An information-enriched digital twin for pipe systems is valuable for facility management and maintenance. Currently, the process of creating a twin of pipes requires much human effort and the available information is also very limited. In this paper, we propose a novel approach that can automatically create a geometric digital twin of pipes from laser scanned point cloud and enrich it with information extracted from photos, such as fluid type, fluid direction, etc. In our method, text detection is applied to extract information on pipes in 2D images. Then photogrammetric point clouds are reconstructed and further used to map the extracted information from 2D camera photos into the 3D laser scanned point cloud. The final information-enriched digital twin contains geometric information, semantic information, along with all information included in code-compliant pipe labels. We compare the output of our approach with the manually created model to prove our approach has convincing performance in pipe detection, pipe reconstruction and the recognition of useful information from labels.