Error correction of spectral reflectance based on constrained linear spectral mixture model

2018 International Conference on Advanced Materials, Intelligent Manufacturing and Automation

材料科学;机械制造;运输工程

Zhang, Haonan^1,2 ; Wen, Xingping^1,2 ; Luo, Dayou^1,2 ; Xu, Junlong^1,2 ; Li, Jinbo^1,2

Faculty of Land Resource Engineering, Kunming University of Science and Technology, Yunnan Kunming, 650093, China^1

Mineral Resources Prediction and Evaluation Engineering Laboratory of Yunnan Province, Yunnan Kunming

650093, China^2

关键词: Classification accuracy;    Linear spectral mixture model;    Quantitative remote sensing;    Remote sensing images;    Remote sensing technology;    Spatial information extraction;    Spectral measurement;    Spectral reflectances;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/382/4/042033/pdf
DOI  :  10.1088/1757-899X/382/4/042033

学科分类：材料科学（综合）

来源: IOP

【 摘 要 】

With the rapid development of remote sensing technology, spatial information extraction from remote sensing images is playing an increasingly important role. However, spectral signals of various ground objects are inevitably recorded in pixel. The classification accuracy is reduced and the development of quantitative remote sensing is restricted. In this paper, four groups of mixed pixel samples are designed, using the ASD FiledSpec3 spectroradiometer and the constrained linear spectral model to analyse the error distribution and the error correction methods are proposed. The results show that the constrained linear spectral model can be used in experiments where the illumination is uniform and the measured objects are smooth, and model has strong applicability. After correction, the average error of each group was reduced by 0.00847, and this method can effectively reduce the error caused by wavelength variation in the actual spectral measurement experiment.