Cameras “See” More with Imec Hyperspectral Imaging
New CMOS-based imagers offer smaller size and cost, but with greater discriminating powers over today’s RGB-based cameras.
Imec, a major researcher in nanoelectronic technology, has announced an evaluation kit for its compact, low-cost, CMOS-based hyperspectral imaging filter system. Hyperspectral technology provides images with greater detail than traditional-visible-spectrum Red-Green-Blue (RGB) camera systems. This additional detail permits the human viewer or machine system to “see” more details about the image (i.e., the materials that make up the image). Application areas for hyperspectral imaging include the medical, optical sorting, remote sensing, and even agricultural markets.
“The added value of hyperspectral imaging is a significant increase in the capability to discriminate between images – to look at the whole spectrum beyond just the visible,” explained Imec’s Senior Researcher, Murali Jayapala. “Now, you have greater discrimination power to distinguish between different kinds of objects.”
How does spectral imaging differ from basic camera systems? The human eye sees visible light in three bands – red, green, and blue. Hyperspectral imaging extends the visible bands into a much wider spectrum, which results in a more complete identification of the materials that make up the scanned object than RBG-based camera systems.
Imec developed the CMOS-based sensor-imaging chip by tiling hyperspectral filter sets onto the sensor chip. In other words, it processes the spectral filters directly on top of the CMOS image sensors at the wafer level. Hyperspectral sensors can be tailored to specific customer requirements during the filter’s design phase.
This approach greatly reduces the size and cost of the overall hyperspectral imaging systems. The resulting camera is able to acquire real-time hyperspectral video.
How does this new sensor impact the world of semiconductor IP? Spectral imaging technology will compete with today’s existing visible RGB-based sensors. This means that new hyperspectral imaging software and algorithms will be needed (e.g., in analogy to current color-correction matrices as offered on Chipestimate.com). Color-correction software is needed to eliminate the color overlap created by most RGB-based channels.