A high-speed image super-resolution algorithm based on sparse representation for MEMS defect detection

A novel high- speed image super-resolution algorithm based on sparse representation for MEMS defect detection is proposed in this paper. Traditional super-resolution algorithms adopt a single dictionary to represent images, which cannot differentiate varieties of image blocks and leads to slow processing speed. Aiming at overcoming this shortage of traditional super-resolution algorithms, image blocks are divided into different categories by local features and each of these categories possesses the corresponding high and low resolution dictionary pairs. Experimental results of different MEMS defects show that the improved algorithm can obtain images of little lower quality with much less processing time, indicating that the proposed algorithm is more suitable for MEMS defect detection.