Research on optimization of crop planting area estimation from remote sensing data

Abstract

It was very important to estimate and forecast the crop production for acquiring the crop planting area and spatial distribution quickly and accurately. As an important method, remote sensing images are usually interpreted to acquire the crop planting area. However, for the reasons of spatial resolution of remote sensing images, many small features and linear features were interpreted into crop, which surely affected the final crop planting area. Thus, subtracting the small features and linear features became a key method to improve the accuracy of crop planting area. In this study, the Zhaodong City of Heilongjiang Province was selected as a study area, and the GF-1 image with 16 meter resolution was used as data source. The maize planting area of interpretation in the study area was optimized by subtracting the linear features using buffer zone method and small features using sampling frames. In order to get the exact maize planting area, three steps were used in this study. Firstly, the maize planting area was interpreted by the object-oriented classification method according to the GF-1 image. Secondly, linear features were subtracted by remote sensing interpretation and spatial calculation of Geological Information System (GIS), and small features were subtracted according to the 7 sampling frames. Thirdly, maize planting area of interpretation in the Zhaodong City were accurately calculated by subtracting the small features and linear features. Four important conclusions were got in this study, which were list as follows. (1) The maize planting area of interpretation in the Zhaodong City was 2199.71km <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> , which mainly distributed in the eastern and southwest region. In the towns of Zhaodong City, Liming Town had the biggest area of maize planting area (153.44km <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> ), and Honghe Town had the biggest percent of maize planting area (88.00%), while Laozhou Town had the smallest area and percent of maize planting area. (2) Linear features in the study area mainly contained road, railway, channel, river and forest belt. Linear features had major effect on the maize planting area of interpretation in the Zhaodong City. There were 70.91km <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> liner features were interpreted into maize, which was taken up 3.22% of maize planting area of interpretation in the study area. (3) Compared with linear features, the small features had less effect on the maize planting area of interpretation in the Zhaodong City. There were 8.36 km <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> small features, which were interpreted into maize. The small features was just taken up 0.38% of maize planting area in the Zhaodong City. (4) The finial maize planting area in the Zhaodong City was 2120.44 km <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> after subtracting the small features and linear features. The error of maize planting area was from 3.26% to 0.46% in the Zhaodong City, and it was from 7.13% to 4.30% for seven sampling frames. Therefore, subtracting the small features and linear features could obviously improve the accuracy of interpreted maize planting area. This study were not only helpful for improving accuracy of crop planting area, but also could supply the researching thoughts and references for other regions.