Spectral Imaging: The Future of Spectroscopy?

Optical spectroscopy provides valuable insights into the interaction between light and matter, and has been applied in many fields. Traditionally, spectrometers have been used to monitor the progress of chemical processes through discrete measurement of cuvettes or small samples at fixed locations. However, with the emergence of new camera technologies and advanced image processing algorithms, new spectral imaging technologies have emerged, thus opening up new applications and replacing the traditional single point monitoring spectrometer in some cases.

The measurement results of microvascular blood flow and oxygenation in skin, muscle and other tissues were studied. These measurements are made from a single discrete location on the surface of the skin or other tissues, and the monitoring changes at these local points provide useful information about hemodynamics during medical or drug intervention. However, tissue perfusion has great heterogeneity, so the development of perfusion imaging technology enables people to better understand the distribution of microvascular flow and oxygenation, as well as spectroscopy. In many cases, single point spectrometer measurement can provide valuable data. However, new spectral imaging technologies are improving our understanding and bringing new insights. Hyperspectral imagers, originally developed to provide satellite and aircraft spectral images, are now increasingly used in medical research, machine vision, food science, material analysis, agriculture and mineralogy.

Remote sensing spectral imaging

Spectroscopy is widely used in plant science and agriculture for overall assessment of plant health or detection and identification of plant diseases. Many researchers use portable spectrophotometer to sample local plant leaves in the field. Generally, plants with photosynthetic activity will absorb red light and reflect green light and near-infrared light. The portable spectrometer can measure a wide spectral range of 350~2500 nm, and can calculate the nutrition index through various spectra. These indexes can reflect the health status of plants or the chlorophyll and nitrogen content and other indicators. However, the detection of small samples of leaves does not represent the growth of the entire region.

Standard color (RGB) aerial photos can provide images of the whole field, but there are discolored areas of leaves in healthy plant fields. At this time, interpretation based on simple RGB color photos is very limited. Hyperspectral imaging using cameras mounted on UAVs (unmanned aerial vehicles or unmanned aerial vehicles) is a more powerful technology, because it can obtain the entire spectrum at each point (pixel) in the image, so it combines the benefits of viewing images with the rich data information provided by spectroscopy. Although multispectral imaging can also provide images, it only contains data in a few spectral bands at each pixel. Whether to use hyperspectral imaging or multispectral imaging depends on what your task is. Images related to general plant health based on nutritional indicators can usually be generated based on several spectral bands, so they can be solved through multispectral imaging. However, applications that rely on subtle differences in spectral characteristics require more finely interpreted hyperspectral data. Therefore, hyperspectral imaging is a more appropriate technology to identify tree species or distinguish more subtle differences between diseased plants and healthy crops.

In remote sensing applications, spectral imaging has obvious advantages over single point spectral technology, but these technologies are highly complementary. The information provided by the airborne camera can be effectively verified through a single point measurement on the ground (the so-called "ground live"). In addition, these two sets of data will increasingly be used in conjunction with hyperspectral images provided by Earth observation satellites. This will provide more detailed information on crop health and environmental issues at multiple levels of spatial and spectral resolution.

Spectral imaging in machine vision

Machine vision is another area that benefits from the increased application of spectral imaging. Most materials can be identified by their interaction with light (reflectivity, transmissivity or absorptivity). Different spectral characteristics can be used to identify materials or separate specific materials from each other. Single point spectrometer measurements may be useful, but spectral imaging can check the spatial distribution of different materials. By identifying spectral features beyond the visible wavelength range, spectral imaging can also go beyond the functions of traditional RGB cameras. Broadband spectral imaging covering the visible and near-infrared wavelength ranges can be used to distinguish different plastics, identify drugs or grade food quality. The progress of real-time classification technology based on machine learning has also begun to be used to support and even automate decision-making, accelerate processes and improve quality control.

The use of spectral imaging in machine vision will depend on the balance between its benefits, equipment costs and ease of implementation. Advances in spectral imaging hardware and software will help simplify ease of use, and the cost depends to some extent on whether multispectral or hyperspectral data is required. In a controlled environment with a limited number of variables, the application of multispectral imaging can reduce the cost. However, hyperspectral imaging may be required in case of the need to detect subtle differences in material or product quality.

Spectral imaging in medical imaging

Returning to medical and surgical applications, spectral imaging can provide significant benefits in several areas. I have mentioned the imaging of microvascular tissue oxygenation, but it also has applications in endoscopy, which can be used for non-invasive disease diagnosis and image guided minimally invasive surgery.

Real-time spectral imaging has great potential in image-guided surgery and cancer diagnosis. Cancer tissue is usually indistinguishable from healthy tissue in the operating room. However, spectral imaging is possible to distinguish the spectral differences between healthy tissues and cancer tissues, and classify different tissues in real time during surgery. By providing color coded images at the molecular and tissue levels, this has the potential to enhance the surgeon's field of vision, which in turn can maximize the removal of tumors without damaging adjacent normal tissues.

So is spectral imaging the future?

The benefits of spectral imaging are significant and extensive. With the development of hardware technology, image analysis methods and computing capabilities, I hope spectral imaging will play an important role in many applications beyond the examples mentioned here. The ability of spectral imaging to detect objects that may be missed by human eyes or ordinary RGB machine vision cameras cannot be underestimated. Single point spectral analysis still plays a vital role, but multispectral and hyperspectral imaging provides an exciting glimpse of the future and makes people confident about health and happiness.