An efficient image denoising algorithm based on double density dual tree discrete wavelet transform for wireless sensor network

Generally, the input images are unavoidably corrupted by Gaussian noise during process of sensing, transmitting and retrieval of images over Wireless Sensor Network (WSN). To suppress the noise and enhance the input image quality, the wavelet based image denoising methods has shown better results in the field of WSN. However, these methods affect the quality of the denoised image due to the poor selection of thresholding technique and the number of decomposition levels. In order to overcome the above mentioned problems and to reduce the impact of Gaussian noise over WSN images, this research work concentrated on the hybrid of double density wavelet and DTCWT based wavelet called Double Density Dual Tree Discrete Wavelet Transform(DDDTDWT). The proposed work is discussed in form of two parts. The first part explains about the simple DDDTDWT based image denoising technique. The second part describes about the proposed DDDTDWT with the combination of Fast Non Local Means Filter (FNLMF). Further, to verify the effectiveness of the proposed image denoising algorithms, two thresholding methods such as hard thresholding using median absolute deviation and bivariate thresholding using adaptive method are utilized. Furthermore, the performance comparison of the existing and the proposed image denoising methods developed for WSN are examined through the simulation results using MATLAB.