Tunable Sequential Filter for Color Image Acquisition and Reconstruction

Technology

Si-based image sensors are the industry standard for most imaging applications. Color imaging is usually achieved by using a color filter array that enables to capture in a single frame the red, green and blue components of the image. Recent improvements in the capturing rate of image sensors and in the image processing capabilities are used for capturing several full frames, each with a different color scheme. Fine control of the dynamic filter spectral response and unique fusion technology of the images provide novel image sensors with improved performance and capabilities.

Need and Advantages

Image sensors are widely in use in numerous kinds of applications. Color imaging is the common implementation since the color information is needed for most uses. The pixel count race seems to be over mainly due to the diffraction limit barrier. Thus, sacrificing resolution for getting the color information is now an issue. Usually, the common Bayer color filter array followed by de-mosaicing algorithms requires a significant penalty in spatial resolution. In addition, by acquiring the different colors sequentially provides also the capability to control the spectrum of each color filter and also to overcome blur and pixel noise effects.The SIS (Smart Image Sensor) approach addresses these problems with the tunable sequential color filtering scheme, in which each color frame is taken in full sensor resolution but in a fraction of the total acquisition time period. Then, the color frames are fused into a color image.The tunable sequential filter offers key advantages, among them:• Full sensor resolution usage. Comparing with a Bayer scheme, for green the sensor provides 2 times more resolution, and for red and blue 4 times more resolution• Flexible transmission spectra, including wide Bayer-like to reduce number of exposures or alternatively many short narrow band exposures for increasing the wavelength resolution• Adaptive exposure time periods for different exposures for obtaining improved image quality• Add exposures for handling  better low light environment by using near infrared spectral bandwidths• Use of standard, low cost and high resolution image sensors without color filter array and infrared filter

• Shorter frame acquisition time periods to produce sharper images by overcoming motion blur effects• Frame-to-frame registration to handle camera and scenes movements• Flexibility that allows image fusion of multiple, non-standard, colors, including NIR. This is extremely relevant for getting an SNR improvement• The composition of the image from few full frame images eliminates the demosaicing process. This improves the image quality and save some computation efforts.

Potential Application

The proposed solution can be applied to any low cost color and/or multispectral imaging system within the VIS-NIR.

Stage of Development

This project has been analyzed theoretically as well as via computer simulations. A detailed design is available as well. It should undergo proof-of-concept phase that includes prototype assembly and testing.

Patents

Patent pending

Supporting Publications

“Tunable Fabry-Perot Etalon as Sequential Filter for Color Image Acquisition and Reconstruction” (M.Sc. dissertation, in preparation)

Project manager

Oren Calfon
VP Business Development, ICT

Project researchers

David Mendlovic
T.A.U Tel Aviv University, Engineering
Physical Electronics

Ariel Raz
T.A.U Tel Aviv University, Engineering
School of Electrical Engineering