Estimation of Soil Cations Based on Visible and Near-Infrared Spectroscopy and Machine Learning
Soil exchange cations are a basic indicator of soil quality and environmental clean-up potential. The accurate and efficient acquisition of information on soil cation content is of great importance for the monitoring of soil quality and pollution prevention. At present, few scholars focus on soil ex...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-06-01
|
| Series: | Agriculture |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2077-0472/13/6/1237 |
| _version_ | 1797596511104663552 |
|---|---|
| author | Yiping Peng Ting Wang Shujuan Xie Zhenhua Liu Chenjie Lin Yueming Hu Jianfang Wang Xiaoyun Mao |
| author_facet | Yiping Peng Ting Wang Shujuan Xie Zhenhua Liu Chenjie Lin Yueming Hu Jianfang Wang Xiaoyun Mao |
| author_sort | Yiping Peng |
| collection | DOAJ |
| description | Soil exchange cations are a basic indicator of soil quality and environmental clean-up potential. The accurate and efficient acquisition of information on soil cation content is of great importance for the monitoring of soil quality and pollution prevention. At present, few scholars focus on soil exchangeable cations using remote sensing technology. This study proposes a new method for estimating soil cation content using hyperspectral data. In particular, we introduce Boruta and successive projection (SPA) algorithms to screen feature variables, and we use Guangdong Province, China, as the study area. The backpropagation neural network (BPNN), genetic algorithm–based back propagation neural network (GABP) and random forest (RF) algorithms with 10-fold cross-validation are implemented to determine the most accurate model for soil cation (Ca<sup>2+</sup>, K<sup>+</sup>, Mg<sup>2+</sup>, and Na<sup>+</sup>) content estimations. The model and hyperspectral images are combined to perform the spatial mapping of soil Mg<sup>2+</sup> and to obtain the spatial distribution information of images. The results show that Boruta was the optimal algorithm for determining the characteristic bands of soil Ca<sup>2+</sup> and Na<sup>+</sup>, and SPA was the optimal algorithm for determining the characteristic bands of soil K<sup>+</sup> and Mg<sup>2+</sup>. The most accurate estimation models for soil Ca<sup>2+</sup>, K<sup>+</sup>, Mg<sup>2+</sup>, and Na<sup>+</sup> contents were Boruta-RF, SPA-GABP, SPA-RF and Boruta-RF, respectively. The estimation effect of soil Mg<sup>2+</sup> (R<sup>2</sup> = 0.90, ratio of performance to interquartile range (RPIQ) = 3.84) was significantly better than the other three elements (Ca<sup>2+</sup>: R<sup>2</sup> = 0.83, RPIQ = 2.47; K<sup>+</sup>: R<sup>2</sup> = 0.83, RPIQ = 2.58; Na<sup>+</sup>: R<sup>2</sup> = 0.85, RPIQ = 2.63). Moreover, the SPA-RF method combined with HJ-1A HSI images was selected for the spatial mapping of soil Mg<sup>2+</sup> content with an R<sup>2</sup> of 0.71 and RPIQ of 2.05. This indicates the ability of the SPA-RF method to retrieve soil Mg<sup>2+</sup> content at the regional scale. |
| first_indexed | 2024-03-11T02:52:25Z |
| format | Article |
| id | doaj.art-0f4dc681746a499bb49e65dc7d5da9ad |
| institution | Directory Open Access Journal |
| issn | 2077-0472 |
| language | English |
| last_indexed | 2024-03-11T02:52:25Z |
| publishDate | 2023-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Agriculture |
| spelling | doaj.art-0f4dc681746a499bb49e65dc7d5da9ad2023-11-18T08:52:12ZengMDPI AGAgriculture2077-04722023-06-01136123710.3390/agriculture13061237Estimation of Soil Cations Based on Visible and Near-Infrared Spectroscopy and Machine LearningYiping Peng0Ting Wang1Shujuan Xie2Zhenhua Liu3Chenjie Lin4Yueming Hu5Jianfang Wang6Xiaoyun Mao7College of Natural Resources and Environment, South China Agricultural University, Guangzhou 510642, ChinaCollege of Natural Resources and Environment, South China Agricultural University, Guangzhou 510642, ChinaGuangdong Academy of Social Sciences, Guangzhou 510635, ChinaCollege of Natural Resources and Environment, South China Agricultural University, Guangzhou 510642, ChinaCollege of Natural Resources and Environment, South China Agricultural University, Guangzhou 510642, ChinaCollege of Tropical Crops, Hainan University, Haikou 570228, ChinaCollege of Natural Resources and Environment, South China Agricultural University, Guangzhou 510642, ChinaCollege of Natural Resources and Environment, South China Agricultural University, Guangzhou 510642, ChinaSoil exchange cations are a basic indicator of soil quality and environmental clean-up potential. The accurate and efficient acquisition of information on soil cation content is of great importance for the monitoring of soil quality and pollution prevention. At present, few scholars focus on soil exchangeable cations using remote sensing technology. This study proposes a new method for estimating soil cation content using hyperspectral data. In particular, we introduce Boruta and successive projection (SPA) algorithms to screen feature variables, and we use Guangdong Province, China, as the study area. The backpropagation neural network (BPNN), genetic algorithm–based back propagation neural network (GABP) and random forest (RF) algorithms with 10-fold cross-validation are implemented to determine the most accurate model for soil cation (Ca<sup>2+</sup>, K<sup>+</sup>, Mg<sup>2+</sup>, and Na<sup>+</sup>) content estimations. The model and hyperspectral images are combined to perform the spatial mapping of soil Mg<sup>2+</sup> and to obtain the spatial distribution information of images. The results show that Boruta was the optimal algorithm for determining the characteristic bands of soil Ca<sup>2+</sup> and Na<sup>+</sup>, and SPA was the optimal algorithm for determining the characteristic bands of soil K<sup>+</sup> and Mg<sup>2+</sup>. The most accurate estimation models for soil Ca<sup>2+</sup>, K<sup>+</sup>, Mg<sup>2+</sup>, and Na<sup>+</sup> contents were Boruta-RF, SPA-GABP, SPA-RF and Boruta-RF, respectively. The estimation effect of soil Mg<sup>2+</sup> (R<sup>2</sup> = 0.90, ratio of performance to interquartile range (RPIQ) = 3.84) was significantly better than the other three elements (Ca<sup>2+</sup>: R<sup>2</sup> = 0.83, RPIQ = 2.47; K<sup>+</sup>: R<sup>2</sup> = 0.83, RPIQ = 2.58; Na<sup>+</sup>: R<sup>2</sup> = 0.85, RPIQ = 2.63). Moreover, the SPA-RF method combined with HJ-1A HSI images was selected for the spatial mapping of soil Mg<sup>2+</sup> content with an R<sup>2</sup> of 0.71 and RPIQ of 2.05. This indicates the ability of the SPA-RF method to retrieve soil Mg<sup>2+</sup> content at the regional scale.https://www.mdpi.com/2077-0472/13/6/1237soil cationsVIS-NIR spectroscopyfeature screeningmachine learn algorithm |
| spellingShingle | Yiping Peng Ting Wang Shujuan Xie Zhenhua Liu Chenjie Lin Yueming Hu Jianfang Wang Xiaoyun Mao Estimation of Soil Cations Based on Visible and Near-Infrared Spectroscopy and Machine Learning Agriculture soil cations VIS-NIR spectroscopy feature screening machine learn algorithm |
| title | Estimation of Soil Cations Based on Visible and Near-Infrared Spectroscopy and Machine Learning |
| title_full | Estimation of Soil Cations Based on Visible and Near-Infrared Spectroscopy and Machine Learning |
| title_fullStr | Estimation of Soil Cations Based on Visible and Near-Infrared Spectroscopy and Machine Learning |
| title_full_unstemmed | Estimation of Soil Cations Based on Visible and Near-Infrared Spectroscopy and Machine Learning |
| title_short | Estimation of Soil Cations Based on Visible and Near-Infrared Spectroscopy and Machine Learning |
| title_sort | estimation of soil cations based on visible and near infrared spectroscopy and machine learning |
| topic | soil cations VIS-NIR spectroscopy feature screening machine learn algorithm |
| url | https://www.mdpi.com/2077-0472/13/6/1237 |
| work_keys_str_mv | AT yipingpeng estimationofsoilcationsbasedonvisibleandnearinfraredspectroscopyandmachinelearning AT tingwang estimationofsoilcationsbasedonvisibleandnearinfraredspectroscopyandmachinelearning AT shujuanxie estimationofsoilcationsbasedonvisibleandnearinfraredspectroscopyandmachinelearning AT zhenhualiu estimationofsoilcationsbasedonvisibleandnearinfraredspectroscopyandmachinelearning AT chenjielin estimationofsoilcationsbasedonvisibleandnearinfraredspectroscopyandmachinelearning AT yueminghu estimationofsoilcationsbasedonvisibleandnearinfraredspectroscopyandmachinelearning AT jianfangwang estimationofsoilcationsbasedonvisibleandnearinfraredspectroscopyandmachinelearning AT xiaoyunmao estimationofsoilcationsbasedonvisibleandnearinfraredspectroscopyandmachinelearning |