Soil Salt Content Inversion Using Remote Sensing And Its Application In Study On Oriental Migratory Locust

Locust hazard is a destructive biological disaster. It is one of the three major natural hazards together with floods and droughts. Once it outbreaks, it will be a severe threat to the safety of local agriculture, farmer's income and the stability of society. Soil physical and chemical property have great effect on the spawn and growth of oriental migratory locust. As for soil salt content, the extent of soil salt content plays a very important role on the choice of places where they lay eggs, hatch and development of spawn. At the same time, the difference of soil salt content accordingly causes the difference of the distribution, coverage level, category of vegetation, and indirectly affect the development of oriental migratory locust. Moreover, soil salinization is the major land degradation problem of the world .Globally, an estimated 9.5×10~8 hm~2 of land are affected by salinity. By virtue of synoptic view of large area in discrete bands of the electromagnetic spectrum, multi-spectral data acquired from space-borne platforms, remote sensing has been found to be useful for detecting, mapping and monitoring of soil salt content.Through the analysis of soil spectral data gained by using hyper-spectral technique, the findings show that with the decrease of NDVI the correlation index between soil salt content and soil spectral reflectance increases. And that soil spectrum is more sensitive to soil salt content over the spectral range of 451.42—593.79 ran. where diagnosis index and correlation coefficient index between soil salt content and soil spectral reflectance is larger than other spectral ranges. Based on soil spectral analysis, the study selected 17 samples with no or little vegetation cover and low soil water content to establish model for predicting soil salt content. Before using the model , to remove vegetation effect on basis of linear mixing model is to replace the spectral information of vegetation in one mixed pixel with that of non-vegetation..But whether or not mixed image cell decomposition is made, the precise of the result of the model is so low that it doesn't meet the need of the practical application. The reason that caused it is the relationship of soil salt content and image feature is not a simple linear model but a complicated non-linear model. So the study introduces an artificial neural network model for soil salt content inversion because it has a superior ability for solving the non-linear problem. In contrast to statistical model, the accuracy of the inversion is greatly improved.In the end, considering the previous study concerned and the analysis of datum of study area, it's obvious to see that the development of oriental migratory locust has an intimate relation with soil salt content .So we predict various areas of development of oriental migratory locust using the result of soil salt content inversion that the artificial neural network model produced.