基于高分辨率卫星遥感的城市典型地物变化监测

发布时间：2018-04-02 19:48

本文选题：高分辨率卫星遥感　切入点：变化检测　出处：《东北大学》2014年硕士论文

【摘要】：高分辨率卫星遥感应用是遥感领域近年来研究的热点,利用高分辨率遥感卫星空间分辨率高,重访周期短,敏捷性强等特点,对城市土地利用情况进行变化检测,完成土地利用情况普查,土地空间数据库更新等任务,对于国家地理国情监测,数字城市工作的开展具有重要意义。目前,还没有针对高分辨率卫星遥感影像最优的变化检测方法,而在使用一般检测方法处理高分辨率遥感影像时,会出现处理效率低下,检测结果破碎化以及变化阈值难以确定等问题。本文面向城市地区土地利用情况动态监测的需求,改进了一种基于典型相关分析方法的变化检测算法,利用典型变量来集中影像各个波段的变化信息,引入交叉窗口均值滤波以去除高分卫星影像在像元级检测时的破碎噪声,并在传统的均值-标差法固定阈值基础上提出动态闽值算法,以降低漏检测率。本文的实验部分利用GDAL等开源函数库对改进的典型相关分析变化检测算法进行了程序实现,并选取沈阳市浑南新区土地覆盖变化显著的某建设区作为研究区,使用2013年7月4日与2013年8月21日两期高分一号卫星影像进行了变化检测应用实验,并对检测结果进行了精度评价,验证了该算法的有效性。最后,实验使用ENVI Feature Extraction模块对研究区变化区域进行了基于面向对象方法的城市典型地物分类,使用ArcGIS对分类后的各种土地覆盖类型变化量与变化率进行了统计分析,并制作了研究区土地覆盖变化情况专题地图。本文针对高分辨率卫星遥感影像数据量大,细节特征丰富等特点,改进并实现了基于典型相关分析方法的变化检测算法,并验证了该算法的有效性。使用面向对象的方法对变化检测结果进行分类,最后对分类结果进行统计分析及制图。研究结果证明本文的理论在基于高分遥感影像的城市地区变化检测中具有一定的应用价值,为后续的基于高分遥感的城市典型地物全自动变化检测提供了实用方法与实践基础。
[Abstract]:The application of high-resolution satellite remote sensing is a hot research topic in the field of remote sensing in recent years. Using the characteristics of high-resolution remote sensing satellite with high spatial resolution, short re-visit period and strong agility, the change of urban land use is detected.It is of great significance to complete the land use survey and update the land spatial database for the monitoring of the national geographical situation and the development of the digital city.At present, there is no optimal change detection method for high-resolution satellite remote sensing images, but the processing efficiency will be low when using general detection methods to process high-resolution remote sensing images.The result of detection is broken and the threshold of change is difficult to be determined.In this paper, a change detection algorithm based on canonical correlation analysis is improved to meet the demand of dynamic monitoring of land use in urban areas.The cross-window mean filter is introduced to remove the broken noise in the pixel level detection of high-score satellite images, and a dynamic threshold algorithm based on the traditional mean-scalar difference method is proposed to reduce the leakage detection rate.In the experiment part of this paper, the improved canonical correlation analysis change detection algorithm is implemented by using GDAL and other open source function libraries, and a construction area with significant land cover change in Hunnan New District of Shenyang is selected as the research area.The experiment of change detection is carried out by using the satellite images of Gaofen 1 on July 4, 2013 and August 21, 2013. The accuracy of the detection results is evaluated and the validity of the algorithm is verified.Finally, ENVI Feature Extraction module is used to classify the typical urban features based on object oriented method, and ArcGIS is used to analyze the change amount and change rate of various land cover types.A thematic map of land cover change in the study area was made.In this paper, the change detection algorithm based on canonical correlation analysis method is improved and implemented, and the validity of the algorithm is verified according to the characteristics of high resolution satellite remote sensing image such as large amount of data and rich details.The object-oriented method is used to classify the change detection results. Finally, the classification results are statistically analyzed and mapped.The research results show that the theory of this paper has certain application value in urban area change detection based on high score remote sensing image, and provides a practical method and practical basis for automatic change detection of typical urban ground objects based on high score remote sensing.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：P237

【相似文献】