摘要
提出一种先对图像进行有效二值化增强,再进行带约束的边缘检测算法。该边缘检测算法设计了百分比滤波器,得到二值图的进一步平滑去噪图,再求取原始图像标准方差图和Canny算子二值图。叠加以上3幅图,移除面积小于阈值的连通区域后,利用改进的Canny算子检测出线粒体边缘。对多幅线粒体电镜图像分别采用经典算子和本文方法检测线粒体边缘。实验结果表明,该算法明显优于经典算子,能有效识别出多数线粒体边缘。
In this paper, an edge detection algorithm is proposed, whose first step is efficient two-val- ue enhancement of images, and the second one is the edge detection algorithm with constraints. The 1 st step of edge detection algorithm is designing a percentage filter to get the further smoothing and de- noising image. The 2nd step is seeking the standard deviation image and the two-value image by using Canny operator of the original image. The 3rd step is overlaying the three images. The 4th step is re- moving the connected region whose area is less than the threshold value. The last step is detecting the edge of mitochondrial by improved Canny operator. The edges of many mitochondrial electron micro- scope images are detected with classical operator and the designed enhancement algorithm proposed in the paper respectively. The experimental results show that the algorithm proposed in the paper is supe- rior to classical operators. It can identify the majority of mitochondria edges effectively.
出处
《重庆理工大学学报（自然科学）》
CAS
2012年第5期94-100,122,共8页
Journal of Chongqing University of Technology：Natural Science
基金
福建省教育厅B类科技项目(JB11190)
福建江夏学院院级科研项目(2011C006)
关键词
边缘检测
二值化
图像增强
线粒体
edge detection
binarization
image enhancement
mitochondria
