Research On Three-dimensional Reconstruction Of Rail Surface By Phase Measurement Of Single Grating Stripe Image

With the development of the economy and the increase of social demands,the task of railway transportation is gradually increasing.The railway is developing towards heavy-duty and high-speed tends.As the most basic and important part of the wheel-rail touch system,the “health” of the rail is directly related to the safe driving of the train.Therefore,the detection of the rail is very important.If the shape and size of the rail defects can be reconstructed,there is no need to subjectively identify and judge these defects through the human eyes,which will save a lot of manpower and financial resources.By searching domestic and foreign references of defects detection and three-dimensional reconstruction of rail surface,the thesis finally decided to use the Phase Measurement Profilometry to carry out three-dimensional reconstruction of rail surface.In the thesis,for the Phase Measurement Profilometry,when using phase-shift method to extract the phase,more than three phase-shift grating stripe images are needed,which makes the whole reconstruction process complicated.It is proposed to extract the phase in the thesis by using single grating stripe image to reconstruct three-dimensional sizes of the rail surface.The main ideas are as follows: realizing the phase shift of the grating stripe by Hilbert transform on the single grating stripe image and extracting the phase by using the derived phase-shift formula to reconstruct the three-dimensional shape of the rail surface.Combined with the traditional Stoilov phase-shift method,simulation and comparative analysis are carried out.The simulation results show that the proposed method can extract the phase of the grating stripe well and realize the three-dimensional reconstruction of the rail surface.The error is not greater than the traditional Stoilov phase shift method,and it takes 48 ms to restore the shape of the rail surface by Stoilov phase shift method.The proposed method only needs 10.5ms,and the method runs faster.The experimental platform is built to verify the phase extraction algorithm in the thesis.Firstly,the industrial camera is controlled to acquire the grating stripe images.According to the characteristics of the rail surface and the bottom of the rail,the maximum brightness sum algorithm are used to extract the grating stripe image of the rail surface area.Then,three widely used filtering algorithms are used to filter the grating stripe image of rail surface.Through analysis and comparison,the grating stripe image of rail surface with the best filtering result is selected.Then,the filtered grating stripe image of rail surface is transformed by Hilbert transform to extract the phase of rail surface grating stripe,and the rail surface is reconstructed by combinating with the relationship obtained by calibration.Finally,the hand-made defects were reconstructed.The experimental results show that the phase extraction algorithm based on Hilbert transform can reconstruct the three-dimensional shape of rail surface and its defects,and meets the requirements of rail three-dimensional detection.