Low complexity wavelet packet compression of bayer images for communication over wireless sensor networks
The purpose of this thesis was to develop a low complexity image compression algorithm for Wireless Sensor Networks (WSN). The main constraints in a WSN node are limited processing power, memory allocation, and battery life. The goal of this thesis was to develop an image compression algorithm that would thrive under this constraint environment. The main focus was to minimize the complexity of the algorithm by taking advantage of the spatial correlation found in Bayer images. The algorithm employs line based 1-D Discrete Wavelet Packets Transform of Bayer images. In addition the image compression method includes the application of a third-order Differential Pulse Coded Modulation, and entropy coding. The generation of the prediction values for a third order DPCM were generated based on statistical data of a set of six Kodak true color images. In addition, the probability density function of every subband was obtained and used to design the Huffman tables used for entropy coding. The proposed algorithm was able to produce comparable PSNR and compression ratio compared to a direct implementation of the JPEG algorithm to a Bayer image. In addition it was found that applying the JPEG algorithm with a very low quality factor does not preserve the color information in the image.
Calzada, Ivan, "Low complexity wavelet packet compression of bayer images for communication over wireless sensor networks" (2010). ETD Collection for University of Texas, El Paso. AAI1483854.