论文部分内容阅读
全球变化背景下,准确获取森林覆盖是监测森林资源动态、实现林业可持续发展的重要基础。为将省级尺度森林资源清查面积资料空间化,以黑龙江省为例,利用1999-2003年该省森林资源清查面积数据,结合2000年500 m分辨率的MODIS数据,构建了基于阈值分割的森林类型遥感识别方法。该方法利用不同地表覆被类型归一化植被指数时间序列的季节分异特征,以森林资源清查面积为标准,设定森林类型的划分阈值,识别了黑龙江省森林类型的空间分布。最后,基于分层随机抽样和精度评价方法,表明森林类型识别结果与地面参考数据具有较高的一致性,总体分类精度为78.1%;特别是季节特征明显的落叶林,精度可达80%以上。本文所构建的方法可将森林清查统计数据进行准确的空间定位,同时结合多期森林资源连续清查资料和遥感信息,可为识别并量化区域生态系统生物量和碳库变化等提供科技支撑。
In the context of global change, accurate access to forest cover is an important basis for monitoring the dynamic changes of forest resources and achieving sustainable forestry development. In order to spatialize the inventory data of forest resources at the provincial level, taking Heilongjiang Province as an example, using the area of inventory of forest resources in the province from 1999 to 2003 and the MODIS data of 500 m in 2000, a threshold-based forest was constructed Types of remote sensing identification methods. The method uses seasonal differentiation of time series of normalized vegetation index for different types of land cover. Based on the inventory of forest resources, the classification threshold of forest types is set, and the spatial distribution of forest types in Heilongjiang Province is identified. Finally, based on the stratified random sampling and the accuracy evaluation method, it shows that the forest type identification results have high consistency with the ground reference data, and the overall classification accuracy is 78.1%; especially the deciduous forests with obvious seasonal characteristics can reach more than 80% accuracy . The method proposed in this paper can accurately locate the forest inventory statistics and provide scientific and technological support for identifying and quantifying the changes in the biomass and carbon stocks of the regional ecosystems in combination with the continuous inventory of forest resources and remote sensing information.