摘要：

应用属性形态学的属性准则级联与组合处理方法对模糊边界目标图像进行滤波,研究面积、面积亮度对比、形状、拓扑高度和后代数等属性形态学滤波原理及滤波适应性。通过最大树节点属性计数方式对模糊边界细胞进行精确统计,利用基于连通域灰度层次的拓扑分布节点属性描述和修剪方法,实现模糊边界分割和粘连目标识别,并采用柔性结构元素对复杂边界图像进行规约性滤波。实验结果表明,与现有目标识别方法相比,该方法能有效排除模糊目标图像中非感兴趣区域的背景干扰,提高对各类细胞的识别精度。

关键词: 属性形态学滤波, 面积亮度对比滤波, 矢量属性, 面积属性, 细胞图像

Abstract:

This paper researches filtering method to identify objects with fuzzy bordary based on cascading and integrated attribute morphology filtering.It focuses on principle of morphological attribute filtering such as the area,area brightness contrast,shape,topological height and descendants,and analyzes theirs filtering adaption.It proposes an counting statistics mode by the node attributes of a maximum tree.It concretely probes the connected region topological distribution node attribute description and purrning methods according to theirs gray level.It is suitable for recognizing overlapped objects and object segmentation with different types of fuzzy boundary,and implements efficient and disciplinary filtering for complex boundary image by flexible structural elements.The experimental result shows that the proposed method clears away area of disinterest background interferences in a fuzzy image efficiently compared with previous object recognition methods,and it promotes the recognition accuracy rate for various cell images.

Key words: attribute morphological filtering, area brightness contrast filtering, vector attribute, area attribute, cell image

中图分类号: 

• TP391.41