复杂背景下扩展目标的分割算法研究

综合利用目标边缘、区域和纹理等多种特征，提出一种新的复杂背景下扩展目标的分割算法。在基于知识的串行边界初分割基础上，运用一种 新的分形分割方法抑制初分割图像中的自然背景团块，最后利用数学形态运算消除背景粘连。实验结果表明该方法能够有效的分割复杂背景下的扩展目标，并能较好地保留目标的形状特征。

SEGMENTATION OF EXTENDED TARGET IN COMPLEX ENVIRONMENT

A new segmentation algorithm, which utilizes boundary, area and texture features for segmenting extended target in complex environment, was proposed. the target was segmented preliminarily by series boundary technique based on knowledge, Then a new fractal segmentation algorithm was used to suppress the complex nature background blocks in the image. And the mathematical morphology method was utilized for eliminating the background conglutination. The experimental results indicated the validity of the method for segmenting the extended target in complex environment and that the method can reserve the shape details of the target perfectly.