Calibration of a stereo rotating line camera system for sophisticated indoor panorama services

Internet portal providers have great interest in providing indoor panorama services which can be effectively utilized for many indoor geospatial applications. Such services require accurate estimation of objection points. In this study, we thus propose an efficient system to generate indoor panorama and a calibration method for the accurate determination of object points. First, we derive geometric relationship between an object point and image points on the overlapped images. Second, we define the calibration parameters consisting of camera intrinsic parameters and system parameters. We then acquire stereo panorama images using this system and estimate the calibration parameters. Finally, we analyze the effectiveness of the proposed method based on the calibration results. The experimental results show that the residuals of image points are within ±2 pixels and the error of object points are within ±1.6 cm. With the accurate parameters determined by the proposed method, we can generate precise indoor 3D models and provide indoor panorama services equipped with sophisticated user interaction.