雷达数据的准无损压缩

Quasi-Lossless Compression of Radar Data

利用不同周期雷达信号之间的相关性,采用二维小波变换,融合信号去噪和数据压缩算法,实现雷达数据的准无损压缩.针对雷达数据噪声较多的特点,在压缩之前通过软阈值滤波去除雷达信号中的大部分噪声,大大降低数据熵值.为评价重建信号质量,设计了形态逼真度指标.实验结果表明,该方法能对航空探测雷达回波数据进行较大比率压缩.

A compound algorithm combining signal de-noising with data compression is presented. Traditional compression methods use some prediction scheme to de-correlate the one-dimensional signal. Due to the large amount of noise, these methods, however, are not so efficient when used to radar data. For both signal de-noising and data compression, wavelet is a good choice. Also the algorithms for both purposes are very similar in the structure. To reduce the entropy of the data, a soft-thresholding is applied to de-noising the signal first, so that a quasi-lossless compression can be implemented. Experiments show that this algorithm demonstrates good performance in radar data compression. In addition, a new index named Morphological Fidelity is proposed to measure the morphological distortion of the reconstructed signal.