论文部分内容阅读
通过卫星影像利用水稻不同物候期的特征对其进行识别是目前在水稻信息提取方面比较有效的方式。首先根据水稻区别于其他植被的显著特征,即水稻生长前期稻田的下垫面含有大量水的特性,将用于提取水域信息的归一化水体指数(normalized difference water index,NDWI)合理地应用在水稻前期的信息提取中,并且利用两个物候期的NDWI做比值,扩大了水稻与其他地物之间的差异。再借助归一化植被指数(normalized differential vegetation index,NDVI)在不同物候期的增长变化特征可以进一步提高水稻信息提取的精度。研究发现充分利用水稻在不同物候期的特殊性,并结合与水稻特性相关的指数,将NDWI和NDVI两种指数同时用于水稻提取,确定合理、准确、有效的提取方法是提高水稻提取精度的主要途径。本文以安徽省来安县的水稻为研究对象,基于2013年7月12日和8月30日获取的两幅高分一号卫星WFV影像数据,利用水稻分蘖期和抽穗期的NDVI和NDWI构建了水稻信息提取模式,有效地提取出了来安县的水稻信息分布并进行制图,最后结合在研究区野外实地考察的信息数据对提取结果进行验证和评价。研究结果表明利用该模式能够快速、准确地从遥感影像上获取水稻分布信息,具有很好的普适性。
It is an effective way to extract information from rice by utilizing the characteristics of rice in different phenophases by satellite imagery. First of all, according to the salient features of rice distinguished from other vegetation, that is, the paddy land contains a large amount of water in the early rice growth stage, the normalized difference water index (NDWI) In the pre-rice information extraction, and using the ratio of NDWI in two phenophases, the difference between rice and other features was expanded. With the help of the normalized differential vegetation index (NDVI) in different phenophase growth and variation characteristics can further improve the accuracy of rice information extraction. The study found that taking full advantage of the special characteristics of rice at different phenological stages, combined with the index related to rice characteristics, NDWI and NDVI two indices simultaneously used in rice extraction, to determine a reasonable, accurate and effective extraction method is to improve rice extraction accuracy main method. Based on the WFV image data of two WST-1 satellites acquired on July 12 and August 30, 2013, rice in Laian County of Anhui Province was used as the research object. Based on NDVI and NDWI at the tillering and heading stages of rice, The rice information extraction model effectively extracted and mapped the rice information in Laian County. Finally, the extraction results were verified and evaluated based on the field data in the field. The results show that this model can quickly and accurately obtain the rice distribution information from remote sensing images and has good universality.