花生种皮颜色智能识别模型的建立与应用  

Establishment and application of peanut seed coat color intelligent recognition system

作　　者：朱树良 赵昆昆 高古腔 屈成鑫 马莹莹 任锐 巩方平 李忠峰 马兴立[1] 张幸果[1] 殷冬梅[1] ZHU Shu-liang;ZHAO Kun-kun;GAO Gu-qiang;QU Cheng-xin;MA Ying-ying;REN Rui;GONG Fang-ping;LI Zhong-feng;MA Xing-li;ZHANG Xing-guo;YIN Dong-mei(College of Agriculture,Henan Agricultural University,Zhengzhou 450002,China)

机构地区：[1]河南农业大学农学院,河南郑州450002

出　　处：《中国油料作物学报》2022年第2期324-330,共7页Chinese Journal of Oil Crop Sciences

基　　金：国家自然科学基金项目(U1704232);河南省产业技术体系(S2012-05-G03)。

摘　　要：为解决花生群体种皮颜色快速鉴定难题,建立一种准确、简便和经济的种皮颜色识别标准,采用阿里云智能云计算平台颜色识别系统,测定黑、紫、粉、白等不同种皮颜色花生的HSV颜色空间(Hue,Saturation,Value colormode)数据,同时结合花生种皮花青素含量数据,构建花生种皮颜色指数(p)与种皮花青素含量(i)相关性模型,并用群体验证其准确性。研究结果表明,p和i相关系数达0.932,根据p值将花生种皮颜色划分为黑色(≥0.6000±0.0238)、紫色(0.6000±0.0238~0.4000±0.0238)、粉色(0.4000±0.0238~0.2000±0.0238)、白色(≤0.2000±0.0238)四个组别;采用AI智能种皮颜色识别模型能够快速、准确鉴定RIL群体后代的种皮颜色,具有准确性高、上下界明显等优势。In order to rapidly identified peanut seed coat color,an accurate,simple and economical seed coat color recognition system was established.In this experiment,the color recognition system of Aliyun intelligent cloud computing platform was used to determine the HSV(Hue,Saturation,Value color mode)color spatial data of peanut seed coat with different colors such as black,purple,pink and white.At the same time,combined with the anthocyanin content data of peanut seed coat,the correlation model between color index(p)and anthocyanin content(i)was constructed,and the accuracy was verified by population.The results showed that the correlation coefficient between p and i was 0.932.According to p value,peanut seed coat color was divided into four groups:black(≥0.6000±0.0238),purple(0.6000±0.0238~0.4000±0.0238),pink(0.4000±0.0238~0.2000±0.0238)and white(≤0.2000±0.0238).Using AI intelligent seed coat color recognition system can quickly and accurately identify the seed coat color of RIL population offspring,with high accuracy and obvious upper and lower bounds.

关 键 词：花生 种皮颜色 HSV颜色空间 颜色识别 

分 类 号：S565.2[农业科学—作物学]