...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Journal of Forest Planning >Deforestation Detection in Kinabalu Area, Sabah, Malaysia by using Multi-sensor Remote Sensing Approach
【24h】

Deforestation Detection in Kinabalu Area, Sabah, Malaysia by using Multi-sensor Remote Sensing Approach

机译：多传感器遥感技术在马来西亚沙巴京那巴鲁地区的森林砍伐检测

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

This paper examines use of multi-sensor remote sensing approach for deforestation detection in the tropics. Multi-sensor satellite data of Landsat-MSS of 1973 and Landsat -TM of 1991 and 1996 were employed. Accuracy of image-to-image registration wasbelow 1 pixel. Relative radiornetric normalization of Landat-MSS 1973 and Landsat-TM 1991 to Landsat-TM 1996 as the reference image was carried out to remove the unwanted variabilities between all the satellite images. Image differencing algorithm with Normalized Difference Vegetation Index (NDVI) was examined for deforestation detection. The performance of the NDVI image differencing algorithm for deforestation detection between 1973 and 1996 was investigated at three test sites covered with reliable ground truths. The accuracy of detection was satisfactory that the algorithm was used in deforestation detection of the whole study area in two change periods i.e. I: 1973-1991 and II: 1991-1996. Although false deforestation pixels in period I were also detected, it can easily be rectified using a land use map of 1984. In total, 2,445 ha of forest, which is almost 1% of the study area, were cleared from 1973-1996 and most of them were deforested in period I (2,090ha). This study concludes that the multi-sensor approach is a useful solution for deforestation detection because of better temporal coverage. It can also provide more satellite data for the application and thus lessen data acquisition problem due to cloud cover which is a consistent problem for the tropics.
机译：本文研究了在热带地区使用多传感器遥感方法进行森林砍伐检测的情况。使用了1973年的Landsat-MSS和1991年和1996年的Landsat-TM的多传感器卫星数据。图像到图像配准的精度低于1个像素。进行了Landat-MSS 1973和Landsat-TM 1991到Landsat-TM 1996的相对无线电标准化，以作为参考图像，以消除所有卫星图像之间的不想要的变化。检查了具有归一化植被指数（NDVI）的图像差异算法以进行森林砍伐检测。在覆盖可靠地面实况的三个测试地点对1973年至1996年之间NDVI图像差分算法在毁林检测中的性能进行了研究。检测的准确性令人满意，该算法在两个变化时期即I：1973-1991和II：1991-1996被用于整个研究区域的森林砍伐检测。尽管在第一阶段中也发现了错误的毁林像素，但可以使用1984年的土地利用图轻松纠正。在1973-1996年期间，总共清理了2445公顷森林，几乎是研究面积的1％，大部分其中一期（2,090公顷）被砍伐。这项研究得出的结论是，由于具有更好的时间覆盖范围，因此多传感器方法是用于森林砍伐检测的有用解决方案。它还可以为应用程序提供更多的卫星数据，从而减轻由于云层覆盖而引起的数据采集问题，这对热带地区来说是一个持续存在的问题。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号