Intelligent Detection System Of High-Speed Rail Station Based On Image Processing

As a developing country, China has a large population. The Chinese national economy developed very fast in recent years. People’s living conditions are improving as well. Their demands of safe, quick, comfortable travel are increasing. Therefore, The Chinese government has been devoted huge government to the Chinese Railway undertakings, especially for the Chinese High-speed Rail since2002and it has been made many remarkable achievements. The development of High-speed Rail construction was very fast in recent years, the facilities about the High-speed Rail are all over our country. In order to satisfy the passengers’need, the construction of the High-speed Rail station tends to be more large-scale and modern. As a result, the insiders are paid more and more attention to the security risks of the tall buildings’roof. Aimed at the conditions of the possible security risks of the tall buildings’roof, this paper put forward a way that applying the image processing technology to detect the foreign bodies or check exception.This paper designs a system which name is intelligent maintenance system of High-speed Rail station that based on Image processing technology. This system completes in the platform of Microsoft Visual Studio2010.The part of image processing has several steps as followed:The detection of the image’s feature point、Using RANSAC to select the accurate match points、Estimate the space transform matrix of the images、Do the operation of perspective transformation to the new image、Do the operation between the original image and the new image、The image’s detection and labeling. The paper writes several corresponding algorithms by means of OpenCV function library. The paper basis on the analysis of the screening and experiment on each step of the algorithm. Such as:In the detection of the image’s feature points, This paper uses two classic Corner detection methods which names are Harris and SIFT to detect、 descript and match the image in different scales. The experimental results show that the SIFT has the properties of scale invariant. This paper uses other two classic Corner detection methods which names are SURF and SIFT to detect the same image and record their operation time. The experimental results show that the SURF has a higher efficiency. This paper uses SURF to detect、descript and match the original image and the new image for getting the relationship of them. Aimed at the condition that there are many pairs of wrong feature point between the original image and the new image, this paper uses an algorithm which is called RANSAC to select the accurate pairs from them. Aimed to get the conversion space of two images, we need to get the transformation matrix between the two images. We can estimate the space transformation matrix between the two images by applying the accurate pairs. Then we can use the transformation matrix to realizing the new image’s Perspective transformation and let the new image registry with the original image. Then we use the image shear algorithm to deal with the two images and let them to have the same size. Then this paper uses the algorithm of pixel subtraction to deal with the two images which have the same size and mark the foreign body. This paper uses the algorithm of image’s erosion to remove the noise and get a better effect of foreign body detection.The test’s results show that in many circumstances in the practical application, such as: the displacement, the environment’s bright or dark, the image’s big or small and the rotation, the Intelligent maintenance system of High-speed Rail station that based on image processing can get a good effect in the actual application.