中文摘要：

在分析计算机视觉、摄影测量学相机模型的基础上，提出了多平面约束的监控视频与2D地理空间数据的几何互映射模型和面向互增强的内容互映射模型，并就其应用进行了介绍。针对几何互映射中存在的地面起伏对映射的影响以及监控视频空间分辨率等问题进行了分析。该模型适合于地面区域为多个平面的情况。研究表明，监控视频与2D地理空间数据可以实现一定精度下的互映射，能够满足视频监控系统与GIS的深度集成。

英文摘要：

In order to achieve the mutual enhancement between surveillance videos and 2D geo-spatial data, a modal including geometric and content mapping is proposed. Firstly, based on the pin-hole camera model of photogrammetry and computer vision, we assume the ground is composed of multiple planes and propose a mapping model. The mapping model has clear physical meaning, convenient and flexible. Through the model, the videos and 2D geospatial data can be mutually projected. Secondly, because the true ground is not flat, undulation has an impact on the mapping between videos and 2D geospatial data, a formula for the impact is provided. Thirdly, the ohiects in videos are larger nearby and smaller at far distances. In order to get the real space of each pixel, we propose an indicator, vide- o spatial resolution and discuss the spatial distribution. The model is suitable for multiple planes but not suitable for complex geographic scenes.