Design, development and application of a satellite-based field monitoring system to support precision farming

The factual base of precision agriculture (PA) - the spatial and temporal variability of soil and crop factors within or between different fields has been recognized for centuries. Field information on seeding suitability, soil & crop nutrition status and crop mature date is needed to optimize field management. How to acquire the spatially and temporally varied field parameters accurately, efficiently and at affordable cost has always been the focus of the researches in the field. Satellite remote sensing has held out much promise for within & between-field monitoring, along with the promising development regarding spatial, temporal and spectral resolution in the last decade. Scientists from all over the world have provided a great deal of fundamental information relating spectral reflectance and thermal emittance properties of soils and crops to their agronomic and biophysical characteristics. This knowledge has facilitated the development and use of various remote sensing methods to detect spatially and temporally varied environmental stresses which limit crop productivity. This can make significant contribution in optimizing crop management as sowing, irrigation, fertilization and harvest. However, gathering, accessing, and processing of remote sensing images from different satellites require high technical skills, not mention the time consumed in processing large amount of images. The lack of comprehensive software platforms to extract useful spatially and temporally varied information from satellite image hindered the wide application of satellite image to support PF. With this back ground, an integrated satellite-based field monitoring system was designed and developed with .Net and IDL (Interactive Data Language). The system consists of 4 primary functional models: 1) satellite image pre-processing model; 2) field seeding suitability evaluating model; 3) soil & crop nutrition status monitoring model and 4) crop mature date predicting model. In the first model, remote sensing images from different sensors can be pre-processed with format conversion, radiation calibration, atmospheric correction and geometric correction. BRDF correction was also provided for images with wide swath. Fusion of images from different sensors can also be implemented in this model to provide images with both high spatial and high temporal resolutions. In the second model, soil moisture and surface temperature will be acquired from satellite images. Together with the information on needs of different crops in seeding, seeding suitability of different fields can be evaluated. In the third model, the soil and crop nutrition status (nitrogen and chlorophyll concentration for crop; available nitrogen and organic matter content for soil) will be mapped with satellite image, and then been transferred to field/pixel scale fertilization prescriptions. In the fourth model, crop canopy/leaf water and chlorophyll content will be quantitatively mapped, along with the digital expression of crop water and chlorophyll content variation at maturing stage, crop mature date will be predicted 20days before the harvest, and will updated with the approaching of crop maturity. The structure, methods and the development of the system are introduced in detail in this paper. A case application of the system in ShuangShan Farm in Northeast China will also be presented as result of the system.