基于机器视觉的组织培养苗自动识别技术

为了使组织培养苗分割移植机器人的视觉系统能够引导机器人实现条状组培苗的自动分割、移植作业，提出了对条状组培苗的图像进行灰度拉伸、中值滤波、最大方差自动取阈值、二值图像的闭运算、细线化、苗节点搜索及切割点定位的自动识别算法。对组培车间普通照明下拍摄的200幅日本黄杨苗(条状组培苗)图像的识别处理结果表明，该算法能够识别出苗的节点，计算出切割点的位置，成功率为91％，且该算法不受图像大小和样本摆放方位的影响，适应性强。

Application of Machine Vision in Automatically Discriminating Tubular Tissue Culture Plantlet

The objective of this research is to develop a automatic discrimination algorithm to find the location of tubular tissue culture plantlet nodes,and then compute the cut location,so the vision system of tissue culture plantlet robot can direct the robot to cut and replant tubular tissue culture plantlets according to the cut location. The algorithm includes gray stretching, median filtering, automatically thresholding, close operation, thinning, node searching, and cut location computing and so on. Experiments performed on Japan boxwood plantlets which are tubular tissue culture plantlets images taken from workshop showed that plantlet nodes were discriminated and cut locations were computed with 91% success ratio in all images of 200 sets of plantlets, and the algorithm has strong adaptability because it wasn't affected by images size and orientation of plantlets.