使用基于Canny理论的算法自动识别及筛选菌落

在微生物研究中，经常需要大量挑选重组子或突变子，工作量十分繁重。本文基于Canny算子，提出了一种算法用以自动识别及筛选菌落，从而减轻这一操作的工作量。该算法能够直接处理彩色图像，充分利用彩色图像信息，能很好的检测彩色图像边缘，从而自动识别菌落。算法既继承了传统Canny算法优点，又针对检测差异菌落图像这一实际目的做了优化。除能进行菌落图像边缘检测外，本算法还使用边缘曲线半径、长短轴比、像素半径比等菌落图像描述参数来表述菌落，以此来区别不同菌落。结果表明该算法能准确识别菌落，对特征不同的菌落，区分效果良好。结合算法开发相应仪器，可以自动化识别并筛选菌落，在需要进行大量的菌落挑选时，该方法具有一定实用价值。

Colonies Detection and Screening with Canny Based Algorithm

Large amount of mutons or recons screening therefore present a Canny operator based algorithm is quite difficult for microbiology research. We to automatically detect and screen different microbial colonies. The algorithm can directly process color images and well detect color edges with full color information. We also optimize microbial colony screening. The algorithm can also employ such image parameters as radius and ratio of long axis and short axis length to describe colony images. Results show that the algorithm can well detect and distinguish colonies of different characters. We can develop an intelligent device with the algorithm, which can automatically screen microbial colony. The algorithm is therefore a very useful tremendous colony screening approach.