High Quality Street-View Image Acquisition System And Application

As the development of technology in virtual reality, fast modeling city in three-dimonsional (3D) are gains more attentions. Modeling city by capturing images has two features: one is the tremendous data scale and time cost, the other is the requirements of more and more precise data. However, there is not a satisfied solution which meets both speed and precision.In this paper we propose a high-quality image acquisition processing system to support fast 3D city modeling. The system consists of high-resolution CCD camera, panoramic camera and high-precision GPS with differential technique, so it is able to acquire images of urban buildings automatically and generate digital orthophoto map for buildings.The pre-adjust step is applied for the hardware devices to obtain higher quality images; and moving path is planned to maximize productivity. Digital orthophoto maps are generated with the acquired images. In comparison with the precious method, the method introduced in this thesis brings great improvement in efficiency without quality reduction. Techniques like Auto Group Image with GPS Data (AGIGD), Auto Image Enhancement (AIE), Auto Find Ground Axis Frame (AFGAF) were used to improve efficiency of system, and avoid inconsistency of manual process.At the end of this thesis, we show the results of the system and introduced the evolution and data acquisition log.