SBIR Phase I: Spatial Color-Interpolation for New Image Sensor

This Small Business Innovation Research Phase I project is aimed at developing a new image sensor. Digital image sensor devices including cameras and camcorders play an ever increasing role in remote sensing applications. The present image sensors, both CCD and CMOS types, used in these devices suffer "aliasing" artifacts. A costly optical "antialiasing" filter can be used upon image sensor to reduce the artifacts but it will also blur the image at the same time. Pioneering works by Yellot and Mitchell indicated other possible solutions to this problem about 20 years ago. With the use of stochastic sampling of the image together with the "Blue Noise Criteria", a new photo-diode layout of image sensor is proposed in this SBIR research project. Due to the discrete nature of image sampling, color interpolation is required for all pixels on the chip when Bayer color filter array is employed. Two standard interpolation methods plus an innovative multi-dimension interpolation method (Dirac-Monte Carlo method) developed by the proposer will be used to perform 2-dimension scattered color data interpolation on the chip for the reconstruction of the image. Image sensors constitute a fundamental part in digital cameras and camcorders. It is expected that this work will lead to further scientific understanding of signal sampling and reconstruction. The new image sensor proposed can be employed in digital camera and camcorder with different level of sophistication and cost. In addition, different sizes of CMOS chips will cover products ranging from consumer electronics (web camera, cell-phone camera, digital camera and camcorder) to deep space detectors.