HJ-1A HSI与Sentinel-2A遥感数据土壤全氮含量反演精度的对比研究  

Accuracy Comparison of Soil Total Nitrogen Content Inversion Based on HJ-1A HSI and Sentinel-2A Remote Sensing Data

作　　者：马驰[1] MA Chi(Department of Surveying and Mapping Engineering,Liaoning Provincial College of Communications,Shenyang 110122,China)

机构地区：[1]辽宁省交通高等专科学校测绘工程系,辽宁沈阳110122

出　　处：《无线电工程》2021年第12期1540-1548,共9页Radio Engineering

基　　金：国家自然科学基金项目(41371332);中国地质调查局项目(1212010911084);辽宁省交通高等专科学校科研项目(lnccrc202126)。

摘　　要：HJ-1A高光谱影像具有较多的光谱波段和较高的光谱分辨率,在土壤成分探测方面潜力巨大。然而,由于高光谱影像各波段间的信息冗余度高,使得对HJ-1A高光谱影像进行数据处理难度较大。为了探究Sentinel-2A遥感数据在土壤全氮含量方面的反演能力,以HJ-1A HSI和Sentinel-2A遥感影像为数据源,以吉林省农安县为研究对象,结合研究区土壤采样的实验室全氮含量化验值,分析两种遥感影像各波段反射率及其数学变换与研究区土壤全氮含量的相关性,提取土壤全氮含量的敏感波段并建立全氮含量的反演模型。通过比较两种遥感影像在土壤全氮含量方面的反演能力与反演精度,探究Sentinel-2A多光谱遥感影像反演土壤全氮含量的可行性。研究结果表明,HJ-1A HSI影像在可见光波段的反射率与农安县土壤全氮含量具有较强的负相关性,并在第28波段达到峰值;将遥感影像反射率进行适当的数学变换及一阶微分变换后可以有效提升与土壤全氮含量的相关性;利用Sentinel-2A遥感影像建立的研究区土壤全氮含量最优反演模型的模型精度与HJ-1A HSI影像的最优反演模型的模型精度相近,表明利用多光谱遥感影像Sentinel-2A反演土壤全氮含量的方法可行。HJ-1A hyperspectral image has more spectral bands and higher spectral resolution,and has great potential in soil composition detection.However,due to the high information redundancy between bands of hyperspectral images,it is difficult to process HJ-1A hyperspectral images.In order to explore the inversion capability of Sentinel-2A remote sensing data in terms of soil total nitrogen content,by taking HJ-1A HSI and Sentinel-2A remote sensing images as data sources,Nong’an County of Jilin Province as the research object,and combined with the laboratory total nitrogen content test values of soil samples in the studied area,the correlation between the reflectance and mathematical transformation of each band of the two remote sensing images and the soil total nitrogen content of the studied area is analyzed.The sensitive band of soil total nitrogen content is extracted and the inversion model of total nitrogen content is established.The feasibility of soil total nitrogen content inversion from Sentinel-2A multispectral remote sensing image is explored by comparing the inversion capabilities and accuracies of the two remote sensing images in soil total nitrogen content.The results show that the reflectance of HJ-1A HSI image in the visible band has a strong negative correlation with the content of soil total nitrogen in Nong’an County,and reaches the peak in the 28th band.After proper mathematical transformation and first-order differential transformation of the remote sensing image reflectance,the correlation with soil total nitrogen content can be effectively improved.The model accuracy of the optimal inversion model of soil total nitrogen content established by Sentinel-2A remote sensing image is similar to that of HJ-1A HSI image,indicating that the method of soil total nitrogen content inversion by Sentinel-2A multispectral remote sensing image is feasible.

关 键 词：HJ-1A HSI Sentinel-2A 全氮含量 定量反演 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置] S151.9[自动化与计算机技术—控制科学与工程]