Research on the Detection Methods of Soil Nitrogen and Organic Carbon Using Hyperspectral Imaging Technology
Soil is an important environment for crop growth,quick and accurately access the information of soil nutrient content is a prerequisite to scientific fertilization.In this work,hyperspectral imagingtechnology was applied for the fast measurement soil organic carbon and totalnitrogen.The soil samples were collectedfrom the north Zhejiang plain farmland.Soil types included paddy soil,coastal saline soil and red soil.The main conclusions were as follows:the methods of Minimum Noise Separation Transform(MNST) and Orthogonal Signal Correction(OSC) were applied for the decomposition and reconstruction of the hyperspectral data,so as to obtain clear images and smooth spectral curves.Three variable selection methods,namely,competitive adaptive reweighted sampling(CARS),successive projections algorithm(SPA) and Monte Carlo–uninformative variable elimination(MC-UVE) were applied to select feature spectral variables of soil nitrogen and organic carbon.Meanwhile,the texture variables of the grayscale images at the feature wavebands were extracted by gray level co-occurrence matrix(GLCM).The information fusion methods for the effective spectral variables and image texture variables were proposed for the establishment of classification models of soil types.An improvement of 10 ％of the correct classification rate was obtained in model performance compared to those when either spectral data or textural data were used alone.The prediction models for each soil type were developed for the prediction of soil nitrogen and organic carbon.The optimal prediction results fororganic carbon and total nitrogen were obtained from the paddy soil samples.The values of coefficient of determination(R2) and residual prediction deviation(RPD) were respectively 0.87,0.85,and 2.8,2.6.The results indicated that the comprehensive application of spectral information and texture information can effectively improve the correct classification rate of soil types and the prediction accuracy for the determination of soil organic carbon and total nitrogen content.
S.Y.Jia Y.J.Wang H.Y.Li Q.Li
China Jiliang University,No.258 Xueyuan Road,Hangzhou,China
国际会议
国际测量技术联合会(IMEKO)2017年国际测量技术研讨会
杭州
英文
310-310
2017-09-04（万方平台首次上网日期，不代表论文的发表时间）