Introduction
Microscopy has long been a cornerstone of scientific discovery, offering researchers the ability to observe and analyse the microscopic world. Traditionally, microscopes have provided spatial information, allowing scientists to visualise the structure and morphology of samples. However, the advent of spectral imaging has elevated microscopy to a new level by introducing spectral information into the imaging process. Spectral imaging allows for the capture and analysis of spectral signatures from samples, providing additional insights into their chemical composition and properties. This breakthrough has empowered researchers to explore a wide range of applications, from life sciences to materials characterisation.
Most microscopic spectral imaging is implemented using a multispectral approach capturing images at just a few specific, pre-defined wavelengths or spectral bands. However, spectral imaging can be extended to hyperspectral imaging, which involves acquiring spectral data at multiple contiguous wavelengths with narrow bands across the electromagnetic spectrum. Hyperspectral imaging provides even more detailed spectral information, allowing for the characterisation of additional sample properties, such as chemical composition or material identification.
In this article, we will delve into the world of spectral imaging for microscopy applications and highlight some of the benefits offered by hyperspectral imaging.
Multispectral Imaging
Multispectral imaging involves capturing images at multiple wavelengths or spectral bands to analyse various aspects of a sample’s composition, structure, and behaviour. The implementation of this technique in microscopy typically includes the following steps:
A suitable light source, such as LEDs, lasers, or filtered white light sources, is used to emit light at different wavelengths. A filter wheel or tuneable filters can also be employed to select specific wavelengths. The filter wheel contains multiple filters, each transmitting light within a specific wavelength range, while tuneable filters offer continuous wavelength selection.
The imaging system is integrated with a microscope, such as a brightfield, fluorescence, or confocal microscope, positioning the light source and filter wheel/tuneable filters in the optical path. The specimen is illuminated with light of a specific wavelength, determined by the selected filter. The camera or detector mounted on the microscope captures the resulting image. This process is repeated for each desired wavelength by rotating the filter wheel or adjusting the tuneable filters.
The acquired multispectral images undergo image processing techniques, such as spectral unmixing or feature extraction algorithms, to analyse and extract meaningful information. Processed data is further analysed and visualised using software tools for tasks like image segmentation, classification, and quantitative measurements of spectral components.
Advancements in technology have led to sophisticated multispectral imaging systems with improved spatial and spectral resolution, sensitivity, and faster acquisition times. These advancements empower researchers to explore complex samples with greater accuracy and detail, facilitating scientific investigations and discoveries.
Figure 1: Multispectral vs Hyperspectral Imaging.
Multispectral imaging provides researchers with a versatile tool for investigating biological samples, analysing tissues, and assessing material properties. The technique offers valuable insights into complex systems and enables a wide range of applications including:
Fluorescence Microscopy
Multispectral imaging is widely used in fluorescence microscopy to study biological samples labelled with fluorescent dyes or markers. By capturing images at different wavelengths, researchers can separate and analyse the emissions from different fluorophores, enabling the identification and localisation of specific biomolecules within the sample. For instance, in immunofluorescence studies, multispectral imaging helps distinguish overlapping fluorescence signals from different antibodies targeting specific proteins, allowing for precise colocalisation analysis and protein interaction studies.
Tissue Imaging and Pathology
In medical research and pathology, multispectral imaging is employed to analyse tissue samples and identify disease markers or abnormalities. By acquiring multispectral images, different tissue components can be differentiated based on their spectral signatures. This enables the identification and characterisation of specific cell types, the detection of diseased regions, and the evaluation of tissue morphology. In cancer research, multispectral imaging aids in the analysis of tumour heterogeneity, helping to identify various cell populations within the tumour, assess their distribution, and study their interactions with the surrounding microenvironment.
Material Analysis and Quality Control
Multispectral imaging is valuable in materials science for the analysis and quality control of various materials and surfaces. It enables the characterisation of material properties, identification of defects, and assessment of surface coatings. For instance, in the semiconductor industry, multispectral imaging is used to inspect integrated circuits, identify manufacturing defects, and evaluate the uniformity of thin film coatings. By acquiring images at specific wavelengths, multispectral imaging can reveal hidden information about material composition, thickness, and structural variations that may not be apparent in standard brightfield or colour imaging.
Hyperspectral Imaging
Hyperspectral imaging involves the acquisition and analysis of images at numerous narrow and contiguous spectral bands, resulting in a highly resolved spectral profile for each pixel of the image. This can be integrated into microscopy applications using several different technological approaches.
Traditionally the line-scan or push-broom approach has been used but newer approaches including tuneable filters, snapshot cameras and Fourier transform based technologies are also being used.
With the line-scan approach, a light source capable of emitting a broad spectrum is combined with an imaging spectrograph or spectrometer device consisting of a slit or aperture, a dispersive element, and a detector. The imaging spectrograph is integrated into the microscope’s optical path, aligning the light source and the sample appropriately. The sample is illuminated by a line of the broadband light and the imaging spectrograph captures the dispersed light, recording the spectral information for each pixel of the image. The line is scanned across the sample, or the sample moved on a stage to create the full hyperspectral image.
Figure 2: Nireos HERA VIS-NIR coupled to a Nikon LV100 upright microscope.
Before image acquisition, calibration procedures are performed to account for system-related variations or distortions. This includes acquiring reference spectra from calibration standards to correct for instrumental effects.
The acquired hyperspectral image stack undergoes image processing and analysis. Techniques such as spectral unmixing algorithms are used to separate and identify the contributions of different components in the sample. Spectral classification methods can be applied to classify regions or objects based on their spectral characteristics. Visualisation tools are employed to display and explore the hyperspectral data, allowing researchers to study fine spectral details and extract valuable insights.
Similar spectral hypercubes can be collected using the tuneable filter, snapshot and Fourier Transform approaches.
Hinalea Imaging Inc has developed an alternative imaging technology using a Fabry-Perot interferometer tuneable filter approach to create multi- or hyperspectral images. A range of filter settings are selected sequentially, and the cameras capture up to 600 different spectral band images within a few seconds. An attractive feature of the tuneable filter is that it can be operated in hyperspectral or multispectral mode. This allows the researcher to explore which wavelengths are important for a specific application and then restrict data acquisition to just those few bands therefore reducing the size of the associated hypercube files.
Figure 3: Spectral image of stained soft tissue cancer with spectral classification using the Hinalea 4250 spectral camera
NIREOS SRL has developed another approach to acquire hyperspectral data. They have developed a time-domain Fourier Transform hyperspectral camera technology based on a common-path interferometer. The HERA VIS-NIR camera covers 400 to 1000nm and provides good resolution both spectrally and spatially. The camera is designed with a large entrance aperture, hence high sensitivity, which enables it to be used in low-light illumination conditions. The camera can be simply connected via the microscope C-mount camera port.
Figure 4: Spectral image of photoluminescence of a spin-coated perovskite semiconductor using the Hera VIS-NIR spectral camera. Nikon LV100 upright microscope. Excitation source: mercury arc lamp with 350nm bandpass filter. Magnification 1000x.
The choice of which technology to use will depend on the specific application and other considerations such as cost and performance. Hyperspectral imaging and multispectral imaging are both valuable techniques, but they offer different advantages. Some of the benefits that hyperspectral imaging provides compared to multispectral imaging include:
Figure 5: Images of a slice of corn (Zea mays) seed acquired with HERA hyperspectral camera from Nireos.
Enhanced Spectral Resolution
Hyperspectral imaging offers a higher spectral resolution compared to multispectral imaging. It can capture a larger number of narrow spectral bands, allowing for more precise characterisation and discrimination of different spectral signatures within a sample. This increased spectral resolution can provide finer details and improved spectral specificity, enhancing the ability to differentiate between closely related fluorophores or spectral features.
Continuous Spectral Coverage
Hyperspectral imaging captures the full spectral range of interest, providing continuous coverage across the entire spectrum. In contrast, multispectral imaging typically captures only a few discrete spectral bands. This continuous spectral coverage allows for a more comprehensive analysis of samples, particularly when studying samples with complex or overlapping spectral features.
Spectral Unmixing and Analysis
Hyperspectral imaging enables spectral unmixing, which is the process of separating and quantifying the contributions of different spectral components within a sample. This unmixing capability allows for accurate quantification and analysis of specific fluorophores or molecular species present in the sample. Multispectral imaging, on the other hand, provides limited information for unmixing and is less suitable for detailed spectral analysis.
Improved Multiplexing
Hyperspectral imaging allows for more effective multiplexing compared to multispectral imaging. Multiplexing refers to the ability to simultaneously detect and distinguish multiple fluorescent labels or probes within a sample. With hyperspectral imaging, it becomes possible to differentiate and visualise a larger number of fluorophores with overlapping emission or absorption spectra, enabling the study of multiple components or biomarkers in a single experiment.
Flexibility and Adaptability
Hyperspectral imaging provides flexibility in terms of selecting spectral bands of interest during data analysis. Researchers can extract specific spectral information from the acquired hyperspectral data, focusing on wavelengths or regions relevant to their research. This adaptability is particularly advantageous when investigating unique spectral features or analysing specific molecular properties in the sample.
Potential for Spectral Mapping
Hyperspectral imaging enables the creation of spectral maps, where spectral information is recorded for each pixel of the image. This capability allows for the visualisation and interpretation of spectral variations within the sample, providing insights into spatial and spectral heterogeneity. Multispectral imaging, with its limited spectral bands, may not capture the same level of detailed spectral information across the sample.
Summary
In summary, hyperspectral imaging offers higher spectral resolution, continuous spectral coverage, spectral unmixing capabilities, improved multiplexing, flexibility in spectral analysis, and potential for spectral mapping. These benefits make hyperspectral imaging particularly valuable for detailed spectral analysis, differentiation of closely related components, and comprehensive characterisation of samples in microscopy applications.
Further Information
An in-depth review of the different technologies used in multi- and hyperspectral imaging systems is provided in this Pro-Lite Technical Note.
The Pro-Lite range of multispectral and hyperspectral imagers is presented here.
