论文部分内容阅读
先通过几何矫正、配准、裁切等预处理获取深圳市辖区的HJ-1-A/B卫星CCD2数据,再选取城市绿地、林地、农用地等样本进行光谱特征分析,并以深圳市基本生态控制线等数据作为辅助,最后利用“决策树-监督”混合分类模式提取深圳城市绿地信息。将分类结果与深圳市生态资源测算的城市绿地面积进行对比,分析发现两类数据的Pearson相关系数值为0.9997,平均误差为8.37%;通过空间分布对比发现,两类数据解译获取的结果基本一致。结果有利于提高城市绿地信息的提取效率,进而为降低区域生态环境动态监测的成本并提高监测的时间精度提供一定的借鉴。
Firstly, the CCD2 data of HJ-1-A / B satellite in Shenzhen area were obtained by geometric correction, registration, cutting and other preprocessing, then the urban greenland, woodland and agricultural land samples were selected for spectral characteristics analysis. Ecological control lines and other data as a subsidiary, and finally use “decision tree - supervision ” hybrid classification model to extract Shenzhen urban green space information. The results of the classification are compared with the urban green area measured by the ecological resources in Shenzhen. The Pearson correlation coefficient of the two types of data is 0.9997 with an average error of 8.37%. According to the spatial distribution comparison, the results of two types of data interpretation Consistent. The results are helpful to improve the efficiency of urban green space information extraction and provide some reference for reducing the cost of dynamic monitoring of regional ecological environment and improving the time precision of monitoring.