Invention:
Deflectometry is one of the potential techniques for freeform surface metrology. The display panel is either LCD, LED, or OLD and it displays a certain pattern onto the freeform surface. The cameras then record how the pattern is distorted on the surface and this is fed into a processing unit where it calculates the surface shape. This invention involves a robust and easy to use deflectometry system for measuring freeform surfaces. The system includes three cameras, two which are positioned at the edge of the display panel and one in the middle. The processing that this invention uses consists of a trained neural network to more accurately calculate the surface shape.
Background:
The two main issues with freeform surface metrology are that systems of this nature are not robust enough and they are often very difficult to use. It’s difficult to use because the location of parts must be precise and information about the surface must be known to be able to make a precise calculation.
The system will be robust as the cameras and display panel are integrated into one unit. This also helps with ease of use because the cameras and display panel will be in the same location relative to each other. This gets rid of the problem with having to know the precise location of components. Lastly, this solution utilizes a deep learning model to obtain the surface shape based on the patterns captured by the cameras. The multiple camera approach allows for a more accurate surface capture.
Applications:
- 3D modeling
- Quality control for manufacturing
Advantages:
