Skip to Content

Non-Contact, In-Process Shape Measurement of Windscreen Glass

Overview | Current Projects | Recent Projects | Laboratory | Research Team

Project Partners

Research Engineers

  • Dr. Kynan Graves

Overview

Optical techniques have been widely used for non-contact measurement of the 3-D shapeof diffusely reflecting surfaces. However, there is no evidence for the successful implementation of a real-time shape measurement system for large specular surfaces, despite the many important industrial applications. The aim of this project is to develop optically based techniques to measure the shape of specular and transparent surfaces in real time in an industrial environment. The main outcome of the research will be a prototype on-line shape measurement system to control the quality of car windscreens.

 

Shape Mesurement using Machine Vision.

Shape Mesurement using Machine Vision.

 

Challenges

There are several challenges involved in this research project. Firstly, conventional (triangulation-based) inspection approaches are, generally, not feasible. This is due to the material properties, particularly with respect to the (specular) surface reflection characteristics and that the material is transparent to optical wavelengths. Further, in the industrial environment, the material may be vibrating and have non-uniform positioning.

Optical Table used for Initial Measurements.

Intrinsic Camera Calibration used to Remove Camera Distortion.

 

Approaches

Image processing algorithms, including intrinsic and extrinsic camera calibration and structured-highlight imaging are used to provide high-accuracy positional measurements.

Shape Mesurement using Machine Vision.

Extrinsic Camera Calibration used to locate Cameras relative to the Windscreen.