Multi-camera visual inspection of abrasives distribution density on electroplated diamond wire saw surface

Electroplated diamond wire saw is widely used in slicing of hard and brittle materials, such as monocrystalline silicon, sapphire, single crystal silicon carbide, and ceramics. The abrasive distribution density on the surface of electroplated diamond wire saw will directly affect its slicing performance, which is the main index of diamond wire saw quality inspection. Four CCD cameras were used to acquire the surface images of diamond wire saw at the same time, and the images were preprocessed by Gaussian filter. The images were expanded based on the cylindrical model, and the image mosaic was completed through the steps of feature point matching, feature point screening, registration model solving, image resampling, and image overlapping area removal. Then the number of abrasives in the whole image was extracted by using the connected region labeling. The results show that the proposed method for detecting the abrasive particle distribution density of the full-surface image of the electroplated diamond wire saw can improve the accuracy of the online detection of the abrasive particle distribution density. Compared with the half surface image results detected by a single camera, the maximum relative error of the detection results of the abrasive distribution density decreases by 29.6%, and the average relative error decreases by 17.3%.