摘要: |
The research described in this report is the fifth stage of an ongoing programme of research with the objective of developing a more objective and repeatable bridge inspection procedure than the current system of manual inspections. The proposed inspection procedure is based on the collection and processing of images of structures to identify defects in highway structures. The research has been funded by, and performed on behalf of, the Transport Research Foundation (TRF). This report describes changes to the image collection system, including the inclusion of a robotic mount system. This requires very little human input other than to perform a brief calibration to determine the camera field of view, and the extent of the scene to be imaged. This dramatically speeds up the image collection procedure which was previously very labour intensive. Possible changes to the manner in which the data is presented to the end user are also described. The use of 3-D modelling and display techniques is discussed, and several pieces of proprietary software are investigated. These software packages can manipulate and display LIDAR data. This data can be converted into 3-D models of the bridge. The 3-D bridge model can be further enhanced by texturing, using the high resolution images collected by the imaging system. The report also discusses the use of a web-based service which takes multiple images of a scene and uses these to create a 3-D model of the scene or object which was imaged. The advantages and disadvantages of using the web-based system as a key part of the inspection regime are discussed. The report concludes by outlining the key components required for a successful automated bridge inspection system based on images, which could provide inspection data at a level comparable to that currently provided by a standard Highways Agency General Inspection. |