The Sar Image Preprocessing And Parameter Inversion Platform Design And Implementation Of Wheat

SAR is a high-resolution imaging radar featuring all-time and penetration. It can work all-time regardless of weather, so it has been applied in the field of earth observation widely. With the guradual deterioration of the environment and frequent natural disasters, people have to monior the disaster situations to develop program of responding to disaster and earth protection policy. SAR can provide continuous, large-scale, multi-frequency and multi-polarization data, from which we can extract the rich information of ecological environment and disaster. As one of the four major food crops, we need to monitor the growing state of wheat by computing the wheat parameter by using SAR image.Backscattering is an important parameter coefficient in the microwave remote sensing. Backscattering coefficient of wheat was obtained from three different methods in this study:calibration of ASAR image, MIMICS model simulation and direct measurement by the scatterometer. By comparing three kinds of backscattering coefficient, we found that the values from calibration of ASAR image and direct measurement were similar, and the simulation value of MIMICS model was almost equal to both of them.Firstly, we introduced principles of radar imaging as well as a series of concepts which were helpful to read image header and process data. We also studied on data formats of3kinds of SAR images, ASAR, PALSAR and GeoTIFF (RADARSAT) and then got their image calibration formulas.Secondly, we explored how to measure backscattering coefficient by using L S、 C and X land-based microwave scatterometer and its modification in Qionglai, Sichuan. The results from regression analysis showed that polarization differences were sensitive to wheat parameters. Based on the results, we built empirical formulas between LAI and backscattering coefficient under different angles.Thirdly, we developed a visual application to simulate backscattering coefficient after encapsulation of MIMICS, then compared simulation value and measured value. We also tested the MIMICS model and established corresponding application conditions.At last, we developed a SAR image pre-processing system to realize input and displaying of3kinds of SAR images. Logarithmic transformation and histogram equalization were used to strengthen the SAR images in low gray value without pre-processing. In order to get target SAR images, we should compress the gray-scale and size of images and then cut out ROI. Finally, we generated interference image and studied on retrieval of wheat parameter.