中文摘要：

水下声信号处理中通常采用匹配滤波器作为线性调频回波的最优检测器，并且根据匹配滤波器输出的峰值位置估计目标距离，但对于径向速度未知的动目标线性调频回波信号，多普勒频偏造成的回波和样本失配将导致匹配滤波器检测性能下降。在Dechirp算法基础上提出了可以用于混响噪声背景下动目标线性调频回波检测的累积归一化算法，仿真测试表明，该算法在混响噪声背景下具有良好的检测性能，并且根据判决向量的峰值位置可以估计目标径向速度。

英文摘要：

Matched filter is used as the optimal detector of LFM （Linear Frequency-Modulated） echo in the domain of underwater acoustic signal processing, and the peak position of the matched filter＇s output can be used in the estimation of target range. But due to the LFM echo of moving target whose radial velocity is unknown, the mismatch of echo and replica caused by Doppler frequency offset will degrade the detection performance of matched filter. Based on the dechirp algorithm, the cumulative normalization algorithm which can be used in the detection of LFM echo under the reverberation noise background is proposed. The simulating test proves this algorithm has well detection performance under the reverberation noise background, and the peak position of the decision can be used in the estimation of target radial velocity.