The Study Of Monitoring Soil Moisture Based On Redarsat-2Remote Sensing Images And BP Artificial Neural Network

Soil moisture is an important parameter of research such as agricultural research and weather analysis. It has significant meaning in hydrological forecasts and crop water monitoring because of its accurate and fast information retrieval soil moisture on the analysis. In the practice of remote monitoring of soil moisture, the accuracy of passive microwave remote sensing monitoring is relatively low, and visible-infrared remote sensing easily affected by the weather. Active microwave remote sensing can monitoring in all-time without affecting of rain clouds and has strong penetrating ability with other characteristics, making up the drawbacks in soil moisture monitoring of passive microwave remote sensing and visible light-infrared remote sensing, providing a new method for monitoring soil moisture, becoming the most promising means of land surface soil moisture remote sensing in recent years.Based on research of active microwave remote sensing in soil moisture studies, selecting required Advance Integral Equation model (AIEM) for conditions to be studied regions, namely the qualified surface roughness of the exposed surface area, using this model simulate the relationship of backscattering coefficient and surface dielectric constant. Using the advantages of BP artificial neural network model nonlinear relationships of the data processing, whilst, compared with other inversion model, AIEM can more effectively simulate the actual surface scattering properties. Selecting AIEM simulated relational data as BP neural network training data, one can achieved using BP artificial neural network to handle the backscattering coefficient and surface soil permittivity relationship.In the conditions of known such as radar system parameters, through NEST, ENVI processing Radarsat-2active microwave remote sensing images obtained backscattering coefficient. Once Advance Integral Equation Model (AIEM) simulated data network training BP artificial neural is completed, the backscatter coefficients of four polarizations (HH,HV,VH,VV) are used as input data inversion soil permittivity, the relationship between soil dielectric constant and soil moisture was represented by a Topp dielectric model, and then getting results in soil moisture information. In this research, we used remote sensing data of Radarsat-2C-band active microwave, selecting relevant test area in Zhaokou Irrigation, inverting bare surface soil moisture information, comparing with the actual sample value; it can be applied to achieve the desired effect.