中倍镜下弱目标分割算子耦合主动视觉的自动显微图像排查机制

针对中倍镜下微生物图像背景复杂,目标微弱,不易分割等难题,提出了弱目标分割算子;并引入主动视觉,设计了二者相融合的微生物自动显微图像排查机制。根据微生物的形态和颜色特征,建立样本相对标准,实现基于主动视觉的弱目标分割。该方案以样本相对标准为判断依据,建立在基于HSV颜色空间中H通道增强与线性变换;再结合自适应阈值分割与形态学图像处理技术,解决弱目标难分割的问题。实验结果表明,相对于传统方法,该方案应用在微生物显微排查上可以达到较好的效果,在工程上应用是可行的。

SThe Examination Mechanism of Automatic Microscopy Image Based on Medium Power Faint Target Segmentation and Active Vision

According to the problem of complicated background of microbe image at medium power and faint target, there is difficult to segment, The faint target segmentation at medium power was proposed in this paper. According to morphological and color features of microbe, sample relative standard was established, realizing faint target segmentation based on active vision,. In this solution, judging based on sample relative standard, using H channel enhance in the HSV color-space ,then using linear conversion based on adaptive threshold segmentation and morphological image processing, to resolve the problem of difficulty segment. Experimental results show that the solution achieve good effect in microscopy for microbe, and is feasible in engineering.