基于小波包变换和形态学滤波的小目标检测算法

Small Target Detection Algorithm Based on Wavelet Packet Transform and Morphological Filtering

为从红外图像中提取出微弱的小目标,提出一种小波包变换和顶帽运算相结合进行图像处理的小目标检测算法。首先利用小波包分解得到原图像不同子带的低频分量和水平、垂直、对角三个方向的高频分量,然后采用软阈值去噪方法滤除高频分量中的噪声并进行小波包重构;最后利用顶帽算法对背景进行处理并进行二值图像分割。算法充分利用小波包多尺度、多方向性特点,在大面积海空背景抑制、小目标增强方面进行了改进,在信噪较低的情况下,能够从不同背景的红外图像中有效检测小目标。

In order to improve the detection effect of small target in infrared image,a small target detection algorithm based on wavelet packet transform and top hat operation is proposed.Firstly,the low frequency components of different subbands and the high frequency components of horizontal,vertical and diagonal directions of the original image are obtained by wavelet packet decomposition.Then,the noise in the high frequency components is filtered by soft threshold denoising method and the wavelet packet reconstruction is carried out.Finally,the top hat algorithm in morphological filtering is used to remove the background and segment the image.The algorithm takes full advantage of the multi-scale and multi-direction characteristics of wavelet packet,improves the large area sea and air background suppression and small target enhancement,and can effectively detect small targets from infrared images with different backgrounds under the condition of low signal-noise.