BGU researchers have developed an ultra-spectral miniature camera. Most of the existing hyperspectral imaging systems are bulky, complicated and expensive, while this camera has the advantages of being compact, simple to operate and cost effective. The paper describing the development was published in the prestigious journal Scientific Reports of the Nature publication group.
Spectral imaging enables the capture of images at many different light wavelengths over a wide spectral range: each pixel in the camera not only gives an image of a particular point on the object but it also gives the spectral content of the light originating from that point. As a result, the actual material of the object can be characterized in high spatial and spectral resolution because of the large amount of information the camera captures. Therefore, ultra-spectral imaging is of significant importance in many applications such as medical diagnosis, biological and environmental research, monitoring of industrial fabrication processes, and remote sensing such as imaging from airplanes and satellites.
The miniature ultra-spectral camera was developed in the labs of the Unit of Electro-Optical Engineering at BGU, a collaboration between Prof. Adrian Stern, Prof. Ibrahim Abdulhalim, Prof. Dan Blumberg, Prof Stanley Rotman and PhD students Itzik August, Yaniv Oiknine and Marwan Abuleil. The camera uses a single pixel liquid crystal device developed in the Liquid Crystals Devices and NanoPhotonics lab of Prof. Abdulhalim together with a special compressive sensing system designed by Prof. Stern's group. Prof. Abdulhalim is a member of BGU’s Ilse Katz Institute for Nanoscale Science and Technology.
The basic principle of the methodology relies on sequentially modulating the spectrum of the light passing through the liquid crystal device when it is placed between two light polarizers by applying a small variable voltage of up to 10 Volts. The liquid crystal device is designed in a way that complies with the compressive sensing theorem, which is a recently introduced revolutionary theorem in the field of data acquisition. Using proper algorithms, hyperspectral and ultraspectral (order of a thousand spectral bands) images were reconstructed with resolution better than the standard limits. The compressive sensing design enables high speed data acquisition and reduced storage memory because it minimizes the collection of redundant data during acquisition, unlike the traditional imaging approach in which as much data as possible is first collected and then the redundant data is discarded by digital compression techniques.
The research was funded by The Ministry of Science, Technology and Space of the State of Israel. Rolic-Switzerland provided the photoalignment material.
doi:10.1038/srep23524
Upper row: A color image generated from a set of 1,000 spectral images at different bands.
Lower row: Four images at different spectral bands (610, 570, 530, 490nm) from the reconstructed ultra spectral cube of 1000 spectral bands.
The acquisition effort for capturing this data is 10 times less than that required with conventional systems (the data captured is 10 times less than the size of the reconstructed spectral cube).
