| Popis: |
For poor target segmentation effect, the fewer detail features of underwater target are always lost in Fourier transform. According to amounts of analyzing statistical, a new underwater target detection method based on wavelet transform is proposed in the paper. First of all, multi-resolution processing of Haar wavelet transform is used to retain target features. Then, fitting an exponential function to confirm the initial smoothing interval and calculating the binary processing threshold through iteration. Thirdly, a statistical eight neighborhood grey information method is used to restrain the interferences around the targets. Finally, according to the above threshold ,the small man-made target could be detected from the background. The processing result is satisfied after a median filtering. |
