【摘要】：耕地是全球地表覆蓋制圖中重要的地表覆蓋類型之一,其變化影響著人類社會經(jīng)濟發(fā)展、糧食安全與生態(tài)環(huán)境保護.現(xiàn)有全球耕地制圖產(chǎn)品空間分辨率較粗,缺乏高空間分辨率的全球耕地數(shù)據(jù)產(chǎn)品.本研究利用全球2000/2010兩期30m空間分辨率遙感影像數(shù)據(jù)集(Landsat TM/ETM+、HJ-1)、MODIS 250m空間分辨率NDVI時間序列數(shù)據(jù)及多種參考資料,針對全球尺度30m影像中耕地提取的難點,提出基于像元、對象和知識3個層次的耕地提取方法,即基于像元尺度多特征優(yōu)化的耕地分類提取、基于對象的耕地自動判別以及基于信息服務和先驗知識的交互式對象處理,完成了全球兩期30m耕地遙感制圖,并對全球耕地的面積等信息進行統(tǒng)計分析.結(jié)果表明,全球2000年和2010年耕地總面積分別為19.03億ha和19.60億ha.精度評價結(jié)果表明,兩期全球30m耕地遙感制圖總體精度均達到92%以上.本研究研制的2000/2010兩期的全球耕地遙感數(shù)據(jù)產(chǎn)品,在空間分辨率和分類精度上均優(yōu)于國際上同類產(chǎn)品,為全球糧食安全、生態(tài)環(huán)境監(jiān)測和全球變化等研究提供了重要基礎(chǔ)數(shù)據(jù).
[Abstract]:Cultivated land is one of the most important land cover types in global land cover mapping. The change of cultivated land affects the development of human society and economy, food security and ecological environment protection. The existing global cultivated land mapping products have coarse spatial resolution and lack of global cultivated land data products with high spatial resolution. In this study, we used Landsat TM/ETM (HJ-1), MODIS 250m spatial resolution NDVI time series data and various reference data) in the global 2000 / 2010 two periods of 30m spatial resolution remote sensing image data, aiming at the difficulty of cropland extraction in 30m image of global scale. A method of farmland extraction based on pixel, object and knowledge is proposed, which is based on multi-feature optimization of pixel scale, automatic discrimination of cultivated land based on object and interactive object processing based on information service and prior knowledge. Two periods of remote sensing mapping of 30m cultivated land have been completed, and the global cultivated land area and other information have been statistically analyzed. The results show that the total cultivated land area in 2000 and 2010 is 1.903 billion ha and 1.96 billion ha., respectively. The results of precision evaluation show that the overall accuracy of the global remote sensing mapping of 30m cultivated land in the two periods is over 92%. The global cultivated land remote sensing data products developed in 2000 / 2010 are superior in spatial resolution and classification accuracy to similar products in the world, so they are global food security. The study of ecological environment monitoring and global change provides important basic data.
【作者單位】： 北京師范大學地表過程與資源生態(tài)國家重點實驗室;國家基礎(chǔ)地理信息中心;安徽理工大學測繪學院;
【基金】：國家高技術(shù)研究發(fā)展計劃項目(編號:2009AA122001) 北京師范大學地表過程與資源生態(tài)國家重點實驗室項目(編號:2013-RC-02) 國家自然科學基金項目(批準號:41301352)資助
【分類號】：S127;F323.211

【相似文獻】

相關(guān)期刊論文 前2條

1 牟炳友;趙文茹;;瑪納斯墾區(qū)農(nóng)業(yè)生產(chǎn)地域?qū)ｉT化遙感制圖方法[J];安徽農(nóng)業(yè)科學;2011年01期

2 ;[J];;年期

，

本文編號：2348017