Surface defect detection of curved mosaic ceramics based on improved coupling denoising algorithm

Mosaic ceramic murals have been valued as remarkable artistic expressions, with three-dimensional patterns providing a greater significance. The surface quality of these ceramics is crucial to the overall aesthetic of the art. However, the detection of surface defects is difficult because of the unique curved surface structure of curved mosaic ceramics. To solve this problem, this paper proposes a visual detection method for curved mosaic ceramics combining an improved coupled denoising algorithm with an adaptive threshold segmentation algorithm. This overcomes the limitations of conventional single-image enhancement algorithms in machine vision, which cannot effectively process images comprising mixed noise signals. The surface defect images of curved mosaic ceramics are first enhanced by combining an optimized Gaussian filter with an improved Fourier convolution transform. Many steps, including adaptive threshold segmentation and feature screening, are then implemented to quickly identify surface defects in curved mosaic ceramics. Afterwards, experiments are conducted, demonstrating that the proposed method achieves an accuracy of 95 % in detecting bulges, scratches, bruises, and cracks on curved mosaic ceramics. Finally, it is shown that the proposed approach is also applicable to other curved-surface products.