Abstract: In order to realize the accurate identification of safflower filaments and locate the center point, a method of identifying the coordinates of the filaments and marking the two dimensional center points is proposed based on the MATALB processing software. According to the color characteristics of the filaments, the color difference method is used to compare and analyze. The method of maximum inter class variance is used to extract the contour of the filaments. An improved morphological processing method is proposed to transform the contour of the filament based on the cluster characteristics of safflower. In the end, a two dimensional center point coordinate is obtained by combining the weighted centroid method and the maximum internal tangential circle method to improve the accuracy. The experimental results show that the method can effectively identify and extract two dimensional center points of filaments images with different maturity, and the accuracy rate is 100%, especially the opening of the 2-3 day safflowers.

Key words: safflower filaments, color characteristics, morphological processing, two dimensional center point

摘要： 为实现红花采摘机器人准确识别红花花丝并进行中心点定位，提出了一种识别红花花丝并标记二维中心点坐标的方法。根据花丝的颜色特征，运用色差法进行对比分析；选用最大类间方差法进行花丝的轮廓提取；根据红花丝团簇状特征提出了一种改进形态学处理方法进行轮廓变换，最后提出一种结合改进质心法和最大内切圆法获取花丝二维中心点坐标的方法，提高其准确度。实验结果表明，该方法可对不同成熟度的花丝图像进行有效识别并提取二维中心点，尤其对开放2~3天的红花，准确率达100%。

关键词: 红花花丝, 颜色特征, 形态学处理, 二维中心点