Color and Image Reconstruction
Fraunhofer Institute for Integrated Circuits
High-quality camera images are not a matter of luck.
Accordingly, in the area of computational imaging, our experts on sensor and camera technology work hand in hand with the scientists developing the algorithms for color and image reconstruction. The resulting knowledge synergies benefit both kinds of work: on the one hand, in-depth knowledge of the underlying hardware is a prerequisite for optimal implementation of signal processing algorithms; and on the other hand, understanding the finer points of image editing helps make informed decisions in designing the associated hardware architecture.
An excellent example of this successful collaboration is our High-speed camera . Originally designed for use in the manufacturing industry, it has also proved valuable to the film and advertising industries on account of its superior image quality. Much of this success is due to demosaicking algorithms specially developed for the camera's image sensor.
In this context, our development work centers on algorithms for
- sensor-based image correction methods (denoising, defective pixel correction)
- demosaicking of one-chip RGB and CMYG sensors
- demosaicking of special-purpose sensors (polarization, RGB color filter array or multispectral sensors)
- correction of color shifts caused by overexposure
- color calibration of single- and multi-camera systems
Set Bookmark