Eliminating Crosstalk by Rearranging Patterned Thin-Film Filters on Image Sensor

A Stanford researcher has developed a simple hardware solution to eliminate a newly identified form of optical 'crosstalk' on image sensors. Three types of crosstalk are commonly identified in image sensors: electrical (diffusing electrons), optical (light arriving at the wrong pixel) and spectral (poor filtering). All three forms of crosstalk reduce spatial and/or spectral resolution, which lowers the commercial value of the camera.

This Stanford innovation addresses a newly identified form of optical crosstalk, unique to image sensors with patterned thin-film filters. The researcher discovered that cavity crosstalk occurs because the filters are made with highly reflective mirrors, which can trap and transport light to neighbouring pixels. The novel, low-cost hardware solution he developed eliminates this crosstalk by rearranging the patterned thin-film filters. The two main applications for this work are spectral imaging and miniature spectrometers.

Figure:

Image from Inventor

Stage of Development – Proof of Concept