Reduced-complexity data acquisition system for image-based localization in indoor environments

摘要: Image-based localization has important commercial applications such as augmented reality and customer analytics. In prior work, we developed a three step pipeline for image-based of mobile devices in indoor environments. the first step, generate 2.5D georeferenced image database using an ambulatory backpack-mounted system originally 3D modeling Specifically, create dense point cloud polygonal model from side laser scanner measurements backpack, then use it to depthmaps by raytracing model. second query is matched against retrieve best-matching image. final pose recovered with respect Since recovery only requires sparse depth information at certain SIFT feature keypoints image, this paper improve upon our previous method calculating values these keypoints, thereby reducing required number sensors data acquisition system. To do so, modified version classic multi-camera scene reconstruction algorithm, eliminating need expensive geometry range scanners. Our experimental results shopping mall indicate that proposed reduced complexity depthmap approach nearly accurate map method.