Multispectral Imaging: An Overview
What is Multispectral Imaging?
Multispectral imaging (MSI) is a sub-technique of spectral imaging. Spectral imaging uses multiple bands across the electromagnetic spectrum, capturing information from visual rays like red, blue, and green light and those non-visible to the human eye or standard camera lenses, such as ultraviolet, infrared, and even X-rays.
Initially developed for military surveillance, MSI isolates and captures information from specific wavelengths along the electromagnetic spectrum, revealing additional information not visible to the human eye. MSI differs from traditional color photography in two ways. First, traditional color imaging, or RGB imaging, relies on overlapping three visible light waves (red, green, and blue) to produce an image. Multispectral imaging isolates and does not overlap wavelengths, allowing the user to look at the effects of individual wavelength subsets. Second, as mentioned before, traditional imaging relies solely on red, green, and blue light to produce images, while multispectral captures information across several waves, such as ultraviolet and infrared.
There are numerous advantages to using multispectral imaging that distinguish it from RGB imaging, such as change monitoring, hidden pattern detection, ultra-high resolution, and composition detail. The ability to monitor subtle, imperceptible changes over time has significant implications in agriculture and environmental monitoring. Moreover, the ultra-high resolution of multispectral images provides users with detailed and accurate data, which is particularly beneficial in fields requiring ultra-precise action, such as vegetation studies. The capacity of MSI to observe light outside the range of human vision and provide insights is widely utilized in weather forecasting, and environmental monitoring, underscoring its practical applications.
Multispectral vs Hyperspectral?
Multispectral imaging (MSI) is often confused with its sister technique, hyperspectral imaging (HSI), so let's clear up any confusion. While MSI isolates and captures light from specific wavelengths on a smaller scale, HSI takes it a step further, capturing light on a larger scale in hundreds of spectral bands. In other words, MSI is typically used to demonstrate land surface features and landscape patterns, while HSI's unique ability to distinguish and characterize materials opens up a world of possibilities in various fields.
Due to its limited array of spectral bands captured, multispectral imaging does not provide the same level of spectral detail as hyperspectral imaging but offers a balanced spatial resolution and spectral information. HSI is much slower than traditional and multispectral imaging and requires significant processing power and storage but provides the highest spectral resolution. However, while it captures a lot of information for users, it has a lower spatial resolution, meaning the scale of the smallest unit in an image is larger. Hence, objects are more challenging to visually differentiate, meaning less detail. Users who are familiar with which specific spectral bands they want to observe and do not require extensive material differentiation would benefit from using MSI, while those who require a comprehensive spectral analysis would benefit more from HSI. For a great overview of MSI vs HSI, check out this overview from GIS Geography.
Credit: ResearchGate
How does it work?
Multispectral imaging relies on the scientific premise that all objects possess unique spectral signatures, meaning they interact with specific wavelengths in their own way. Objects such as vegetation, land structures, and animals reflect, emit, and absorb electromagnetic waves differently, meaning MSI allows users to determine their object of interest's physiological and chemical properties.
Multispectral imaging involves several steps to collect, synthesize, present, and interpret data. First, using sensors sensitive to specific segments of the electromagnetic spectrum, observers can detect the energy emitted or absorbed by objects within a specific spectral band. Analysts can employ filter wheels or a dual-band mirror to separate light before it reaches the sensors, which reflect or transmit light differently based on the wavelength.
In order to provide a comprehensive view across multiple spectral bands, the MSI system captures multiple images of a given setting, each in a spectral band of interest. This can be done in two ways. The first method involves subsequent photo taking, using a single sensor camera and a filter wheel. The second method, which is more complex and requires precise setup, involves simultaneous photo-taking utilizing an array of cameras, each outfitted with a specific spectral filter. While the single-sensor camera method is slower, it is a simpler and more cost-effective approach.
Once the images have been collected, they are corrected for atmospheric interference (rain, fog, or clouds) and sensor noise (environmental factors, sensitivity metrics). After these corrections or limitations have been noted, analysts can examine the spectral data and draw conclusions.
What is the electromagnetic spectrum?
The electromagnetic spectrum represents the various radiation segments, delineated by wavelength, from long and slow radio waves to short and high-frequency gamma rays. Each segment is characterized by the different wavelengths and frequencies influencing how objects receive or repel radiation. Here is a rundown of the seven segments:
Radio waves: Identified in the late 1880s by Heinrich Hertz, radio waves are the longest rays on the spectrum with wavelengths greater than 1 millimeter. Some of these waves are as long as football fields! As given in the name, radio waves are usually used for telecommunications because they can freely traverse our atmosphere.
Microwaves: A versatile ray on the spectrum used in microwaves to cook food, display weather projections on TV, and survey the earth's surface from afar. These waves have wavelengths ranging from 1 millimeter to 1 meter in length. Divisible into several subcategories, such as medium-length waves known as C bands, which can penetrate through clouds and smoke to observe the earth's surface; L bands–which are used in car and mobile GPS systems; and other subtypes, X, Ku, and more, are used for communications with satellites.
Infrared: Discovered in 1800 by William Hershel, infrared rays are used in thermal imaging, monitoring the earth's surface, and observing cooler space phenomena such as cool stars and nebulae. Infrared rays are longer than visible light, with wavelengths of 15 micrometers to 1 millimeter, so they can pass through dust and gas. A subset of infrared rays, near-infrared waves, are instrumental in observing and analyzing vegetation health and soil composition. They measure an object's behavior in response to infrared waves (how it absorbs or reflects) rather than an object's production of infrared. Another subset, short-wave infrared, is leveraged for its ability to bypass atmospheric haze, which is helpful in geological mapping and mineral identification. Mid-infrared waves have slightly longer wavelengths and are commonly used in thermal imaging. It helps measure temperature differences and thermal anomalies, applicable in activities that need thermal data, such as military surveillance, volcanic activity, and environmental monitoring.
Visible light: The only segment of the electromagnetic spectrum the human eye can see. If sunlight is put through a prism, one can observe the entire range of visible light, starting with red, with the longest wavelength (700 nanometers), through all the colors of the rainbow to violet, which has the smallest wavelength (400 nanometers). Visible light is instrumental in environmental studies, which use blue light to study the atmosphere and bodies of water; green light, which is used to analyze plant health and deep water bodies; and red light, which is used to monitor and assess chlorophyll absorption, a marker of plant health.
Ultraviolet: Commonly used in forensic analysis and art restoration, ultraviolet light comes from the sun, has a wavelength ranging from 100 nanometers to 400 nanometers, and can be divided into three subcategories: UV-A, UV-B, and UV-C. UV-C rays are the most harmful of the three but are fortunately all absorbed by our atmosphere. UV-B is also detrimental and is known to cause DNA and cellular level damage in organisms, but it is also absorbed mainly by our atmosphere. Ultraviolet rays were discovered in 1801 by Johann Ritter, who knew photographic paper developed better in blue light and decided to expose the paper to light beyond violet, thus proving the existence of UV rays.
X-rays: Discovered and implemented in medicine in 1895 by a German scientist named Wilhelm Conrad Röntgen, who realized that by firing X-rays through a body, due to the density of bone, an image would appear showing only the bones and any foreign object. They have wavelengths of 0.03 to 3 nanometers and are commonly used in medicine, construction, and material science. These rays, while helpful, can be hazardous to human health if exposed in large quantities.
Gamma rays: The smallest but most potent, with waves shorter than 10^-11 meters. They occur during nuclear explosions, lightning strikes, and radioactive decay. Because of their size, these rays cannot be captured nor reflected by mirrors, so observation is much more difficult. Gamma rays are instrumental in observing other planets and their surfaces, allowing scientists to identify which elements are present.
Credit: iStock by Getty Images
Now that we've established the different types of electromagnetic waves let's look deeper into what helps analysts determine what they are looking at:
Refraction occurs when light enters a new medium and changes its trajectory as it leaves. The most common example is a prism, where when a singular beam of light enters it, the waves are separated, and the colors of the rainbow appear. The separation is caused by the variance in wavelengths, causing them to be refracted at different angles.
Reflection occurs when light waves hit the surface of an object and bounce off. Observers note how much light is reflected and which colors are absorbed and use their findings to classify objects. For example, flat objects like mirrors tend to reflect almost all waves, and the waves reflected will tell observers what color the object will appear visually. The object's or organism's properties determine which waves are absorbed and reflected.
Absorption occurs when the photons (tiny particles that contain waves of electromagnetic radiation) from incoming light interact with the molecules within the object, causing them to vibrate and generate heat. Darker objects, such as tarred roads, will absorb more sunlight and thus create more heat.
Scatter occurs when light bounces off an object in a non-uniform way and is dependent on the length of the wave and the composition of the object it interacts with. For example, light scattering is why the sky appears blue. As sunlight traverses our atmosphere, it interacts with the molecules. Because the wavelengths for blue and violet are the smallest, they scatter with much more intensity than red or orange, giving our sky a blue hue. The sky is an example of elastic scattering, which occurs when the particles are smaller than the wavelengths they interact with. Inelastic scattering is the inverse, where the particles are more significant than the wavelengths. Inelastic scattering is most commonly seen in clouds. As sunlight passes through clouds, the dust, water, and ice particles scatter the light equally, giving clouds their white color.
Diffraction occurs when light bends around an object or passes through an opening. Not to be confused with refraction, diffraction is used to observe the molecular structure of objects by analyzing the way light diffracts when it passes around. A common example of natural light diffraction is a rainbow, which occurs when light passes through raindrops, and due to the variation in position and size, light waves are separated into individual colors.
Credit: The Rockefeller Institute
Applications of Multispectral Imaging
Agriculture:
Multispectral imaging is instrumental in agriculture. With its capacity to see beyond the range of visible light, multispectral imaging can help farmers detect problems before they can see them visually. Regarding crop health, MSI can detect diseases early on so farmers can treat the infected area before it spreads to the rest of the crop. This detection is possible by monitoring plants and their reluctance in the near-infrared segment of the electromagnetic spectrum as stress begins to show there first. Regarding irrigation, MSI can also show which areas are getting overwatered or not watered enough. Farmers can also gain more accurate insights into plant count and any problem areas using advanced imaging instead of manually searching large fields.
Credit: ABJ Drone Academy
Medicine:
In the medical field, recent advancements have elevated MSI to a crucial diagnostic tool. Using near-infrared, medical professionals can peer into tissues and study their molecular composition without the need for invasive biopsies. This non-invasive approach allows doctors to identify malignant tissues based on the way light scatters when in contact with compromised tissue. The MSI process saves time and eliminates the need for tissue removal, reducing invasive operations and patient discomfort. This patient-friendly aspect of MSI underscores its importance in the medical field.
Art Conservation:
For researchers, MSI is not just a tool but a gateway to a world of hidden secrets. It provides a unique opportunity to enhance faded writing, uncover hidden text written in nonvisible inks, and even reveal the underlayers of artwork. This allows users to delve deeper into the changes in art strategy over history and how the decisions the artist(s) made evolved without compromising the original work.
Credit: Springer Open
Multispectral imaging (MSI) is a complex technology that bridges the gap between the visible and invisible, offering clarity and detail across various fields. By capturing data across spectral channels, MSI allows users to see the world in ways traditional imaging cannot. Compared to hyperspectral imaging, which provides highly detailed spectral information, MSI offers a balanced approach, delivering essential data with simplicity and efficiency. This straightforward and efficient approach has revolutionized agriculture, healthcare, and environmental monitoring. As the technology continues to evolve, the potential applications for MSI will undoubtedly expand, opening new frontiers.
It's worth reiterating that multispectral imaging is a highly complex and nuanced technology, and this article is just the beginning of your journey. Learning how objects interact with light intuitively takes years of study and experience, and it's perfectly normal to have questions after reading this article. We're here to support your curiosity if you're intrigued by MSI, HSI, and other types of imaging and data collection. Let us know, and we'll be happy to provide more in-depth articles and resources for your continued learning.
Best,
Cece and the Nova Team