Abstract:

【Objective】 According to disadvantage of traditional method combing remote sensing and sampling method which only obtains the regional area, this paper brought up a new method for optimizing wheat distribution with the whole regional area.【Method】 Incorporating the advantages of soft and hard classification, the wheat extent can be divided into three parts, pure, mixed and non-wheat pixels. Based on the regional winter area, the threshold was set and the winter area was obtained, which improved not only the regional accuracy but also spatial distribution accuracy. 【Result】 The result shows that RMSE of proposed method is 0.13, which is better than hard classification result and soft classification result, which is 0.16 and 0.20, respectively. All above verifies the proposed method is effective. 【Conclusion】 Winter wheat area method with regional area control provides a good basis for the large scale crop planting area survey.

Key words: regional area')">regional area, spatial distribution optimization, hard classification, soft classification, winter wheat, remote sensing