南极地区Landsat8卫星影像的阴影检测方法
发布时间：2018-11-20 21:00
【摘要】：阴影在光学遥感影像中普遍存在,会对后续一系列图像信息提取等过程产生影响.在南极地区,由于其特殊的地理位置和环境,光学遥感影像中存在的阴影不可忽视.本研究在已有的阴影检测方法的基础上,根据Landsat8卫星数据的特点构建了一个新的阴影检测方法:首先对所有数据进行反射率转换处理,利用南极地区成像时的阴影在各波段的反射率特征,选取第1波段Coastal/Aerosol(0.43~0.45μm)和第5波段NIR(0.85~0.88μm)建立了能够有效识别阴影的阴影检测指数R_(51),并选取不同地物的训练样本进行统计分析,根据阈值提取法提取阴影为二值化图像,并从影像中剔除少量易与阴影混淆的冰面融池.通过选取南极不同地区的影像进行精度评价,发现该方法针对Landsat8数据在南极地区的阴影检测具有较高的提取精度和较好的普适性.本文利用阴影检测方法对收集的覆盖南极地区的1 127景Landsat8影像进行阴影提取与统计、分布特征分析,最终提取阴影面积约15 000km~2,主要分布于裸岩山体集中区和大型冰川所在地.
[Abstract]:Shadow is widely used in optical remote sensing images, which will affect a series of subsequent image information extraction process. In the Antarctic region, because of its special geographical location and environment, the shadow in optical remote sensing image can not be ignored. Based on the existing shadow detection methods, a new shadow detection method is constructed according to the characteristics of Landsat8 satellite data. Based on the reflectivity characteristics of shadows in the Antarctic region, the shadow detection index R _ (51), which can effectively identify shadows, is established by selecting the first band Coastal/Aerosol (0.430.45 渭 m) and the 5th band NIR (0.85 0.88 渭 m). The training samples of different ground objects were selected for statistical analysis, and the shadow binary image was extracted according to the threshold extraction method, and a small amount of ice melt pool which was easily confused with shadow was removed from the image. It is found that this method has higher extraction accuracy and better universality for shadow detection of Landsat8 data in Antarctic region by selecting images from different regions of Antarctica for accuracy evaluation. In this paper, the shadow detection method is used to extract and calculate the shadow of 1 127 Landsat8 images covering the Antarctic region, and the distribution characteristics are analyzed. Finally, the shadow area is about 15 000 km / m ~ 2, which is mainly distributed in the bare rock mountain concentration area and the location of large glacier.
【作者单位】： 北京师范大学全球变化与地球系统科学研究院北京师范大学极地研究中心;
【基金】：国家重点基础研究发展计划资助项目(2012CB957704)
【分类号】：P237
本文编号：2346008
[Abstract]:Shadow is widely used in optical remote sensing images, which will affect a series of subsequent image information extraction process. In the Antarctic region, because of its special geographical location and environment, the shadow in optical remote sensing image can not be ignored. Based on the existing shadow detection methods, a new shadow detection method is constructed according to the characteristics of Landsat8 satellite data. Based on the reflectivity characteristics of shadows in the Antarctic region, the shadow detection index R _ (51), which can effectively identify shadows, is established by selecting the first band Coastal/Aerosol (0.430.45 渭 m) and the 5th band NIR (0.85 0.88 渭 m). The training samples of different ground objects were selected for statistical analysis, and the shadow binary image was extracted according to the threshold extraction method, and a small amount of ice melt pool which was easily confused with shadow was removed from the image. It is found that this method has higher extraction accuracy and better universality for shadow detection of Landsat8 data in Antarctic region by selecting images from different regions of Antarctica for accuracy evaluation. In this paper, the shadow detection method is used to extract and calculate the shadow of 1 127 Landsat8 images covering the Antarctic region, and the distribution characteristics are analyzed. Finally, the shadow area is about 15 000 km / m ~ 2, which is mainly distributed in the bare rock mountain concentration area and the location of large glacier.
【作者单位】： 北京师范大学全球变化与地球系统科学研究院北京师范大学极地研究中心;
【基金】：国家重点基础研究发展计划资助项目(2012CB957704)
【分类号】：P237
【相似文献】
相关期刊论文 前3条
1 周家香;朱建军;张红亚;;基于C1C2C3彩色不变特征的阴影检测与辐射恢复研究[J];工程勘察;2007年06期
2 虢建宏,田庆久;一种IKONOS影像阴影检测与去除的方法[J];遥感信息;2004年04期
3 ;[J];;年期
相关硕士学位论文 前1条
1 王军利;基于彩色航空影像的阴影检测算法研究[D];武汉大学;2005年
，本文编号：2346008
本文链接：https://www.wllwen.com/kejilunwen/dizhicehuilunwen/2346008.html
