论文部分内容阅读
研究了水声图像高速传输信号处理方法,它包括两个方面,一方面是水声相干通信信号处理方法,其中:(1)多普勒频移补偿,在数据包的前后两端插入已知线性调频(Chirp)信号,拷贝相关后求互相关,估计相对多普勒平均频移。在自适应判决反馈均衡器中加上自适应相位补偿器,采用快速自优化最小均方(LMS)算法,与其对应的速度容限优于常用的二阶锁相环相位补偿器的。两种补偿方法联合工作时,性能优良。(2)带有分集合并器的自适应判决反馈均衡器的算法是快速自优化的LMS算法,计算量小,性能优良。(3)自适应判决反馈均衡器与Turbo-网格编码调制(TCM)译码器级连、迭代算法。研究了基于软输出维特比(SOVA)方法的新型的比特-符号转换器,用它时误比特率(BER)比常规编码、映射方法的近似小2个数量级。另一方面是抗误码的图像压缩方法。本文基于数字小波变换和定长编码方法,研究了声图像的压缩。它包括:(1)选用CDF9/7小波进行小波变换。(2)对小波系数子带能量进行统计分析,三层小波分解是合适的。(3)对不同能量的子带采用不同的量化步长。(4)采用定长编码算法。结果表明声图像压缩比特率为0.85。当BER小于10~(-3)时,图像质量完好。当BER小于10~(-2)时,图像中出现少量小黑白点。在上述基础上研制了水声通信机,频带为(7.5～12.5)kHz,接收声呐阵为8基元等距线阵,信号为QPSK和8PSK。在中国千岛湖进行了湖试,采用SOVA硬迭代算法,达到了低BER。传输一幅256×256×8的声图需时约7s。传输距离与传输速率之积为55 km kbps。
The method of signal processing for high speed transmission of underwater acoustic images is studied. It includes two aspects, one is underwater acoustic coherent communication signal processing method, in which: (1) Doppler frequency shift compensation, insert known before and after the data packet Chirp signals are correlated with each other after the copy is correlated, and the relative Doppler average frequency shift is estimated. In the adaptive decision feedback equalizer with adaptive phase compensator, using fast self-optimizing least mean square (LMS) algorithm, its corresponding speed margin is better than the commonly used second-order phase-locked loop phase compensator. The two compensation methods work together, excellent performance. (2) The adaptive decision feedback equalizer with diversity combiner is a fast self-optimizing LMS algorithm, which has a small amount of computation and good performance. (3) adaptive decision feedback equalizer and Turbo-Trellis coded modulation (TCM) cascade, iterative algorithm. A new type of bit-to-symbol converter based on Soft-Output Viterbi Ratio (SOVA) method is studied. The bit error rate (BER) is 2 orders of magnitude smaller than that of conventional coding and mapping methods. On the other hand, it is an error-resistant image compression method. In this paper, based on digital wavelet transform and fixed-length coding method, the compression of acoustic image is studied. It includes: (1) Selection of wavelet transform CDF9 / 7. (2) Statistical analysis of wavelet coefficient sub-band energy, three-layer wavelet decomposition is appropriate. (3) Different quantization steps for different energy subbands. (4) using fixed-length encoding algorithm. The result shows that the audio image compression bit rate is 0.85. When the BER is less than 10 ~ (-3), the image quality is intact. When the BER is less than 10 ~ (-2), a small number of small black and white dots appear in the image. On the basis of the above developed underwater acoustic communication machine, the band is (7.5 ~ 12.5) kHz, the receiving sonar array is 8 yuan equidistant linear array, the signal is QPSK and 8PSK. A lake test was conducted at Thousand Island Lake in China, using the SOVA hard iterative algorithm to achieve a low BER. Transmission of a 256 × 256 × 8 sonogram takes about 7s. The product of transmission distance and transmission rate is 55 km kbps.