Assessment of ground surface roughness based on computer vision technology

Aiming at different ground surfaces of hardened bearing steel GCr15, this paper made experimental research on assessment method of surface roughness based on computer vision technology, Firstly, some pre-processing of the ground images should be carried out to eliminate noise and get more detail information, including image enhancement and median filtering. Then the method of power spectrum analysis transformed representation of processed image from spatial domain to frequency domain by adopting two-dimensional Discrete Fourier Transform. Gaining the mean power spectrum named E and its corresponding radius r, we made efforts to seek the direction in which the arithmetic average surface roughness Ra varied according to E and r. After that the variation rule can be regarded as an assessment basis of ground surface roughness.