基于颜色特征的草莓成熟度识别技术研究  

Study on recognition technology of strawberry maturity based on color feature

作　　者：赵玲[1] 周桂红[1] ZHAO Ling ZHOU Gui-hong(College of Information Science ＆ Technology, Agricultural University of Hebei, Baoding 071001, China)

机构地区：[1]河北农业大学信息科学与技术学院,河北保定071001

出　　处：《河北农业大学学报》2017年第2期97-101,共5页Journal of Hebei Agricultural University

基　　金：河北农业大学理工基金项目(LG201406)

摘　　要：草莓成熟度识别技术是草莓采摘机器人顺利完成采摘任务的关键。本研究以"丰香"草莓为研究对象,首先提取不同成熟度草莓的颜色特征,计算草莓图像在HIS颜色空间模型下H分量的均值和方差;然后计算草莓果实红色部分着色面积所占比例;最后,建立BP神经网络,以H分量的均值、方差和草莓红色着色面积比作为BP神经网络的输入,训练完成后,判断待测样本的成熟度等级。通过本研究实验方法划分成熟度等级与经验丰富的果农判断结果对比,实验正确率可以达到90%以上。For the strawberry harvesting robot, identification technology of strawberry maturi- ty is one of the key technologies. In this paper,＂Feng Xiang＂strawberry was taken as the re- search object ,The color characteristics of strawberry with different maturity degrees were ex- tracted,and the mean and variance of the H component were calculated in the HIS color space model. Then the proportion of the red part of the color area was calculated. Finally, the BP neural network was established,taking the mean value,variance of H component and red col- oration area ratio of strawberry as the inputs of BP neural network. After the training process being completed, the maturity grade of the tested sample was validated. Compared with the judging result of experienced strawberry farmers, the experimental accuracy is above 90%.

关 键 词：草莓 成熟度 颜色特征提取 着色面积 BP神经网络 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]