Abstract: In order to improve the inspection efficiency of the elevator traction wheel groove wear, a method based on image processing technology for identifying abnormal wear of elevator traction wheel groove was proposed. The relative discrete feature parameters of the traction wire ropes were extracted to characterize the cumulative abnormal wear of traction wheel groove, and the modified Canny operator was used to denoise and detect edges of the forward and lateral images of traction wire ropes. On this basis, the edges of ropes in the forward and lateral images were extracted by random Hough transform and least squares method. Finally, the abnormal wear degree of the traction wheel groove was calculated according to the number of pixels covered in images of tranction wire ropes. The results show that this proposed method can accurately evaluate the abnormal wear state of the traction wheel groove, and that compared with traditional measurement methods, this method can not only improve the inspection efficiency but also save the test cost.

Key words: traction wheel groove, abnormal wear, Canny operator, random Hough transform, edge detection