基于优化光谱指数的土壤有机质含量估算  

Estimation of Soil Organic Matter Content Based on Optimized Spectral Index

作　　者：尼加提.卡斯木 茹克亚.萨吾提 师庆东[2,3] 买合木提.巴拉提 米热阿地力.库尔班 苏比努尔.居来提 NIJAT Kasim;RUKEYA Sawut;SHI Qingdong;MAIHEMUTI Balati;MIREADILI Kuerban;SUBINUER Julaiti(Institute of Arid Ecology Environment,Xinjiang University,Urumqi 830046,China;Key Laboratory of Oasis Ecology,Ministry of Education,Xinjiang University,Urumqi 830046,China;College of Resources and Environmental Sciences,Xinjiang University,Urumqi 830046,China)

机构地区：[1]新疆大学干旱生态环境研究所,乌鲁木齐830046 [2]新疆大学绿洲生态教育部重点实验室,乌鲁木齐830046 [3]新疆大学资源与环境科学学院,乌鲁木齐830046

出　　处：《农业机械学报》2018年第11期155-163,共9页Transactions of the Chinese Society for Agricultural Machinery

基　　金：国家自然科学基金项目(41762019;U1703237);新疆大学博士科研启动基金项目(BS160232);新疆大学大学生创新训练项目(201710755099)

摘　　要：为了寻求估算土壤有机质含量的最佳光谱参数,实现土壤养分无损监测,使用ASD Field-Spec3型高光谱仪对野外采集的土壤样品进行室内光谱测定,并通过重铬酸钾氧化容量法测定土壤样品有机质质量比;利用两波段优化算法对构建的新算法(SOMCI/ND)进行波段优化,筛选基于不同光谱数据(原始光谱反射率及其对应的4种数学变换)运算下的最敏感波段组合,从而建立土壤有机质质量比高光谱估算模型。结果表明:通过归一化光谱指数(IND)和概念指数(ICI)比值构建的新算法(SOMCI/ND)优化后与土壤有机质质量比之间的相关性显著提高,在光谱原始数据及其平方根、倒数变换形式下,相关系数绝对值达到0. 82,且敏感的组合波段集中在2 220~2 240 nm和2 160~2 195 nm。基于平方根波段优化的估算模型效果最佳,估算精度R2P为0. 84,RMSEP为2. 24 g/kg,RPD为2. 89。对光谱数据的适当数学变换有利于优化光谱指数,更好地估算土壤有机质质量比,进一步实现土壤有机质质量比的高精度动态监测。The rapid monitoring of soil organic matter content based on hyperspectral data is of great significance for evaluating soil fertility.The best spectral parameters for predicting soil organic matter content were tried to find and non-destructive monitoring of soil nutrients was achieved.ASD Field Spec3 spectrometer was used to measure the indoor spectra of soil samples collected in the field,and the organic matter content of soil samples was measured by the potassium dichromate oxidation capacity method;the nitrogen planar component index(SOMCI/ND)was optimized by two-band optimization algorithm.Band optimization,screening the most sensitive spectral parameters of different spectral data(the original spectral reflectance and its corresponding four mathematical transformations),thus establishing a hyperspectral estimation model of soil organic matter content.The results showed that the correlations between soil organic matter content and the new algorithm(SOMCI/ND)optimized by the normalized spectral index(I ND)and conceptual index(I CI)ratios were significantly improved.The raw data in the spectrum and its square root and reciprocal transformation form,the absolute value of correlation coefficient reached 0.82,and the sensitive combination bands were concentrated in 2 220~2 240 nm and 2 160~2 195 nm.The prediction model based on the square root band optimization had the best effect.The prediction accuracy was R 2 P of 0.84,RMSEP of 2.24 g/kg and RPD of 2.89.Therefore,the appropriate mathematical transformation of the spectral data was conducive to optimizing the spectral index to better estimate the soil organic matter content,and further achieve high-precision dynamic monitoring of soil organic matter.

关 键 词：土壤有机质 光谱分析 高光谱 优化光谱指数 

分 类 号：O433.4[机械工程—光学工程] S27[理学—光学]