Near Infrared

Abstract

Silicon-based camera sensors exhibit significant sensitivity beyond the visible spectrum (400-700nm). They are  able to capture wavelength up to 1100nm. Near-infrared is the portion of the radiation spectrum that ranges from 700 to 1100 nm. Even though this radiation can be captured by silicon, it is usually considered noise and therefore is removed by putting a filter (hot-mirror) in front of the sensor.

Retaining instead of  eliminating NIR information improves certain tasks in digital photography and computer vision. Intrinsic properties of the NIR part of the spectrum make the scene sharper and less affected by atmospheric haze. NIR images are also “transparent” to a number of colorants or paints and as such can see through the first layer to reveal the surface underneath. Exploiting the inherent correlation and de-correlation of the different waveband images can lead to improved color rendering and object recognition. 

Near-infrared imaging

Combining near-infrared and visible images

Automatic face smoothing – setup at the swiss camera museum, vevey, switzerland

Material-based object segmentation

Removing shadows from images

Shadow Detection

Semantic Image Segmentation

Color Study and Deblurring