脉冲压缩定点处理的量化噪声分析

Quantization Noise Analysis of Fixed-Point Pulse Compression

针对合成孔径雷达成像算法的定点化处理,提出了数字脉冲压缩定点化处理量化噪声的数学模型,采用线性时不变系统的性质计算脉冲压缩定点运算的量化误差.给出了该模块的舍入量化噪声详细分析过程,并推导出系统输出噪信比与系统字长、FFT长度等参数之间的关系.对获取的回波数据采用不同的系统字长进行成像,仿真结果很好地证明了所述理论的正确性.这样可根据理论结果选择合适的位长,用定点处理代替浮点运算处理减少运算的复杂度及资源的占用.

To realize the fixed-point processing of synthetic aperture radar, a model of quantization noise due to the fixed-point processing is proposed. By using the property of linear time invariance system, a quantization noise model is built and detailed analysis of rounding quantization noise is given. Also the relations between noise-to-signal ratio and word length and the length N of FFT are respectively deduced. The result of imaging of radar echo data by different word length shows that the proposed method is correct. Thus suitable word length can be chosen to replace float point computing which can reduce the computing complexity and the cost of resource.