摘要： 针对图像分析中区域填充算法的自动化和通用性要求，对种子填充算法进行改进，提出了反向注入式种子填充算法。与常规的区域填充算法相比，该算法的特点是包括初始种子点在内的所有填充区域均完全由计算机自主、高效地处理，实现了任意复杂区域填充的普遍适用性，解决了扫描线算法和种子填充算法的不足。该算法能一次完成包含多个区域的填充问题，在处理多而密集区域的填充问题上具有很高的效率。

关键词: 区域填充, 多边形填充, 图像分析, 复杂边界

Abstract:

An Injected Conversely Filling(ICF) method is presented in this work to meet two general problems for area-fill methods: full automation and general applicability. Compared with other methods, an excellent advantage of ICF method is able to fill any complicated area easily and automatically, including initial seeds finding. Besides these, as ICF has ability to fill all the separated blocks in a same area at a time, it has a great virtue on multi area filling.

Key words: area filling, polygon filling, image analysis, complicated contour

中图分类号: 

• TN919.81