Synthetic Aperture Radar Raw Data And Practical Compression Algorithm

As a high-resolution microwave remote sensing instrument, the Synthetic Aperture Radar (SAR) has been more and more widely used. The data compression is one of the most important digital signal processing stages of the SAR.The traditional compression algorithm is the Block Adaptive Quantization (BAQ). The theoretic foundation of it is the distribution of the raw SAR data. But in fact, the raw data is NOT Gauss distributed, especially when there is some saturation with the receiver. In this case, the bad effect to the BAQ can't be neglected.Aiming at the problem of the BAQ in the partly saturation SAR data, a model is constructed according to the SAR equipments. Before the signals enter the receiver, they are Gauss distribution indeed. But when they go through the uniform quantizer, the distribution has changed. But we can find the distribution by the mean value of the data absolute value. Then we can find the quantization levels and the reconstruct levels of the BAQ. These floating levels are much more consistent with the data distribution, so a certain higher SQNR (Signal to Quantization Noise Ratio) is obtained.In common signal processing stage, if the input signals exceed the limit of the uniform quantizer, the output signals become the limit level, the minimum of the signals. To the signals which distribution is unknown, this method is certain. But the original data of SAR is doubtless and we can calculate it by the mean value of the uniform quantized data. Then better output signal levels are given.There are many kinds of compression algorithms, from time domain to frequency domain, from scalar to vector. The effect is better and better. But on the other hand, only a few algorithms can be applied into practice. Because the data rate is very high along with the widely usage of SAR, the algorithm must be fast and simple. The compression quality is not so important as the rate. So the high-quality algorithms are hard to apply.Vector Quantization is an effective data compression technology. Even there are many high-speed algorithm of it, but it is not so fast to apply to SAR. Focus on the search speed of the vector quantization, a very high speed algorithm is presented. The table looking up and the table structure is the kernel of this algorithm. Based on the algorithm,a data compression scheme is constructed. Also some other techniques are bound to it to supply a high speed.The high-quality algorithms have been discussed frequently but practical algorithms are still very limited. So this thesis puts the emphases upon the factual problems.