提出了多结构形态学滤波算法,对小目标图像进行背景抑制,接着引入信息熵进行分割,从缩小阈值搜索范围和推导递推关系式两方面给出快速分割算法。通过不同滤波算法进行背景抑制,并采用经典的阈值分割方法和基于信息熵的分割方法对形态学处理后的图像进行分割,对实验结果进行详细分析,表明本文多结构形态学滤波算法对图像的信噪比改善最明显,分割算法处理后的图像含噪声较少且目标分割完整,是一种有效、稳定的目标检测方法。
A multi-structure morphological filter is proposed for IR image processing by which the background of small-target image can be effectively suppressed.Then the information entropy theory is introduced for image segmentation and the rapid segmentation algorithm is given out.Background suppression is done with different filtering algorithms and image segmentation is done through morphological filtering,by adopting the classical segmentation algorithm and the information entropy algorithm.It′s shown by simulation that the SNR of the collected images are improved considerably and the segmentation results contain little noise and the targets are integral.Analysis is made to prove the effectiveness and stabilization of the proposed method.