基于正交小波变换的目标检测方法

Target Detecting Method Based on Orthogonal Wavelet Transform

摘要:

理论分析表明,独立高斯噪声经过正交小波变换后保持方差和独立性不变。基于 Mallat的小波多分辨分析,通过对小波系数进行平方律处理,建立了基于正交小波变换的恒虚警率检测器模型,推导了相应的虚警和检测概率公式,分析了噪声未知情况下小波系数序列长度对检测性能的影响,并给出了合适的长度值。实验结果表明,所提出的检测器能满足不同虚警概率和杂波背景的要求,具有较好自适应性。

Abstract:

Theoretical analysis shows that independent Gaussian noises keep the independence and variance after the or.thogonal wavelet transform. Applying a squarer to the output of the wavelet transform, an OW-CAFR (orthogonal wavelet.constant false alarm rate) detector based on Mallat algorithm was proposed. The probability equations of false alarm and de.tection were also deduced. The influence of the length of wavelet coefficients on detection performance was analyzed, andthe appropriate length was given. Finally, the experiment results of signal detection demonstrated that the proposed methodcould detect the targets under different false alarm rate and cluster background effectively, which showed the adaptability.