高分卫星在气候适应型试点城市绿地变化监测中的应用

绿地建设是气候适应型城市建设的重要内容之一，为客观评价建设情况，基于高分系列多光谱数据，以常德城区作为研究区域，选用包括纹理特征的26项特征变量，构建CART（Classification And Regression Tree，分类与回归树）决策树进行面向对象监督分类，使用Khat方法完成分类精度评价。精度评价数据表明，2014至2019年6年间的总体分类精度大于90%，Kappa系数大于0.8，绿地提取的用户精度与制图精度达到89%，验证了该方法在城市绿地的提取工作的上具有可行性。分析常德城区近6年的绿地变化监测结果，得出城市绿地面积从2014年的149.01km2上升至2019年的166.79km2，面积增加17.78km2，增长率11.93%，经与实际变化趋势比对一致。结果表明，基于CART决策树的面向对象高分遥感解译方法，具有一定参考价值，可为气候适应型试点城市绿地变化监测评估提供有效依据。

Application of GF Satellites in Monitoring Green Space Change in Climate-Adapted Pilot Cities: A Case Study of Changde

Green space construction is one of the important contents of climate-adaptive city construction. In order to objectively evaluate the construction situation, based on GF multispectral data, Changde urban area was used as the study area, and 26 feature variables including texture features were selected to construct a CART decision tree. Through the Khat method to complete the classification accuracy evaluation of object-based supervised classification, the accuracy evaluation showed that the overall classification accuracy is greater than 90%, the Kappa coefficient was greater than 0.8.Also the user accuracy and mapping accuracy of green space extraction reached 89% during the past 6 years from 2014 to 2019, which verified that the method is feasible in the extraction of urban green space. Analyzed the monitoring results of the green space change in Changde urban area in the past 6 years, it was concluded that the urban green space area had increased from 149.01km2 in 2014 to 166.79km2 in 2019, with an area increase of 17.78km2 and a growth rate of 11.93%. The results showed that the object-based remote sensing interpretation method based on the CART decision tree of GF data had certain reference value, which provided an effective basis for the monitoring and evaluation of green space changes in pilot cities with climate adaptation.