关键词：
单芽段;蔗种;机器视觉;甘蔗茎节;切割位置
摘 要：
为实现甘蔗单芽段蔗种的自动切割,利用机器视觉技术识别甘蔗茎节。对甘蔗图像进行均值滤波处理,以处理后的甘蔗图像为基础,在其HSV颜色空间的H分量上进行阈值分割、数学形态学处理以及最大面积选择,提取中心坐标;再利用矩形模板在G-B色差分量图上以一定步长沿甘蔗中心横坐标移动,计算每一步所覆盖甘蔗的平均灰度值,找到最大平均灰度值对应的位置,最终确定甘蔗茎节位置。试验结果表明,在36组组合试验下,较优组合为:模板宽度为6、步长为6,其识别率为90.77%,平均时间为0.481 539s。
译 名：
Recognition of sugarcane nodes based on local mean
作 者：
Huang Yiqi;Huang Tisen;Huang Meizhang;Yin Kai;Wang Xiaobo;College of Mechanical Engineering,Guangxi University;
关键词：
single bud period;;sugarcane seeds;;machine vision;;sugarcane nodes;;cutting position
摘 要：
In order to realize the automatic cutting of sugarcane seeds in single bud period,machine vision technology was used to recognize the sugarcane nodes.Based on the image which had been processed with mean filter,threshold segmentation,mathematical morphology processing and maximum area selection were conducted on H component of HSV color space,and then,center coordinates were extracted.Average grey values under the cover of every step were calculated,using rectangular template moves on the GB component color chart with fixed step size,the corresponding position of maximum average grey value was found to finally determine the position of sugarcane nodes.Test results show that the better combination within 36 testing groups is template width is 6,step size is 6,recognition rate is 90.77% and average time is 0.481 539 s.
