Conventional image sensors in commercial cameras have a limited dynamic range of brightness values they are able to receive and compute. Thus, images of real-world scenes may contain areas that are under- or over-exposed, which can obscure important details and hinder other processing techniques. A commonly used solution to this problem is the acquisition of multiple images of the same scene using different exposure levels, which are then combined to form a single image with high dynamic range. However, the acquisition of multiple images limits the usability of this method in scenes that change in time and precludes real-time implementation. This technology is a system to produce high dynamic range images from low dynamic range sensors using a single acquisition. An optical mask placed in series with the image sensor passes normalized light intensities to the image sensor, preserving image features that would ordinarily be obscured by the limited dynamic range of the sensor.
Typically, light intensities received by an image sensor outside its dynamic range are assigned values corresponding to the highest or lowest intensity limit of the sensor. Using this technology, light intensities taken from a single image acquisition are normalized by an optical mask before being received by the image sensor. The normalized intensity values can then be mapped to a low dynamic range display while preserving all pertinent information. This technology describes techniques for processing optical intensities in addition to methods of mask construction and implementation. Processing can be performed using several different methods and may be used with both CCD- and CMOS-based image sensors. Furthermore, masking may be performed using a translucent optical mask or with liquid crystal technology, providing many opportunities for flexible hardware implementation.
Patent Issued (US 6,864,916)
Patent Issued (US 7,924,321)
Tech Ventures Reference: IR MS98/12/08