| 基于不同卷积神经网络模型的红壤有机质高光谱估算 |
| |
| 引用本文： | 钟亮, 郭熙, 国佳欣, 徐喆, 朱青, 丁萌. 基于不同卷积神经网络模型的红壤有机质高光谱估算[J]. 农业工程学报, 2021, 37(1): 203-212. DOI: 10.11975/j.issn.1002-6819.2021.01.025 |
| |
| 作者姓名： | 钟亮 郭熙 国佳欣 徐喆 朱青 丁萌 |
| |
| 作者单位： | 1.江西农业大学国土资源与环境学院, 南昌 330045;2.江西省鄱阳湖流域农业资源与生态重点实验室, 南昌 330045 |
| |
| 基金项目： | 国家重点研发计划项目（2017YFD0301603）；国家自然科学基金项目（41361049）；研究生创新专项基金（NDYC2020-S008） |
| |
| 摘 要： | 以卷积神经网络（Convolutional Neural Network，CNN）为代表的深度学习方法因具有强大的特征学习能力已被广泛应用于计算机视觉、自然语言处理等领域，但在土壤高光谱遥感领域研究较少。为探究其在小样本数据集下，通过高光谱数据估算土壤有机质（Soil Organic Matter，SOM）的可行性，以江西省奉新县北部为研究区，248个红壤样本为研究对象。对比分析深度学习方法CNN、多层感知器（Multilayer Perceptron，MLP）、常用的机器学习方法随机森林（Random Forest，RF）和支持向量机（Support Vector Machine，SVM）在不同光谱预处理下的建模效果，在此基础上分别建立5种各具特点的CNN结构模型，以探讨不同网络结构的建模效果，包括最早提出的LeNet-5、具有大卷积核的AlexNet-8、采用小卷积核的VGGNet-7、含有Inception结构的GoogLeNet-7以及使用残差学习的ResNet-13。此外，还探讨了VGGNet模型在5种不同网络深度下的模型效果。结果表明：在使用原始光谱的情况下，CNN模型依然能够取得较好的建模效果（相对分析误差>2.5）；浅层CNN结构优于深层建模效果，超参数较小的卷积核、步长和池化范围有助于提取更多的特征数量，提高建模精度；VGGNet-7网络结构在所有模型中表现最为突出，在训练集上决定系数为0.895，均方根误差为4.145 g/kg，相对分析误差为3.447，在验证集上决定系数为0.901，均方根误差为4.647 g/kg，相对分析误差为3.291，具有极好的模型估测能力；680、1 360、1 390、1 920、2 310 nm及其附近是VGGNet-7建模过程中所提取的SOM重要特征波长。因此，CNN能够简化光谱预处理过程，在土壤高光谱遥感小样本建模中具备可行性，具有非常广阔的应用前景，VGGNet-7可以应用于红壤地区通过高光谱数据快速、准确的估算SOM含量。
|
| 关 键 词： | 土壤 模型 卷积神经网络 有机质 高光谱 |
| 收稿时间： | 2020-07-23 |
| 修稿时间： | 2020-10-10 |
Hyperspectral estimation of organic matter in red soil using different convolutional neural network models |
| |
| Zhong Liang, Guo Xi, Guo Jiaxin, Xu Zhe, Zhu Qing, Ding Meng. Hyperspectral estimation of organic matter in red soil using different convolutional neural network models[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2021, 37(1): 203-212. DOI: 10.11975/j.issn.1002-6819.2021.01.025 |
| |
| Authors: | Zhong Liang Guo Xi Guo Jiaxin Xu Zhe Zhu Qing Ding Meng |
| |
| Affiliation: | 1.College of Land Resources and Environment, Jiangxi Agricultural University, Nanchang 330045, China;2.Key Laboratory of Poyang Lake Watershed Agricultural Resources and Ecology of Jiangxi Province, Nanchang 330045, China |
| |
| Abstract: | Deep learning represented by Convolutional Neural Networks (CNN) has been increasing rapidly in recent years, due to its powerful feature learning for computer vision and natural language processing. But there are few studies in the field of hyperspectral remote sensing in soil. Therefore, this study aims to estimate Soil Organic Matter (SOM) using hyperspectral images in small sample dataset, thereby to investigate the modeling effects of different network structures. A total of 248 red soil samples were collected from the northern Fengxin county, Jiangxi province, China. A geospectrometer was used to capture the spectral data. The original spectral data was resampled at 10 nm intervals, after removing the edge bands of 350-399 nm and 2451-2500 nm with a low signal-to-noise ratio. A total of 205 original spectral bands and their derivative transformation were obtained as input data, while the SOM content as output data of the model. Firstly, the modeling effects of CNN were compared, such as Multilayer Perceptron (MLP), Random Forest (RF) and Support Vector Machine (SVM) under different spectral pretreatments. Five CNN structures were established, including the earliest LeNet-5, AlexNet-8 with large convolutional core, VGGNet-7 with small convolutional core, GoogLeNet-7 with inception structure, and ResNet-13 with residual learning, particularly on the modeling effects of VGGNet model at five depths. Secondly, all models were evaluated using random deactivation (Dropout) and early stopping to prevent overfitting of the model by three indicators: decision coefficient (R2), Root Mean Square Error (RMSE) and Relative Analytical Error (RPD). Finally, the black box of CNN model was explained. The results showed that: 1) Due to the strong capability of feature learning in CNN models, the RPD of each CNN model in the validation set was greater than 2.5 in the case of the original spectral data, indicating excellent prediction capability and a better way to predict SOM content using hyperspectral images. 2) In the comparison of different network structures, an optimal model was determined in the network structures of LeNet-5 and VGGNet-7 with small convolutional nuclei, step length, and pooling range of hyper parameters, although the later GoogLeNet-7 and ResNet-13 both incorporated special structures. Therefore, the setting of some hyper parameters in the CNN model can be more critical than the network structure. In different depths, the model was prone to overfitting and unstable, as the network depth increased, where the shallow CNN structure was better than the deep one. 3) An optimal model was achieved in the VGGNet-7 network structure with the excellent model estimation power: R2 was 0.895 and RMSE was 4.145 g/kg on the training set, while R2 was 0.901, RMSE was 4.647 g/kg and RPD was 3.291 on the verification set. 4) The wavelengths of 1390, 680, 2310, 1360, 1920 nm and its vicinities were the important for SOM and they were extracted from the process of VGGNet-7 model establishment. The CNN can be expected for very broad application prospects, due to its simple spectral pre-processing, and feasibility in small samples of soil hyperspectral remote sensing. Therefore, the VGGNet-7 can be applied to the red soil area for rapid and accurate estimation of SOM content using hyperspectral data. |
| |
| Keywords: | soils models convolutional neural network organic matter hyperspectrum |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《农业工程学报》浏览原始摘要信息 |
|
点击此处可从《农业工程学报》下载全文 |
|
