Retrieval of soil salt content based on remote sensing

soil salinization is one of the most important problems of land degradation and the basic environmental problem in arid and semi-arid regions. The remote sensing technology can rapidly and timely provide the information about properties, geographical distribution and extent of soil salinization. Taking the city Huanghua of Hebei Province in China as the study area and through the analysis on the data of soil spectrum measured in field, it was found in this study that vegetation affects greatly the spectral response of soil for salt content, and at the same time the spectrum ranging from 451.42 nm to 593.79 nm is much sensitive to the variation in soil salt content and, therefore, based on the analysis of soil spectrum, the relevant statistic model for predicting soil salt content was constructed. However, due to rather complicated non-linear relations existed between image features and soil salt content, the results of soil salt content retrieved from the statistic model is not so ideal. For this reason, an artificial neural network model(BP model) was constructed and applied in the retrieval of soil salt content. Because of its superior ability for solving the non-linear problem, the BP model provided a much better accuracy in retrieval of soil salt content compared with the results from the statistic model.