An Overview on Linear Unmixing of Hyperspectral Data

Abstract

Hyperspectral remote sensing technology has a strong capability for ground object detection due to the low spatial resolution of hyperspectral imaging spectrometers. A single pixel that leads to a hyperspectral remote sensing image usually contains more than one feature coverage type, resulting in a mixed pixel. The existence of a mixed pixel affects the accuracy of the ground object identification and classification and hinders the application and development of hyperspectral technology. For the problem of unmixing of mixed pixels in hyperspectral images (HSIs), the linear mixing model can model the mixed pixels well. Through the collation of nearly five years of the literature, this paper introduces the development status and problems of linear unmixing models from four aspects: geometric method, nonnegative matrix factorization (NMF), Bayesian method, and sparse unmixing.

Noninvasive hemoglobin sensing and imaging: optical tools for disease diagnosis

“…For spectral imaging methods, analysis can be time-intensive due to the large amount of data collected, which can be problematic for clinical translation in which the real-time display of biomarker data is often desired. Analysis methods vary depending on the biomarkers targeted and the type of tissue imaged from the simplest techniques such as linear spectral unmixing 198 , 199 to more complex methods such as multivariate analysis and machine learning. 26 , 72 , 164 , 165 Linear spectral unmixing determines the type and concentration of chromophores present based on input reference spectra for oxy and deoxy hemoglobin, from which oxygen saturation can be calculated.…”

Quantifying multiple stain distributions in bioimaging by hyperspectral X-ray tomography

“…4 ). A range of unmixing methods have previously been explored in other fields of spectral imaging 44 , in order to better classify material composition and improve spatial resolution. The methods typically focus around linear unmixing routines, applied to fields such as energy-dispersive X-ray diffraction 45 and fluorescence microscopy 46 , while more recently machine learning approaches have been studied for 'blind' spectral unmixing 47 .…”

Supplementary document for Unsupervised segmentation of biomedical hyperspectral image data: tackling high dimensionality with convolutional autoencoders - 6120812.pdf

“…Furthermore, in an experimental setting it is possible for a probed region to contain both fat and muscle (i.e. at a muscle/fat interface, or thicker samples), whereas each pixel in our simulated HSIs individually contain only fat or muscle [14,15]. Lastly, bone fragments may be present in some cuts of meat, and so there is no guarantee that substances other than fat and muscle will not be present in a given sample.…”