{"title":"利用基于变换的动态光谱指数和随机森林估算玉米冠层氮含量","authors":"Shuting Yang, Jianbei Li, Ji Li, Xuejian Zhang, Cong Ma, Zhengyu Liu, Mengyan Ren","doi":"10.3390/su16188011","DOIUrl":null,"url":null,"abstract":"The monitoring of maize health status is crucial for achieving sustainable agricultural development. Canopy nitrogen content (CNC) is essential for the synthesis of proteins and chlorophyll in maize leaves and, thus, significantly influences maize growth and yield. In this study, we developed a CNC spectral estimation model based on transform-based dynamic spectral indices (TDSI) and the random forest (RF) algorithm, enabling the rapid monitoring of CNC in maize canopy leaves. A total of 60 maize canopy leaf samples and the corresponding field canopy spectra were collected. Subsequently, the canopy spectra data were transformed using centralization transformation (CT), first derivative (D1), second derivative (D2), detrend transformation (DT), and min-max normalization (MMN) methods. Three types of band combination methods (band difference, band ratio, and normalized difference) were used to construct the TDSIs. Finally, the optimal TDSI was selected and used as the independent variable, and the measured CNC was used as the dependent variable to build a CNC spectral estimation model based on the RF algorithm. Results indicated that (1) TDSIs can more accurately characterize the CNC in maize, with a correlation coefficient approximately 102% higher than those of raw spectral bands. (2) The optimal TDSIs included TDSI1247,1249CT-RI, TDSI625,641CT-NDI, TDSI540,703D1-RI, TDSI514,540D1-RI, TDSI514,530D1-DI, TDSI540,697D1-NDI, TDSI970,1357D2-DI, TDSI523,1031D2-NDI, TDSI617,620DT-RI, and TDSI2109,2127MMN-NDI. (3) The CNC spectral estimation model based on the optimal TDSIs, and the RF algorithm achieved accuracy indices with R2 and RPIQ of 0.92 and 4.99, respectively, representing a maximum improvement of approximately 67.27% over the traditional CNC spectral estimation model (based on the R2 value). This study provides an approach for the rapid and accurate estimation of CNC in maize, contributing to the sustainable development of agriculture.","PeriodicalId":3,"journal":{"name":"ACS Applied Electronic Materials","volume":null,"pages":null},"PeriodicalIF":4.3000,"publicationDate":"2024-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Estimating the Canopy Nitrogen Content in Maize by Using the Transform-Based Dynamic Spectral Indices and Random Forest\",\"authors\":\"Shuting Yang, Jianbei Li, Ji Li, Xuejian Zhang, Cong Ma, Zhengyu Liu, Mengyan Ren\",\"doi\":\"10.3390/su16188011\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"The monitoring of maize health status is crucial for achieving sustainable agricultural development. Canopy nitrogen content (CNC) is essential for the synthesis of proteins and chlorophyll in maize leaves and, thus, significantly influences maize growth and yield. In this study, we developed a CNC spectral estimation model based on transform-based dynamic spectral indices (TDSI) and the random forest (RF) algorithm, enabling the rapid monitoring of CNC in maize canopy leaves. A total of 60 maize canopy leaf samples and the corresponding field canopy spectra were collected. Subsequently, the canopy spectra data were transformed using centralization transformation (CT), first derivative (D1), second derivative (D2), detrend transformation (DT), and min-max normalization (MMN) methods. Three types of band combination methods (band difference, band ratio, and normalized difference) were used to construct the TDSIs. Finally, the optimal TDSI was selected and used as the independent variable, and the measured CNC was used as the dependent variable to build a CNC spectral estimation model based on the RF algorithm. Results indicated that (1) TDSIs can more accurately characterize the CNC in maize, with a correlation coefficient approximately 102% higher than those of raw spectral bands. (2) The optimal TDSIs included TDSI1247,1249CT-RI, TDSI625,641CT-NDI, TDSI540,703D1-RI, TDSI514,540D1-RI, TDSI514,530D1-DI, TDSI540,697D1-NDI, TDSI970,1357D2-DI, TDSI523,1031D2-NDI, TDSI617,620DT-RI, and TDSI2109,2127MMN-NDI. (3) The CNC spectral estimation model based on the optimal TDSIs, and the RF algorithm achieved accuracy indices with R2 and RPIQ of 0.92 and 4.99, respectively, representing a maximum improvement of approximately 67.27% over the traditional CNC spectral estimation model (based on the R2 value). This study provides an approach for the rapid and accurate estimation of CNC in maize, contributing to the sustainable development of agriculture.\",\"PeriodicalId\":3,\"journal\":{\"name\":\"ACS Applied Electronic Materials\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":4.3000,\"publicationDate\":\"2024-09-13\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"ACS Applied Electronic Materials\",\"FirstCategoryId\":\"93\",\"ListUrlMain\":\"https://doi.org/10.3390/su16188011\",\"RegionNum\":3,\"RegionCategory\":\"材料科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"ENGINEERING, ELECTRICAL & ELECTRONIC\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"ACS Applied Electronic Materials","FirstCategoryId":"93","ListUrlMain":"https://doi.org/10.3390/su16188011","RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENGINEERING, ELECTRICAL & ELECTRONIC","Score":null,"Total":0}
Estimating the Canopy Nitrogen Content in Maize by Using the Transform-Based Dynamic Spectral Indices and Random Forest
The monitoring of maize health status is crucial for achieving sustainable agricultural development. Canopy nitrogen content (CNC) is essential for the synthesis of proteins and chlorophyll in maize leaves and, thus, significantly influences maize growth and yield. In this study, we developed a CNC spectral estimation model based on transform-based dynamic spectral indices (TDSI) and the random forest (RF) algorithm, enabling the rapid monitoring of CNC in maize canopy leaves. A total of 60 maize canopy leaf samples and the corresponding field canopy spectra were collected. Subsequently, the canopy spectra data were transformed using centralization transformation (CT), first derivative (D1), second derivative (D2), detrend transformation (DT), and min-max normalization (MMN) methods. Three types of band combination methods (band difference, band ratio, and normalized difference) were used to construct the TDSIs. Finally, the optimal TDSI was selected and used as the independent variable, and the measured CNC was used as the dependent variable to build a CNC spectral estimation model based on the RF algorithm. Results indicated that (1) TDSIs can more accurately characterize the CNC in maize, with a correlation coefficient approximately 102% higher than those of raw spectral bands. (2) The optimal TDSIs included TDSI1247,1249CT-RI, TDSI625,641CT-NDI, TDSI540,703D1-RI, TDSI514,540D1-RI, TDSI514,530D1-DI, TDSI540,697D1-NDI, TDSI970,1357D2-DI, TDSI523,1031D2-NDI, TDSI617,620DT-RI, and TDSI2109,2127MMN-NDI. (3) The CNC spectral estimation model based on the optimal TDSIs, and the RF algorithm achieved accuracy indices with R2 and RPIQ of 0.92 and 4.99, respectively, representing a maximum improvement of approximately 67.27% over the traditional CNC spectral estimation model (based on the R2 value). This study provides an approach for the rapid and accurate estimation of CNC in maize, contributing to the sustainable development of agriculture.