The Curvelet Compressed Sensing Denoising Algorithm for Tobacco Insect Images

In tobacco insect image denoising, since the uniform threshold in iterative thresholding algorithm cant separate the image from noise completely in each step scale causes much noise is still presented in the reconstructed image, an improved curvelet thresholding compressed sensing thresholding algorithm is proposed. In the proposed algorithm, a curvelet wiener filtering denoising operator designed is used in tobacco insect image sparse transform for removing most noise information. Then an improved curvelet thresholding is proposed to remove remaining noise for further improving the reconstructed image quality. The experimental results on all kinds of tobacco insect images show that the proposed algorithm can remove noise from the noisy tobacco insect image effectively, preserves mores detailed features. Even when compression ratio is low, it still has a higher peak signal to noise ratio.