|
|||||
|
|
| RGB与HSI色彩空间下预测叶绿素相对含量的研究 | |
| 引用本文： | 孙玉婷,王映龙,杨红云,周琼,孙爱珍,杨文姬.RGB与HSI色彩空间下预测叶绿素相对含量的研究[J].浙江农业学报,2018,30(10):1782. |
| 作者姓名： | 孙玉婷 王映龙 杨红云 周琼 孙爱珍 杨文姬 |
| 作者单位： | 1.江西农业大学 计算机与信息工程学院,江西 南昌330045;2.江西省高等学校农业信息技术重点实验室,江西 南昌 330045;3.江西农业大学 软件学院,江西 南昌 330045 |
| 基金项目： | 国家自然科学基金(61562039,61363041,61462038); 江西省教育厅科技项目(GJJ160374,GJJ170279) |
| 摘 要： | 为探明RGB与HSI两种色彩空间下水稻叶色图像参数与叶绿素相对含量(SPAD)之间的关系,应用支持向量机的方法预测水稻叶片的SPAD值,为快速精准获取植物SPAD值提供理论基础,同时为科学施肥提供理论指导。水稻田间试验于2015—2017年在江西农业大学农学试验站和江西省成新农场进行,供试水稻品种为金优458(JY458)、中早35(ZZ35)和两优培九(LYP9),每个水稻品种均设计4组不同的氮素水平。通过对获取的水稻图像进行叶色参数提取以及叶绿素仪测量的SPAD值来分析水稻叶色图像参数与SPAD值之间的关系,并用支持向量机的方法建立相关模型预测SPAD值。结果显示,较RGB色彩空间下三种水稻品种在HSI色彩空间上预测值的均方根误差分别减少了0.067 5(JY458)、0.020 0(ZZ35)和0.154 2(LYP9),平均相对误差比RGB色彩空间下分别减少了0.084 2(JY458)、0.133 5(ZZ35)和0.238 2百分点(LYP9)。水稻叶片在两种不同色彩空间下的叶色图像参数和水稻叶片SPAD值之间存在显著性相关(P<0.05),利用改进的网格搜索算法优化支持向量机的方法建立水稻叶片SPAD值预测模型,其预测结果误差小,为快速准确无损获取植物SPAD值的预测提供了一种新方法。 |
| 关 键 词： | 水稻 SPAD值 支持向量机 HSI RGB |
| 收稿时间： | 2018-01-23 |
Prediction of SPAD in rice leaf based on RGB and HSI color space |
|
| SUN Yuting,WANG Yinglong,YANG Hongyun,ZHOU Qiong,SUN Aizhen,YANG Wenji.Prediction of SPAD in rice leaf based on RGB and HSI color space[J].Acta Agriculturae Zhejiangensis,2018,30(10):1782. | |
| Authors: | SUN Yuting WANG Yinglong YANG Hongyun ZHOU Qiong SUN Aizhen YANG Wenji |
| Institution: | 1.School of Computer and Information Engineering, Jiangxi Agricultural University, Nanchang 330045, China; 2.Key Laboratory of Agricultural Information Technology of Colleges and Universities in Jiangxi Province, Nanchang 330045,China; 3.School of Software, Jiangxi Agricultural University, Nanchang 330045, China |
| Abstract: | The relationship between the leaf image parameters and the SPAD values of rice leaf under the RGB and HSI color spaces was studied. The method of support vector machine (SVM) was used to predict the SPAD value of rice leaf, which provided a theoretical basis for rapid and accurate acquisition of plant SPAD value by using machine vision technology, and provided theoretical guidance for scientific fertilization. The experiment was conducted at Agricultural Experiment Station of Jiangxi Agricultural University and Chengxin farm in Jiangxi Province during 2015 to 2017. And the tested rice varieties were JY458, ZZ35 and LYP9. Four different nitrogen levels were designed for each rice variety. The relationship between the color parameters of rice leaf image and the SPAD value was analyzed by extracting the leaf color parameters and measuring the SPAD value. The model to predict the SPAD value was established by using SVM. The results showed that compared with RGB, the root mean square error of the predicted values of the three rice varieties based on the HSI color space was reduced by 0.067 5 (JY458), 0.020 0 (ZZ35) and 0.154 2 (LYP9), respectively. The average relative error was lower than the RGB color space. They were reduced by 0.084 2% (JY458), 0.133 5% (ZZ35) and 0.238 2% (LYP9), respectively. There was a significant correlation between the leaf color image parameters and the SPAD values of rice under the two color spaces. By optimizing support vector machine with improved grid search algorithm, the prediction model of SPAD value of rice leaves was established. The prediction error was small, which could meet the demand of agronomic scientific research, and also provided a new method for the prediction of plant SPAD value. |
| Keywords: | rice SPAD value support vector machine HSI RGB |
| 本文献已被 CNKI 等数据库收录！ | |
| 点击此处可从《浙江农业学报》浏览原始摘要信息 | |
| 点击此处可从《浙江农业学报》下载免费的PDF全文 | |