| 摘要: |
| 【目的】冬小麦种植面积的提取对保障粮食安全和估产工作具有重要意义，已有冬小麦面积监测方法存在的所需数据量多、时间滞后等问题亟待解决。【方法】基于入冬前的2017年11月21日（分蘖期）和2017年12月24日Landsat8 OLI影像，将MIR、NIR和RED 波段进行HSV 变换，并计算地物的NDVI；利用全国土地利用图提取耕地与非耕地两类地物，统计分析两类地物NDVI值、H波段值的关系并设置阈值，初步提取疑似小麦种植区；利用小麦两个时相S值增大的特点准确提取小麦种植区域。【结果】利用多时相遥感数据中NDVI、H和S差别提取的试验区冬小麦种植面积，与地面调查、县区统计年鉴数据有较高的一致性。【结论】HSV阈值划分方法适用于冬小麦种植面积提取，能够提高小麦面积估算的时效性。 |
| 关键词: 冬小麦 面积提取 HSV 色彩模型 早期监测 遥感影像 |
| DOI：10.12105/j.issn.1672-0423.20190603 |
| 分类号: |
| 基金项目:国家自然科学基金项目“综合前期光谱和上茬作物时序遥感数据的冬小麦播期监测方法研究”（4167011560） |
|
| Area extraction of winter wheat based on HSV transformation |
|
Zhao Ye,Li Cunjun※,Zhou Jingping,Jing Xia,Jing Weibin
|
|
1.Beijing Research Center for Information Technology in Agriculture ，Beijing 100097，China;2.Xi’an University of Science and Technology，Shaanxi Xi’an 710054，China
|
| Abstract: |
| ［Purpose］The extraction of winter wheat planting area is of great significance to guarantee food security and yield estimation. The existing methods need a lot of data and time lag to be solved. ［Method］Based on the images of Landsat8 OLI on November 21，2017（tillering stage）and December 24，2017 before winter，HSV transform was performed on MIR，NIR and RED bands，and NDVI of ground object was calculated. The national land use map was used to extract the cultivated land and non-cultivated land features，and the relationship between the NDVI value and H-band value of the two features was statistically analysed and a threshold was set to preliminarily extract the suspected wheat planting area. The wheat planting area was accurately extracted with the increase of S value of two time phases.［Result］The planting area of winter wheat in the test area extracted by using the difference of NDVI，H and S in the multitemporal remote sensing data has a high consistency with the data of ground survey and county statistical yearbook.［Conclusion］HSV threshold division method is applicable to the extraction of winter wheat planting area and can improve the timeliness of wheat area estimation. |
| Key words: winter wheat area extraction HSV color model early monitoring remote sensing image |
