Hyperspectral Imaging
(Prof. Jan G. Korvink, Dr. Vlad Badilita, )
Hyperspectral imaging (HSI) is a technology which combines 2D spatial imaging and spectroscopy to provide detailed information about the composition of a target. Hyperspectral imaging enables the acquisition of spatially-resolved spectral information by acquiring a spectrum for each pixel in an image using many contiguous narrow spectral bands. The acquired spatial and spectral information forms a 3D data set, which is often referred to as the “hyperspectral cube”.
Due to the large amount of information it provides, hyperspectral imaging has emerged in recent years as a non-invasive, non-destructive, and contact-free analysis tool. Applications for hyperspectral imaging continue to grow from remote sensing to agriculture, industrial sorting, medical diagnostics, food safety, and environmental monitoring.
Figure 1: Schematic representation of the pushbroom hyperspectral imaging technique.
One of the main methods for acquiring hyperspectral data is the “pushbroom” technique. Using a linear slit aperture and a diffractive optical element, the spectral information can be acquired in a line-by-line fashion by means of spatial scanning (Fig.1).
Current hyperspectral imaging systems typically come with a large size and high cost. In addition, the spatial scanning, required for pushbroom systems, is usually achieved through relative motion between the target and the system, for example using an external scanning stage, which further increases costs and narrows application scenarios. The large size, the high costs, and the need for relative motion has so far limited the use of hyperspectral imaging to applications in research and industry, and prevented its utilization on a wider scale in everyday life.
This project aims at developing a portable, compact, and low-cost HSI pushbroom system through miniaturization of key optical components using microfabrication techniques. One key component is a miniaturized scanning slit to enable a compact pushbroom HSI system, which does not require relative motion for data acquisition.