针对洪水灾害分析在速度、准确性和及时性等方面的需求,该文在研究空间自相关分析的基础上,提出了一种基于局部自相关统计的洪水灾害影像分析方法。首先,对影像进行掩膜处理,去除云层干扰;其次,采用局部空间统计的方法对影像进行统计分析;然后,通过密度分割的方法提取水体,将影像分为水体和陆地两类;最后,将3幅影像分类的结果进行空间叠加分析,分析洪水灾害影响情况。以2013年嫩江流域3个时期的影像为实验数据,设计了仿真实验。实验结果表明,该方法可以较准确地对大面积洪水影响区域进行分析。 In view of the demand of flood disaster analysis in terms of speed, accuracy and timeliness, this paper proposes a method of image analysis of flood disaster based on local autocorrelation statistics based on the study of spatial autocorrelation analysis. Firstly, the image is masked to remove the cloud interference. Secondly, the statistical analysis of the image is performed by the method of local space statistics. Then, the water body is extracted by the density segmentation method, and the image is divided into water body and land. Finally, The results of three image classification were analyzed by spatial superposition to analyze the impact of flood disasters. Taking the images of three periods in Nenjiang River Basin in 2013 as experimental data, a simulation experiment was designed. Experimental results show that this method can accurately analyze the impact of large-scale floods.