摘要
从林业行业应用出发,针对国产GF-1卫星数据的特点,采用决策树分类方法对影像采伐信息进行提取,探索了国产GF-1卫星数据在森林资源采伐信息提取方面的应用关键技术,为国产卫星数据的行业推广应用奠定理论基础。研究结果表明:决策树分类方法能够更有效地对伐区图斑分类提取,经过Kappa分析,测试区伐区影像提取精度从大到小分别为:水域分类精度96.71%,森林分类精度89.79%,伐区分类精度82.49%,公路分类精度75.30%,农田分类精度71.18%,建筑分类精度56.26%,其他分类精度53.57%。总体分类精度为87.34%。
Based on the application of forestry industry,based on the characteristics of domestic GF-1 satellite data,the decision tree classification method is used to extract image harvesting information,and the key technology of domestic GF-1 satellite data in forest resource harvesting information extraction is explored to lay a theoretical foundation for the promotion and application of domestic satellite data.The research results show that the decision tree classification method can extract the cutting pattern more effectively.After Kappa analysis,the accuracy of image extraction in the test area is from large to small;the classification accuracy of the waters is 96.71%;the forest classification accuracy is 89.79%;the cutting area classification accuracy is 82.49%;the highway classification accuracy is 75.30%;the farmland classification accuracy is 71.18%;the building classification accuracy is 56.26%;and the other classification precision is 53.57%.The overall classification accuracy is 87.34%.
作者
吴雪娇
王威
王芳
Wu Xuejiao;Wang Wei;Wang Fang(Academy of Forestry and Grassland Investigation and Planning,Beijing,100714,China)
出处
《绿色科技》
2019年第14期225-227,227,共3页
Journal of Green Science and Technology
关键词
决策树分类
GF-1卫星影像
森林采伐
信息提取
decision tree classification
GF^1 satellite imagery
forest harvesting
information extraction
