Rain-streaks Detection and Removal in Single Image Using Curvelet Transform

This study implements a new way to address the issue of rain streaks detection and elimination from a single picture based on the transform of Curvelet. This approach depends on a decomposing of the rainy image into different scales and sub-bands frequencies by using the curvelet transform. Features have been extracted from each sub-band frequency and the neural network will classify these features into "rain" or "non-rain" signatures. The reconstructed image is obtained without the sub-bands that have the rain signature. The findings from the experiments indicate that the proposed approach improves the visualizing quality as well as PSNR and outperforms previous rain removal algorithms.