节点文献
基于区域的图像识别算法研究
Research on Region-Based Image Recognition Algorithms
【摘要】 提出一种基于"分水岭"区域分割方法的图像识别算法。针对分水岭算法"过分割"的问题,进行小图像区域消解和有选择的输出关键区域,提出3种区域选择方法——均值方差法、主灰度法、双值法。实验表明,算法对于背景不太复杂的图像识别效果较好。
【Abstract】 In this paper,an image recognition algorithm based on watershed regional segmentation method is proposed.To restrain the over-segmentation which results from watershed method,small regional degradation and selective output of the key regions are adopted.Three regional selection methods are used,which are named as Mean-Variance Method,Dominant-Gray Method and Two-Value Method.The simulation results demonstrate that the proposed method is efficient for the less complicated image recognition.
- 【文献出处】 电脑知识与技术 ,Computer Knowledge and Technology , 编辑部邮箱 ,2010年29期
- 【分类号】TP391.41
- 【下载频次】204