Research On High Speed Railway Equipment Identification Based On Edge Detection

With the great development of high-speed railway in our country,the safety of railway transportation has become the focus of attention.It is imperative to adopt technical means to deal with this problem by manually monitoring the inefficient operation of HSR equipment.The purpose of image processing is to denoise,split,recover,enhance and extract features by using computer.Image edge is an important feature of the image,but also an important feature of the image.From the 1960 s,image edge detection gradually attracted attention and became a hot spot.For a long time,many scholars have put forward many classical algorithms for image edge detection.However,the conventional edge detection algorithms are not robust to noise and lack of adaptability,and there is still much room for improvement.Traditional edge detection algorithms are based on the differential operator to calculate the gradient of the image,This paper first discusses the current situation of edge detection at home and abroad as well as the basic processing steps of edge detection.Secondly,the advantages and disadvantages of traditional edge detection algorithms including Roberts operator,Sobel operator,Prewitt operator,LoG operator and Canny operator are compared respectively.After that,we improved Canny operator and Laplacian operator and applied it to the high-speed image edge detection.The work of this paper is mainly reflected in the following aspects:(1)Comparing the traditional edge detection algorithms,including the first-order differential operator,the second-order differential operator,the LoG algorithm and the Canny algorithm,the advantages and disadvantages of the image are compared through three types of images.(2)Aiming at the defect of Canny operator,the Gaussian filter is changed to an adaptive median filter,and the threshold of the algorithm is chosen by Otsu algorithm.Process small,fake edges in high speed railway equipment by using linear edge enhancement techniques.(3)The basic theory of wavelet transform is introduced.By contrasting different threshold processing functions and threshold selection methods,different wavelet bases are respectively processed for image denoising to find the most suitable processing method for high-speed railway images.Finally,the local adaptive BayesShrink threshold combined with Daubechies wavelet base can bring better denoising effect,combined with Laplacian edge detection operator can get better edge detection.