top of page
Writer's pictureHanna Vitória

Spectral Imaging: Concepts and Applications

Have you ever imagined how many mysteries surround us that our eyes cannot perceive?

 

When we take a photo with our smartphone, we are capturing only a small portion of the electromagnetic spectrum: visible light. But what lies beyond it? What secrets are hidden outside the reach of our vision? Could invisible information be used to detect diseases early or identify adulterated substances in food? These questions, which seem taken from a science fiction movie, are much closer to our reality than we imagine.


Spectral imaging emerges as an innovative technology capable of revealing the "invisible," translating the physical and chemical information of objects into the digital realm. Whether verifying the quality of the coffee we consume daily, monitoring a plantation to prevent harvest losses, or diagnosing health problems even before visible symptoms appear, spectral analyses transform reflected light into valuable data.


Let’s understand how all this works:


Differences Between Multispectral and Hyperspectral Imaging

Hyperspectral and multispectral imaging capture a much broader range of the electromagnetic spectrum, going beyond the visible spectrum. These technologies encompass ultraviolet light (UV, 200–400 nm), visible light (VIS, 400–780 nm), near-infrared (NIR, 780–2500 nm), and mid-infrared (MIR, 2500–25,000 nm), far surpassing the limitations of conventional cameras.


These images use reflected light interacting with materials, capturing structural and spectral details resulting from processes like absorption, scattering, and transmission. Thus, they make visible the information that our eyes can never see, opening doors to applications that can transform various fields of knowledge.


Spectral images can be classified as multispectral and hyperspectral, each with distinct characteristics:


Differences between Multispectral and Hyperspectral images. Data Cube, Spectrum
(Sabin et al., 2024)

Multispectral Images

  • Number of Bands: Typically capture 3 to 10 spectral bands, which are broader. Common bands include red, green, and blue (RGB) colors, along with some near-infrared bands. 💡

  • Spectral Resolution: Bands are broader (generally between 20 to 100 nm), resulting in less detailed spectral discrimination.

  • Advantages:

  • Easier data acquisition as it doesn’t require high processing power.

  • Generally lower equipment costs.

The multispectral technologies developed by OpenScience use 20 to 40 wavelengths.

Hyperspectral Images

  • Number of Bands: Capture dozens or even hundreds of narrow spectral bands (typically between 10 to 20 nm), forming a data cube and enabling highly detailed analyses.

  • Spectral Resolution: The high spectral resolution allows for deep, detailed analysis and the precise identification of different materials based on their unique signatures.

Characteristic

RGB Images (from smartphones)

Multispectral Images

Hyperspectral Images

Number of Bands

3 (RGB)

3 to 10

Hundreds

Spectral Resolution

Low

Medium

High

Cost

Low

Medium

High

How Spectral Imaging Works


How does Multispectral and Hyperspectral image capture work? Point Scan, Push Broom, Spectral Scan, Snapshot
(Sabin et al., 2024)

The techniques for capturing spectral images vary according to the type of scanning and sensor arrangement. The main categories are:

  • Point Scanning (Whiskbroom): The sensor scans the sample point by point, collecting the full spectrum for each point. This approach is common in laboratory bench radiometers and offers high spectral resolution.

  • Line Scanning (Pushbroom): A line of sensors collects data from a single line while the system or object moves. This technique is widely used in remote sensing (by satellites or drones) and industrial applications on production lines.

  • Spectral Scan: Involves spectral scanning of an object or area, capturing different wavelengths sequentially. This can be performed using devices like spectrometers, which analyze reflected or emitted light.

  • Snapshot: Captures all spectral data in a single snapshot, allowing for the simultaneous acquisition of multiple spectral bands for each pixel in the image. Ideal for moving objects, as it eliminates scanning time issues and motion-related distortions. However, it often requires more complex and expensive systems due to the need for advanced technology to capture all data simultaneously.


Practical Applications

Multispectral and hyperspectral images have applications in various fields, from remote monitoring of agricultural areas to disease detection, human tissue analysis during surgeries, evaluation of pharmaceutical homogeneity, fraud detection in artworks, forensic analysis, and much more.


On the left we see Imajan and on the side we see the Multispectral Microscope developed by OpenScience. Behind the equipment are some images captured with the equipment.
On the left we see Imajan and on the side we see the Multispectral Microscope developed by OpenScience. Behind the equipment are some images captured with the equipment.

Applications by OpenScience

OpenScience is a pioneer in developing innovative and accessible solutions, promoting the democratization of spectral analyses across diverse markets. Its standout technologies include:


Low-Cost Multispectral Microscopy

This advanced system combines high resolution with portability, making it ideal for resource-limited environments or field use. It has been used to identify adulterants in roasted and ground coffee, such as coffee husks, corn, barley, and açaí seeds. The equipment provides rapid and non-destructive analyses, effectively detecting impurities and fraud.


Hyperspectral Technology in the Coffee Production Chain

OpenScience developed a system that uses spectral images to classify coffee beans in the laboratory of a major Brazilian industry. This automated equipment performs detailed analyses of defects, impurities, beverage quality, moisture, and defect location directly from the image in approximately 30 seconds. The process is entirely objective, ensuring efficiency and precision.


Multispectral Equipment - Imajan

Imajan is a portable and economical multispectral device designed for analyzing grains and vegetables. It has been used to evaluate tomato quality and classify soybean, coffee, and bean grains, accurately identifying defects and impurities directly from captured images.


Conclusion

Spectral imaging represents a powerful analytical tool capable of providing rich information about the composition and properties of various materials on scales ranging from macro to microscopic. This approach is revolutionizing fields such as agriculture, the food industry, healthcare, geology, and others, enabling more precise diagnoses, more efficient processes, and higher-quality products.


OpenScience stands out as a leader in democratizing these technologies, offering accessible and impactful solutions. By connecting scientific innovation, technological development, and industrial applicability, the company reaffirms its commitment to making spectral analysis an essential resource for solving problems.


References

Sabin, G. P., Soares, F. L. F., Freitas, D. L. D. D., Silva, H. V. D. O., Antunes, C. M. M. M. O., Mohamed, E. A., Teixeira, C. A., Assis, C., Cardoso, V. G. K., & Volochen, M. (2024). Hyperspectral imaging applications. In F. A. N. Fernandes, S. Rodrigues, & E. G. A. Filho (Eds.), Chemometrics (pp. 91-123). Elsevier. https://doi.org/10.1016/B978-0-443-21493-6.00005-8

3 views

تعليقات


bottom of page