How does it work – some physics, chemistry and optics
Spectroscopy, in short, is the study of how electromagnetic radiation (light) is emitted and/or reflected by various objects/materials. Differences in energy potential and length of emitted/reflected waves are measured. In order to be able to study that it is necessary to separate the light into different ranges of electromagnetic waves and then to measure the intensity of each of them.
To make this possible beam dispersion must be done – for example with a prism. We did that in physics lesson, however now we are dealing with light that goes beyond the visible spectrum. These can be ranges in tens and hundreds of nanometres (ultraviolet) and on the other hand – expressed even in microns (infrared).
Different molecules and atoms react in a very different way to light. How they absorb, emit or reflect light depends on their structure. Each molecule/atom has a unique chemical structure and thus has a unique reaction to light. They leave their individual and unique “fingerprint”, also called the spectral signature.
Hyperspectral sensors capture these signatures in the form of a set of images. Each of these images is a part of the electromagnetic spectrum (spectral band).
Then these images are superimposed on each other creating the so-called data cube containing spatial dimensions x and y as well as dimension λ – wavelength (also known as spectral dimension).
These data are analysed by specialists supported by advanced software, therefore – without further going into details – it is worth noting that only recently we have been able to fully use this data in vision systems.
From space down to Earth
Hyperspectral imaging has begun its “career” in applications in large systems installed in planes and satellites. In the early 1970s the LandSat I satellite was the first ever to provide multispectral images of the Earth’s surface.
In the following years, the conquest of space and the arms race between world power fuelled the development of vision technologies. It resulted in sending into space a numerous of satellites with increasingly sophisticated observations systems.
Recent decades mean a rapid development of civil projects. Satellites, planes and UAVs transmit gigantic amounts of data from multispectral sensors every second.
We are proud to say that Scanway team is also constantly working on the development of this technology by creating the ScanSAT satellite (click).
Available today compact and cheap spectrographic systems have a huge number of applications – and they provide significantly more information that LandSat did. All this with the size of your hand.
These devices find practical applications in the control of industrial processes and quality control but also in many other areas, such as:
- Geography, geology, cartography
Multispectral data allows to obtain much more complete information about the area than traditional satellite photography (RGB or monochromatic). Based on the analysis of light reflected by various parts of the terrain, conclusions can be drawn regarding the type of rocks, soil composition and humidity, melting of ice layers or the type of vegetation in a given area. It also allows the search of new sources of mineral resources.
Multispectral satellite images allow, among others, to study the distribution of water vapor concentration as well as the temperature distribution of not only gases, but also soil and water.
- Precision agriculture and forestry
In agriculture of appropriately high resolution, hyperspectral images allow identification of specific plant species (including weeds), their condition, flowering phase, pollen, parasitic infections or the degree of soil infestation. Multispectral and hyperspectral images are used in the study of the distribution of plant populations (including oceanic flora) or forest stands in forest areas.
Hyperspectral imaging is also a great tool for remote (satellite) detection of environmental pollution. Chemical substances, especially organic ones, have characteristic emission and absorption spectra of electromagnetic waves. Hyperspectral imaging allows the identification of traces of pollutant emissions, on land, in water and air (including smog analysis), as well as monitoring of greenhouse gas emissions and gases adversely affecting the ozone layer. It also supports sorting of waste – e.g. identification of recyclable plastic.
- Rescue and searching for objects
Multispectral imaging enables identification of objects invisible in the image from traditional colour cameras or during direct observation.
- Medicine and biotechnology
Multispectral images are also applicable in laboratory conditions in biology and medicine, such as for example analysis of wounds, neoplastic changes or microscopic analysis of tissues.
- Food production
Multispectral imaging is increasingly used in the study of raw materials and ready-made food products – to detect defects, damage, contamination, parasitic or fungal infections and residues of plant protection products.
- Pharmaceutical production
Vision systems improve quality of both raw materials (including active substances) as well as finished products and their packaging. They allow to ensure a homogeneous distribution of the active substance in the tablet mass or measurement of the humidity level.
Multispectral imaging is also a tool which improves the investigation work and collecting the evidence. It allows quick detection of micro traces of specific organic or chemical substances, detection of forgery (e.g. banknotes) and supports the work of graphologists.
- History of art and archaeology
In studies of antique works of art (e.g. paintings and books) multispectral imaging is a perfect tool for non-invasive testing of authenticity or the invisible content of a work.
- Defence and security
Hyperspectral systems are ideally suited for identifying targets and revealing camouflaged objects.
What is waiting for use in the future
We are ahead of times in which multispectral observations were reserved for academia, military or space scientific research centres and laboratories.
Until recently the limitations were still the experimental nature of the solutions being developed.
But now the technologies that have allow human eyes to “reach where eyes cannot reach” are now available almost at your fingertips. Human eye never had so much power.
Of course, the devices itself are easily accessible for everybody. The snag lies in their respective combination and adaptation to the requirements of a particular situation. Therefore, many practical challenges (especially in industry) in which modern vision systems can help, require dedicated solutions, tailored exactly to a given situation and characteristics of the tested objects. We deal with such solutions at Scanway. We develop systems prepared for specific types of observations, industries or types of production lines.
However, modern tools mainly provide raw data.
Therefore, the next challenge is to process and analyse the information obtained.
And in this place, we make a full circle – we return to the man and his abilities. Not necessarily the sight itself, but the cognitive abilities and knowledge and experience based on them.
Although we are increasingly using big data solutions, artificial intelligence, machine learning and neural networks, we still ned people – teams of specialists with broad, multidisciplinary knowledge to analyse and, above all, to draw final, important conclusions from the processed data. They are the last link in the chain, especially where ready-made solutions are not enough. Exactly this team works at Scanway.
The future looks optimistic. Practical applications for vision systems are basically everywhere. The most important thing is that they can be used not only for commercial applications, but also can contribute to the broadly understood good of humanity: improve environmental protection, anticipate weather more successfully, investigate geological phenomena, improve people’s safety or support medicine progress. In everyday work, the implementation of this mission ensures the greatest satisfaction.