Utilizing near-infrared and colour information in an automotive wide-dynamic-range night vision system for driver assistance

Abstract

Vision-based driver assistance systems utilizing both colour and nearinfrared (NIR) can significantly reduce the risk of traffic accidents by improving night visibility of the roadway and obstacles. The night vision system described here combines within a single imager wide dynamic range (WDR) technology with sensitivity extended into the NIR, and colour capability. The multiple-slope CMOS pixel covers a dynamic range of more than 120 dB. Adaptive control algorithms ensure detection reliability by maintaining the incremental signal-tonoise ratio (iSNR). Colour sensing is based on a modified colour filter array (CFA) with a 4th channel to capture monochrome information from the imager’s entire sensitivity range including the NIR component of the headlight spectrum that is invisible to the driver. Virtual NIR filter algorithms extract relevant colour information that is then fused to the luminance information, resulting in up to 70% higher sensitivity than conventional colour cameras.

Original document

The different versions of the original document can be found in:

• http://dx.doi.org/10.1007/978-3-642-16362-3_15

• http://link.springer.com/content/pdf/10.1007/978-3-642-16362-3_15.pdf,

http://dx.doi.org/10.1007/978-3-642-16362-3_15

• https://link.springer.com/chapter/10.1007/978-3-642-16362-3_15,

https://academic.microsoft.com/#/detail/2097235200