Yoshiki Kageyama
Examination of the paddy rice growth history extraction by the distinction of fixture farmland using time series MODIS data
Atsushi Rikimaru Kenta Sakata
This research aims at extraction of the paddy rice growth information on the time series in Japan using the MODIS data which is observing the large area by high frequency. In in Japan, it is assumed by the method of extracting the unique region of fine weather first like [ paddy rice growth time has many cloud obstacles, and ] previous research that a clear region can hardly extract. Then, in order to extract as many clear regions as possible, the effect by combining two satellites of Aqua and Terra was examined, and the data extraction method of the farmland view domain was also aimed at examination of a solution supposing the state with many clouds. This research acquired the satellite imagery and extracted the fixture farmland domain from the characteristic of change of the near-infrared (NIR) reflectance of a fixture farmland domain, and the reflectance of red (R). Furthermore, the cloud index picture was combined and the extraction in consideration of cloud of the history information of a fixture farmland domain was tried. In this research, the earth surface reflectance data of the MODIS sensor carried in the earth observation satellite, and Band1 and Band2 were used. There is the feature of change in the near-infrared rays (NIR) of a fixture farmland domain, and the value of red (R). In order to transplant rice seedlings from the end of April to the middle of May and for a growth phase to come around in mid-July, the difference (NIR-R) value of near-infrared rays and red increases. The time of mowing will start from the end of August, and a value will decrease at the end of September (NIR-R). The distinction picture of fixture farmland which took cloud into consideration using the fixture farmland distinction picture and cloud distinction picture which were created was created. The extraction range was defined and the average of the NDVI value of area and the NIR-R value which deducted the cloud distinction domain in fixture farmland within the limits was extracted. It was shown clearly that time series change of the vegetation in each fixture farmland domain in 2013 Kanto and Niigata Prefecture corresponds compared with transition of the vegetation of a cultivation calendar. The cloud obstacle was reduced by comparing and choosing a picture without the influence of cloud from both the pictures of the observation on the same day of the artificial satellites Aqua and Terra, and lack of sampled data has been improved. The influence of cloud was able to be eliminated in May 1, 2013 to August 31, the paddy rice fixture farmland domain was able to be extracted, and a time series growing condition change has been grasped for every partial area. As a result of carrying out comparison verification of the growing condition change in a cultivation calendar and satellite analysis, transition of the amount of vegetation of fixture farmland corresponded mostly at the every place point.
Please use "previous botton" to return the previous page