摘要
针对浮选采集图像光照不均、分割难的问题,基于分数阶微分、谷底边界扫描及区域合并的思想提出了一种新的浮选气泡分割算法。该方法首先利用分数阶微分理论构建相应的掩模算子,再通过确定最优阶数,进而对浮选气泡图像的分数阶掩模增强,接着将谷底边界扫描与区域合并相结合对气泡进行分割。整个算法包括:图像预处理、分数阶微分增强、谷底边界扫描、端点检测及联结、区域合并等步骤,算法可以一次性检测出所有谷点,进而高效精准地提取气泡。与基于相似性与不连续性算法做了对比试验,试验结果表明:(1)新算法能较好的减少过分割现象,具有更好的分割精度;(2)其运行速度提高了50~100倍,满足实际生产中对实时性的要求。
This paper proposed an image segmentation algorithm for mineral froth images,based on fractional differential,valley edge detection and region merge on shape. In this algorithm,first,flotation bubble image be enhanced by construct corresponding mask operator and the optimal order. Then,flotation bubble image segmented via valley edge detection and region merge. The algorithm was tested for a number of froth images and fractional differential and valley edge detection. The accuracy and processing speed of the studied algorithm are better than that of the watershed algorithm. The test results showed that the algorithm performs satisfactorily.
出处
《南阳理工学院学报》
2015年第4期48-52,共5页
Journal of Nanyang Institute of Technology
基金
国家自然科学基金资助项目(61170147)
关键词
浮选图像
分数阶微分
谷底边界
图像分割
flotation image
fractional differential
valley-edge
image segmentation
