Automatic 3D Surface Inspection of Sewer Pipes Using 3D Image Processing

The focus of this research project is the automatic 3D reconstruction of sewer pipes.

Image sequences are acquired with a mobile robot inside a sewer pipe using a wide angle or fisheye lens. The images are used to estimate the pose of the camera inside the pipe directly from the images without markers. Dense multi-stereo matching techniques are used to estimate the 3D geometry of the sewer pipes and detect bursts and irregularities inside the pipes.

Our focus of research is the robust camera pose estimation and the estimation of dense depth maps from very long image sequences. Our matching approach has to handle heavy illumination changes and the particular fisheye geometry of the lens system.

Our industry partner, IBAK Helmut Hunger GmbH & Co. KG, is a leading producer of sewer pipe inspection systems.

Image Image

The project is funded by:

Please contact Prof. Dr.-Ing. Reinhard Koch for more information on the IBAK project.

Created by tos. Last Modification: Monday 11 of June, 2012 09:43:43 CEST by sandro.