基于分光测色反演哈密瓜叶片叶绿素与含水率研究

doi:10.3964/j.issn.1000-0593(2024)08-2296-07

摘要
参考文献
相关文章 (15)

全文: PDF (5521 KB)
输出: BibTeX | EndNote (RIS)

摘要：为了快速、精准地预测哈密瓜叶片含水率和叶绿素含量，提高对哈密瓜作物精准管理水平，采用分光测色技术，选用生长期、开花期和结果期三种不同生长阶段的哈密瓜叶片作为实验对象，分别研究9:00—10:00、14:00—15:00和20:00—21:00三个不同采集时间段的叶片温度、叶片含水率和叶绿素含量与色彩空间CIELAB特征值的相关性变化。采用最小二乘法（LS）对不同样本叶片的温度、含水率和叶绿素含量与颜色特征值存在的变化规律进行数据预处理，选用拟合度最优的特征值进行回归分析和预测模型验证。结果表明：①叶片温度、叶片含水率和叶绿素含量在不同参数情况下，都会使叶片存在不同的颜色特征值变化。②含水率为84%~93%的叶片，其温度和叶绿素含量都与叶片含水率呈负相关性。③叶绿素含量和叶片含水率与色彩空间CIELAB存在一定的线性相关，随着叶片含水率的上升，L呈上升趋势，表现为叶片颜色逐渐变浅呈淡绿色，随着叶绿素的增加，L呈下降趋势，表现为叶片颜色逐渐加深呈黑绿色，所有的样本数据均存在L、B为正值。④通过模型预测与评估，选用随机森林（RF）、偏最小二乘法（PLS）、支持向量机（SVM）和LASSO都能很好地对叶绿素含量进行有效的预测。在叶绿素预测模型中，RF预测性能最好，R²_c=0.939、RMSE_C=0.868和MAE=0.686，R²_p=0.915、RMSE_P=1.194和MAE=0.942。⑤通过模型预测与评估，选用随机森林(RF)、偏最小二乘(PLS)、自适应增强与套路回归(ADABOOST)和多项式回归（POLYNOMIAL）都能很好地对叶片含水率进行有效的预测。在叶片含水率预测模型中，POLYNOMIAL预测性能最好，R²_c=0.884、RMSE_C=0.005 9和MAE=0.005 2，R²_p=0.920和RMSE_P=0.006 2和MAE=0.005 7。采用分光测色的方法，能够有效、快速地测定叶片含水率与叶绿素含量，有望为无损、快速和精准地测定叶片含水率与叶绿素含量提供一种可选的可行测定方案。

关键词：哈密瓜；叶绿素；含水率；CIELAB彩色空间；模型回归

Abstract：In order to predict the water content and chlorophyll content of cantaloupe leaves quickly and accurately and improve the accurate management level of Cantaloupe crops, the leaves of cantaloupe in three different growth stages, namely the growing stage, the flowering stage, and the fruiting stage, were selected as experimental research objects by using the spectrophotometry technology. The correlation changes of leaf temperature, leaf water content, and chlorophyll content with LAB eigenvalues of color space were studied in three different collection periods: 9:00—10:00, 14:00—15:00, and 20:00—21:00, respectively. The least square method (LS) was used to preprocess the changes in temperature, water content, chlorophyll content, and color eigenvalues of different samples, and the eigenvalues with the best fit were selected for regression analysis and prediction model verification. The results showed that① Leaf temperature, leaf water content, and chlorophyll content had different color eigenvalues under different parameters. ② For leaves with 84%~93% moisture content, leaf temperature, and chlorophyll content were negatively correlated with leaf moisture content. ③ the chlorophyll content and leaf water content and color space LAB, there is a linear correlation. As the leaf water content -rises, L is on the rise, and the color becomes shallow gradually with light green leaves; with the increase of chlorophyll, L has a downward trend, showing the leaf color gradually deepens with black-green, exists in all types of sample data, L B positive. ④ Through model prediction and evaluation, random forest (RF), partial least squares (PLS), support vector machine (SVM), and LASSO can be used to predict chlorophyll content effectively. Among the chlorophyll prediction models, RF had the best prediction performance, R²_c=0.939, RMSE_C=0.868 and MAE=0.686, R²_p=0.915, RMSE_P=1.194 and MAE=0.942. ⑤ Through model prediction and evaluation, RF, PLS, AdaBoost, and polynomial regression (POLYNOMIAL) can effectivelypredict leaf water contents. In the prediction model of leaf moisture content, the POLYNOMIAL prediction performance is the best, R²_c=0.884, RMSE_C=0.005 9 and MAE=0.005 2, R²_p=0.920 and RMSE_P=0.006 2 and MAE=0.005 7. The spectrophotometry method can effectively and rapidly determine leaf water and chlorophyll content, which is expected to provide an optional feasible method for nondestructive, rapid, and accurate determination of leaf water and chlorophyll content.

Key words：Hami melon; Chlorophyll; Water content; LAB; Model regression

收稿日期: 2022-08-09 修订日期: 2023-11-28

中图分类号:

S126

基金资助: 国家自然科学基金面上项目(61367001)，新疆维吾尔自治区教育厅自然科学重点项目(XJEDU2020I009)资助

通讯作者: 郭俊先 E-mail: junxianguo@163.com

作者简介: 李龙杰，1996年生，新疆农业大学机电工程学院硕士研究生 e-mail: 1018793176@qq.com

引用本文:

李龙杰，史勇，郭俊先. 基于分光测色反演哈密瓜叶片叶绿素与含水率研究[J]. 光谱学与光谱分析, 2024, 44(08): 2296-2302.
LI Long-jie, SHI Yong, GUO Jun-xian. Inversion of Chlorophyll and Water Content of Hami Melon Leaves Based on Spectrophotometry on Study. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2024, 44(08): 2296-2302.

链接本文:

https://www.gpxygpfx.com/CN/10.3964/j.issn.1000-0593(2024)08-2296-07 或 https://www.gpxygpfx.com/CN/Y2024/V44/I08/2296

[1] WANG Dong, SHEN Kai-cheng, FAN Ye-man, et al(王东, 沈楷程, 范叶满, 等). Transactions of the Chinese Society for Agricultural Machinery(农业机械学报), 2022, 53(4): 192.
[2] GAN Hai-ming, YUE Xue-jun, HONG Tian-sheng, et al(甘海明, 岳学军, 洪添胜, 等). Journal of South China Agricultural University(华南农业大学学报), 2018, 39(3): 102.
[3] LI Yong-mei, WANG Hao, ZHAO Yong, et al(李永梅, 王浩, 赵勇, 等). Acta Agriculturae Zhejiangensis(浙江农业学报), 2022, 34(4): 781.
[4] LI Hong, ZHANG Kai, CHEN Chao, et al(李红, 张凯, 陈超, 等). Transactions of the Chinese Society for Agricultural Machinery(农业机械学报), 2021, 52(2): 211.
[5] GAO Sheng, WANG Qiao-hua(高升, 王巧华). Chinese Optics(中国光学), 2021, 14(3): 566.
[6] SUN Xu-dong, HAO Yong, ZHANG Guang-wei(孙旭东, 郝勇, 张光伟). Journal of Chinese Agricultural Mechanization(中国农机化学报), 2015, 36(2): 150.
[7] JI Rong-hua, ZHENG Li-hua, DENG Xiao-lei, et al(冀荣华, 郑立华, 邓小蕾, 等). Transactions of the Chinese Society for Agricultural Machinery(农业机械学报), 2014, 45(8): 269.
[8] ZHANG Jun-yi, GAO De-hua, SONG Di, et al(张俊逸, 高德华, 宋迪, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2022, 42(5): 1514.
[9] MA Chun-yan, WANG Yi-lin, ZHAI Li-ting, et al(马春艳, 王艺琳, 翟丽婷, 等). Transactions of the Chinese Society for Agricultural Machinery(农业机械学报), 2022, 53(6): 217.
[10] HE Gui-fang, WU Jian, PENG Jian, et al(何桂芳, 吴见, 彭建, 等). Journal of Northwest Forestry University(西北林学院学报), 2022, 37(1): 25.
[11] ZHANG Hui-chun, ZHANG Meng, BIAN Li-ming, et al(张慧春, 张萌, 边黎明, 等). Transactions of the Chinese Society for Agricultural Machinery(农业机械学报): 2022, 53(4): 313.
[12] CHENG Li-zhen, ZHU Xi-cun, GAO Lu-lu, et al(程立真, 朱西存, 高璐璐, 等). Acta Horticulturae Sinica(园艺学报), 2017, 44(2): 381.