A Ceramic Image Restoration Scheme Based on Cloud Platform

Aiming at the poor performance of traditional clustering algorithm in image inpainting, an improved algorithm based on Hadoop distributed platform is proposed for ceramic image inpainting. For the defect that the threshold parameters can not be determined automatically in the process of salient edge structure extraction, we propose a threshold descent extraction algorithm based on Canny edge detection. Then, the improved algorithm is combined with fuzzy clustering and parallelized on Hadoop platform, which is applied to crack extraction, fitting and repair of ceramic image. The experimental results show that such scheme improves the final repair effect of the damaged edge, reduces the running time of the algorithm, and greatly improves the overall efficiency of the restored image.