２時期の高分解能SARデータを用いた水田利用形態の分類

抄録

This study was aimed at developing the classification methodology for the planting condition of paddy fields utilizing two time-series data acquisitions of the high-resolution SAR images of TerraSAR-X satellite during the rice transplanting and growing periods. We investigated whether it was possible to conduct classification of the paddy fields for the earlier transplanting of rice, the usual transplanting of rice, the later transplanting of rice, the direct seeding and also for the soybean and wheat crops.
The methodology performed the cluster analysis for the study area using two SAR images of the rice transplanting period and the rice growing period. At first, we created polygons for each agriculture field. The classification was conducted for the earlier transplanting of rice, usual transplanting of rice, later transplanting of rice, direct seeding of rice, and for the soybean and wheat crops by the ratio for pixels in each cluster inside a polygon.
In the proposed classification methodology, we were able to extract the direct sowing the soybean and wheat crops in the study area. The transplanting of rice (earlier, usual and later periods), wheat and soybean were clearly classified by the developed classification methodology. In addition, we understood that each field was not misclassified if there was a difference of the transplanting of about 20 days between earlier transplanting of rice and later transplanting of rice.