• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Asian conference on remote sensing;ACRS >Classification of vegetation in Japan using MODIS by machine learning method
【24h】

Classification of vegetation in Japan using MODIS by machine learning method

机译：基于MODIS的日本植被分类机器学习方法。

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

The classification of land use and land cover(LULC) from remotely sensed imagery can be conducted with two general image analysis approaches: i) pixel-based classifications, and ii) object-based classifications. While pixel-based analysis has long been the mainstay approach for classifying remotely sensed imagery. However, object-based analysis has become increasingly commonplace over the last years. Topography and geology are factors to characterize the distribution of natural vegetation. Topographic contour (elevation, slope, slop direction) determines the living conditions of plants such as soil moisture, sunlight, and windiness. The similar topographic conditions exhibit the similar distribution of vegetation unless natural disturbances or artificial disturbances being occurred. A vegetation map of Japan was developed using an object-based segmentation and pixel-based approach, included in machine learning package, with topographic information and climate information (monthly average temperature, monthly average precipitation) with MODIS remotely sensed data. The results of four methods were compared: i) object-based considering climate information ii) pixel-based considering climate information ni) object-based not considering climate information, and iv) pixel-based not considering climate information. Through the comparison, the object-based classification is more effective to produce a vegetation map than the pixel-based classification. In addition, the classification accuracy with considering climate information was higher than not considered. In a study on areas with elongated topographical features like Japan, it is necessary to consider horizontal distribution due to the latitude of vegetation and vertical distribution due to altitude.
机译：可以通过两种通用的图像分析方法对遥感影像中的土地利用和土地覆盖（LULC）进行分类：i）基于像素的分类，以及ii）基于对象的分类。尽管基于像素的分析长期以来一直是对遥感图像进行分类的主要方法。但是，基于对象的分析在最近几年变得越来越普遍。地形和地质是表征自然植被分布的因素。地形轮廓（海拔，坡度，坡度方向）决定了植物的生存条件，例如土壤湿度，阳光和风。除非自然干扰或人为干扰发生，否则相似的地形条件显示出相似的植被分布。日本的植被图是使用基于对象的分段和基于像素的方法开发的，该方法包括在机器学习包中，带有地形信息和气候信息（月平均温度，月平均降水量）以及具有MODIS遥感数据。比较了四种方法的结果：i）考虑气候信息的基于对象的信息; ii）考虑气候信息的基于像素的信息; ni）不考虑气候信息的基于对象的信息;以及iv）不考虑气候信息的基于像素的信息。通过比较，基于对象的分类比基于像素的分类更有效地生成植被图。此外，考虑气候信息的分类准确性高于未考虑的分类准​​确性。在对诸如日本这样具有长形地形特征的区域进行的研究中，有必要考虑植被的纬度引起的水平分布和海拔高度引起的垂直分布。

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