Image Compression Of Synthetic Aperture Radar In Wavelet Domain

As a high-resolution microwave imaging radar which isn't influenced by terrain, time and weather etc., Synthetic Aperture Radar (SAR) takes an important place in the field of microwave remote sensing. SAR images are usually big and contain a large amount of data. So it is a key problem in SAR image data processing that how to compress the image to reduce data amount effectively so that it can be saved or transmitted conveniently.Compared with general image, SAR image has such properties: degraded by speckle noise, has low correlations between pixels; includes abundant texture information and a great deal of point targets; different users care about different content and objective. So SAR image has lower compression ratio than general optic image, and users have different requests for image quality. Researching on compression methods of SAR image, we should consider not only compression ratio but also fidelity.Because of above properties, some good compression methods for optic image such as JPEG based on Discrete Cosine Transform (DCT), didn't achieve satisfied effects. In recent 10 years, Wavelet Transform (WT) is widely used in the field of image compression. The essential of WT is multiresolution analysis, and WT has good local characters both in time domain and frequency domain. So using WT to compress SAR image is a hotspot in the field of SAR image processing. Here we make an attempt.The highlight contents as follows:1) Theories and methods of image compression, especially widely used DCT and international image compression standard-JPEG based on DCT are investigated.2) Wavelet theories and fast algorithm for image compression by WT-Mallat algorithm, JPEG2000 based on Discrete Wavelet Transform (DWT) are discussed.3) The choice of wavelet bases is discussed. Spatial-Orientation Tree (SOT) in wavelet domain and image compression algorithms based on SOT include EZW and SPIHT are studied. Speckle noise and compression of SAR image supplement each other, removal of speckle noise can enhance correlations between pixels and compressibility of SAR image. The generative mechanism and conventional reduction methods of speckle noise, soft threshold and hard threshold methods in wavelet domain are investigated.4) A SOT structure based method for SAR image compression is proposed, which integrated speckle noise removal and EZW algorithm. Results of compression of large numbers of Airborne SAR images validate the proposed method is efficient and better than JPEG and EZW algorithm, which reduced the infection of speckle noise and improved compressibility of SAR image.