The Research Of Image Enhancement Algorithm Based On NSCT Transform

From imaging to completely reception, images of the human receives are often interfered by various factors like the disadvantage of hardware and deteriorated some new noise in transmitting procedure. These issues may reduce the quality of the obtained image which has a small dynamic range, low-contrast, blurred edge and noise pollution. And that could influence subsequent analysis of the image processing. As an important technical, image enhancement could improve the quality of the image according to different specific requirements. At the same time, it could weaken the imformation that we don’t need as far as possible in the image. So, it is effectively achieved the purpose of image enhancement.In this paper, a new remote sensing image enhancement method using mean filter and unsharp masking methods based on non-subsampled contourlet transform(NSCT) in the scope for greyscale images is proposed. First, the initial image is decomposed into the NSCT domain with a low-frequency sub-band and several high-frequency sub-bands. Secondly, linear transformation is adopted for the coefficients of the low-frequency sub-band. The mean filter is used for the coefficients of the first high-frequency sub-band. Then, all sub-bands were reconstructed into spatial domains using the inverse transformation of NSCT. Finally, unsharp masking was used to enhance the details of the reconstructed image. The experimental results show that the proposed method is superior to other methods in improving image definition, image contrast and enhancing image edges.Next, a new medical image enhancement method is proposed in this article. Firstly, the initial medical image is decomposed into the NSCT domain with a lowfrequency sub-band, and several high-frequency sub-bands. Secondly, linear transformation is adopted for the coefficients of the lowfrequency sub-band. An adaptive thresholding method is used for denoising the coefficients of the high-frequency sub-bands. Then, all sub-bands were reconstructed into spatial domains using the inverse transformation of NSCT. Finally, unsharp masking was used to enhance the details of the reconstructed image. The results of experiment show that the proposed method is superior to other methods in image entropy, EME, and PSNR.