CAJ | 학술논문

针对全球范围内只能依据单季相Landsat影像数据提取耕地的现状,在现有基于影像光谱与纹理特征的监督分类方法基础上,引入NDVI时间序列数据,解决部分单季相影像造成的耕地与其他类型的混分问题.实验表明,利用NDVI时间序列数据能有效地反映耕地的物候特征,弥补使用单季相Landsat影像提取耕地的缺陷,较好地完善全球耕地的提取方案,提高了耕地提取精度.
침대전구범위내지능의거단계상Landsat영상수거제취경지적현상,재현유기우영상광보여문리특정적감독분류방법기출상,인입NDVI시간서렬수거,해결부분단계상영상조성적경지여기타류형적혼분문제.실험표명,이용NDVI시간서렬수거능유효지반영경지적물후특정,미보사용단계상Landsat영상제취경지적결함,교호지완선전구경지적제취방안,제고료경지제취정도.
@@@@According to the present situation of doing cropland classification using single-season Landsat image, this paper introduces NDVI time series data and resolves the confusion between cropland other land cover types in single-season Landsat images. Experiments show that this method could reflect the cropland’s phonological characteristics, complete the global cropland classification scheme and improve the accuracy of cropland classification.