[目的]本试验利用GF-1遥感影像估测土壤有机质含量.[方法]该文对扶余市耕作区土壤进行采样,在实验室化验土壤样品的有机质含量,分析GF-1各波段反射率及其变换形式与土壤有机质含量的相关性,确定有机质的敏感波段,建立土壤有机质含量的单波段与多波段估测模型,旨在通过比较估测模型的精度和稳定性,确定研究区土壤有机质含量的最优估测模型.[结果]F-1各波段反射率与有机质含量均呈显著负相关,且在第3波段达到最大值,其相关系数为-0.805,均方根误差为0.362;将反射率进行幂、指数变换以后可以有效提高与有机质含量的相关性,相关系数分别提高至-0.886和-0.872,均方根误差下降至0.283和0.342;利用前3个波段反射率指数变换建立起的多元估测模型,模型判定系数R2达到0.851,检验样本的均方根误差降低至0.172,表明此模型的估测精度较高、稳定性较好.[结论]GF-1遥感影像可以作为估测土壤有机质含量的遥感数据源,并为使用GF-1遥感影像估测土壤成分等方面的研究提供参考.%[Objective] This paper aimed to test the application of GF-1 remote sensing images for estimating soil organic matter content.[Method] Taken the farming area soil of Fuyu city as tested samples,the organic matter contents of soil samples in laboratory tests were detected,the correlation of GF-1 band reflectance and transforming form with the content of soil organic matter were analyzed to determine organic matter sensitive bands,the spectral estimation models of a single band or multi band of soil organic matter content were established,and the area estimation model of the optimal content of soil organic matter in the study was chosen compared with the accuracy and stability of these estimation model.[Result] The GF-1 band reflectance had a significantly negative correlation with organic matter content and reached the maximum value in the third band,the correlation coefficient was-0.805,and the RMS error was 0.362;The reflectivity of power index would transform and could effectively improve the correlation with the content of organic matter,its correlation coefficients were increased to -0.886 and-0.872,and root mean square error decreased to 0.283 and 0.342;The 3 yuan to establish reflectance index transformation estimation model,model determination coefficient(R2) reached 0.851,and the RMS error of test samples was reduced to 0.172,which indicated that this model had a high accuracy,good stability estimation.[Conclusion] The GF-1 remote sensing image could be used as a remote sensing data source for estimating soil organic matter content and would provide a reference for the study of GF-1 remote sensing image estimation of soil components.
展开▼
