Optimal Entropy Coding Method for Lloyd-max Quantizer

Abstract

The invention discloses an optimal entropy encoding method for a Lloyd-Max quantizer. The method comprises the steps of according to a probability density function of raw data, calculating probabilitydistribution of an output electrical level; according to the bit of the Lloyd-Max quantizer, choosing to expand Huffman coding or to calculate a Huffman code combined with the Huffman coding and theprobability distribution of the output electrical level, and obtaining Huffman code libraries of different bits; besides, utilizing the Huffman code libraries of different bits to obtain the Huffman code of the output electrical level. According to the method, it is unnecessary to conduct statistical pretreatment on a quantitative electrical level and can encode the quantified data in real time. Compared with a fixed-length code, the encoding efficiency is improved, and at the same time, the encoding efficiency of a low-bit quantizer output electrical level is further improved; compared with traditional quantizer fixed-length encoding, the fixed-length code only needs to be replaced by the Huffman code, hardware does not need to be modified, and surplus calculation quantity is also not needed.

Description

technical field [0001] The disclosure belongs to the field of synthetic aperture radar data compression, and relates to an optimal entropy coding method for a Lloyd-Max quantizer. Background technique [0002] Synthetic aperture radar is an all-weather and all-weather radar for telemetry on the surface of the planet. Due to the application of synthetic aperture technology, the radar can obtain high-resolution images. At the same time, due to the high-resolution characteristics of radar, the amount of radar data has increased significantly, which conflicts with the limited data storage capacity and effective downlink bandwidth. Therefore, effective compression should be carried out before the raw radar echo data is downloaded. [0003] Aiming at the problem of SAR raw data compression, the FBAQ (Flexible Block Adaptive Quantization) algorithm is commonly used in engineering at present. The FBAQ algorithm has the advantages of simple hardware implementation and low computati...

AI Technical Summary

Problems solved by technology

However, the probability and statistical characteristics of the vast majority of SAR raw data approximately obey the Gaussian distribution. After encoding with a fixed-length code, the probability distribution of the quantized output level is uneven, and its entropy value is smaller than the code length of the fixed-length code. The coding redundancy, the coding efficiency is low

[0005] In theoretical analysis, in view of the probability distribution of output levels, some literature proposes that different output levels can be counted first, and the proportion of different output levels can be obtained as its probability distribution before Huffman coding (Hoffman Coding). This method requires statistical preprocessing of the quantized data, which not only brings a large encoding delay, loses the real-time encoding ability of the output level, but also needs to make a reasonable setting for the statistical output level, which increases the The amount of calculation and hardware overhead is difficult to apply in engineering

Images