一种基于时频分布的信号检测的新方法

A New Algorithm of Signal Detection Based on Time-frequency Distribution

针对在低信噪比的情况下检测接收信号中是否含有目标信号,提出了结合数学形态学滤波的新方法。首先将接收到的信号变换到时频域,然后设阈值将时频图转化为二值图像,再经过形态学滤波提取有用的几何目标,接着统计目标点的总能量,最后与蒙特卡洛实验所得的门限进行比较,从而检测出是否含有目标信号。仿真结果表明,对于在时频平面具有比较规则的图像特征的信号,该方法能够有效的检测出低信噪比条件下接收信号中是否存在目标信号。

A new algorithm based on time-frequency （TF） is presented to detect the target signal in the received signal by morphology filtering from low signal to noise ratio （SNR） environments. First, transforms the received signal form time domain to time- frequency plane, and then thresholding process is operated on the TF image to convert the complex valued image to a binary one. Uses morphological filtering to obtain the structure, and then statistics the total energy of the point of target. At last, statistic value is compared with the threshold, which is obtained from the Monte-Carlo tests, so detects if it has target signal. The result of experiments shows that the proposed method is effective to detect the target signal with special morphological structure in the TF image in low SNR.