binocular image calibration. A laser triangulation method is the used of the laser triangulation principle, the use of camera calibration device, measuring more than more than a dozen groups of image pixel coordinates and the corresponding contact rail geometric position, through the camera’s linear imaging model, according to Matlab find the image pixel coordinates transformation between the department and the contact rail geometric location, and finally selected10different image pixel coordinates transformation between image pixel coordinates system has been obtained and the contact rail geometry location, contact rail geometric spatial location. Track geometrical position will seek contact with the actual location of the error analysis to verify the validity and accuracy of the calibration method; another laser rangefinder calibration method is the use of calibration devices were measured in10groups of image pixel coordinate system of uv which corresponds to the contact rail geometric spatial location p, h, using least squares support vector machine to fit the processing, and finally the fitted values of the same pixel position and the actual measurement value error analysis, to verify the calibration method for the effectiveand accuracy.Binocular imaging calibration method can be divided into a linear array camera calibration and surface array camera calibration method. Linear camera calibration by linear image processing, extraction of feature points, using the principle of binocular imaging to obtain image pixel coordinate system and contact rail geometry of the spatial location of the conversion relationship, proved the same paragraph, when the inspection vehicle at different speeds through contact rail, linear array image acquisition is very different, so the calibration does not apply to the contact rail testing; plane array camera calibration method using camera linear imaging model, image pixel coordinate system and the location of the contact rail geometric space relationship through the analysis and study of the relationship, to know the calculation of the calibration method is too complex, real-time is not high, does not apply to the contact rail test.According to the selection of different types of cameras, calibration method based on binocular vision can be divided into two classes:a calibration method for binocular vision based on line-scan digital camera and a calibration method for binocular vision based on array camera. A calibration method for binocular vision based on line-scan digital camera extract the feature point by image processing algorithm, then uses the principle of binocular vision to get the relationship of the pixel coordinate system and the horizontal distance and vertical distance between camera lens and contact rail. This paper does camera calibration test at first, then uses this the calibration method in the online and detection system of contact rail and experimental analysis. A calibration method for binocular vision based on array camera uses the same performance and at the same horizontal position array cameras and camera model to get the relationship of the pixel coordinate system and the horizontal distance and vertical distance between camera lens and contact rail. But the calibration method of solving process is too complex, and this paper only introduces the method of solution process.