Hyperspectral brain imaging

Push-broom hyperspectral imaging is a technique used to collect full spectral and spatial information of a target line by line. This imaging method that reconstructs images as a function of wavelength has proven to be useful in industrial, airborne, and satellite applications such as target recognition and classification. More recently, a foray into applying this technique to applications like diagnosis and health monitoring has displayed huge potential in revealing insights into medical conditions and treatments. However, the hyperspectral imagers used for these studies have been costly and inaccessible. Hence, this project aims to investigate the feasibility of applying a low-cost hyperspectral imager to microscopic medical applications.

This project developed materials and methods used to construct, calibrate, and test a low-cost hyperspectral camera. The push-broom hyperspectral imaging system was validated by comparing the data obtained from imaging ischaemic and non-ischaemic human fingers. Results show that images of the target objects could be reconstructed, albeit with insufficient resolution to accurately differentiate the fingers. Due to the limited useable finger images obtained using the camera, images from an open-source database were used to evaluate the efficacy of different methods used to perform preprocessing and end-member analysis. It was concluded that these methods were successful proofs of concepts that the device and aspects of processing strategies worked. This serves as a starting point to further investigate how the hyperspectral camera can be integrated into a common laboratory microscope for micro-scale medical imaging.