细胞位姿视觉识别的研究-《计算机技术与发展》

文章信息/Info

Author(s):: YUAN Ling; DU Qi-Hang; College of Automation Science and Technology, South China University of Technology

摘要:: 由于环境的非结构化，视觉引导和视觉伺服技术是实现细胞自动注射的重要途径。其中细胞位置和姿态的识别是该技术应用中的重要环节。针对卵细胞浆内精子注射（ICSI）的操作，提出了一种显微镜下细胞位姿的计算方法。首先根据二值图像的区域几何特征，检测出细胞区域，确定细胞的中心位置，然后依据倒置式光学显微镜的成像特点，通过图像中细胞膜区域的定向区域灰度特征，判断细胞的俯仰角是否为零。对俯仰角不为零的情况，通过Hough圆检测识别出细胞膜内轮廓包围的区域中的极体，最后通过空间几何关系计算出细胞的姿态角。实验结果表明，该方法能有效地计算出ICSI过程中细胞位姿调整阶段中的细胞位置和姿态信息

Abstract:: Vision-guided and visual servoing are important solutions to automatic cell injection due to the non-structured environment,in which the identification of the position and orientation of a cell acts as a vital part. Present a novel calculating method to work out this problem in an inl~a-cytoplasmic sperm injection（ICSI） process. The cell＇ s area and centroid were detected by the geometric feature of regions in the binary image,and the elevation was determined whether to be zero by directional gray-level feature in the cell membrane, according to the imaging characteristic of an inverted light microscope. A hough transform for circle detection was carried out in the next step to inspect the polar body in closed region enclosed by the inner contour of the membrane when the elevation was judged non-zero. Attitude angles were calculated according to three-dimensional geometry eventually. An experiment was carried out with detailed descrip- tion to the results, which proved the effectiveness of the proposed method