To verify the reliability of the algorithm for speculating physical variable from 1km satellite data, it is necessary to use the ground observational data to calculate physical value in a wide survey unit (1km × 1km) . Since it is difficult to conduct the ground observation within whole wide survey unit in short time, the minimal survey area which can represent a wide survey unit has to be determined. The research is armed to find the minimal representing area of a wide survey unit through the car survey and radio-controlled helicopter survey.
Flat and homogeneous grassy plain in Mongolia is chosen as survey land. In case of car survey, 100 observational points are measured. VCR (Vegetation Coverage Ratio) values are calculated from spectrum information and used to compute the moving-averages of VCR. With the increasing of point number, the standard variation of moving-averages becomes smaller and smaller. When the variation approximately disappears after N points, the moving-average of VCR is considered as the observational value which can represent the value in wide survey unit. The total area size of N points at this time will be the minimal representing area of the wide survey unit.
Moreover, the result from the car survey may be obtained occasionally from ground surface in same situation. To confirm the result, additional observational data are taken from higher sky using radio-controlled helicopter. Each observational point has about minimal representation area. The standard variation of moving-averages in this case is very small. So it can be said that the method used in the research is a good way to calculate the minimal representing area of a wide survey unit.
View full abstract