HyperSpectral Imaging

Hyperspectral Imaging

Hyperspectral Imaging Spectrometers can acquire spectra at a pixel, over a line, or over an entire plane to produce a three-dimensional hyper cube of spectral information as shown above. Applications include chemical analysis, pharmaceutical identification, vegetation growth monitoring, forensic detection of harmful substances, and atmospheric studies/monitoring.

Hyperspectral imaging is an advanced remote sensing technique that combines imaging and spectroscopy to capture and analyze detailed spectral information of a scene or object. Unlike traditional imaging techniques that capture color information in just a few broad bands (such as red, green, and blue), hyperspectral imaging records data in hundreds or even thousands of narrow, contiguous spectral bands across the electromagnetic spectrum.

The key components and characteristics of hyperspectral imaging are as follows:

Spectral Bands: Hyperspectral sensors capture a wide range of spectral bands, typically spanning from the ultraviolet (UV) to the shortwave infrared (SWIR) region. Each band represents a specific wavelength or narrow range of wavelengths, providing detailed spectral information about the objects or materials in the scene.
Spatial Resolution: Hyperspectral imaging combines spectral information with high spatial resolution. It captures detailed spatial features and textures of the scene, allowing for precise identification and analysis of objects.
Spectral Signature: The spectral signature of an object refers to its unique spectral response or reflectance across the captured spectral bands. Hyperspectral imaging enables the characterization and differentiation of materials based on their spectral signatures, aiding in various applications such as vegetation monitoring, mineral exploration, and target detection.
Material Identification: By analyzing the spectral signatures, hyperspectral imaging enables the identification and classification of materials within a scene. It can discriminate between different types of vegetation, geological formations, land cover classes, and other objects of interest based on their distinct spectral properties.
Subpixel Analysis: Hyperspectral imaging allows for subpixel analysis, meaning it can identify and analyze materials within individual pixels, even if they occupy a small fraction of the pixel area. This capability is particularly valuable in applications such as vegetation health assessment, where subtle changes within a pixel can be indicative of specific conditions or diseases.
Applications: Hyperspectral imaging finds applications in various fields, including agriculture, environmental monitoring, geology, forestry, food quality assessment, defense, and remote sensing. It provides valuable information for land cover mapping, crop health monitoring, mineral exploration, pollution detection, object recognition, and more.

To acquire hyperspectral data, specialized sensors or cameras are used, typically mounted on aircraft or satellites. These sensors capture the electromagnetic radiation reflected or emitted from the scene, with each spectral band recorded as a separate image. Advanced data processing techniques, including spectral analysis, classification algorithms, and dimensionality reduction, are then applied to extract meaningful information from the vast amount of hyperspectral data.

Overall, hyperspectral imaging offers a powerful tool for precise and detailed analysis of materials and environments, enabling a wide range of applications across various disciplines.

Types of Hyperspectral Imagers

Different types of hyperspectral imagers are used to capture hyperspectral data, each with its own advantages and disadvantages. Here is an overview of some common types of hyperspectral imagers (see Figures next page):

Point Scan Imagers: Point scan imagers use a fiber probe with built-in illumination and receiver to sequentially scan different spatial locations in a scene. The probe is placed on the object of interest, and spectral data is collected point by point.
1. Advantages:
  1. High spectral resolution (up to 8000 wavelengths).
  2. Compact and portable design.
  3. Can be used for targeted measurements on specific points of interest.
  4. Suitable for small-scale or localized applications.
2. Disadvantages:
  1. Slow data acquisition due to the sequential scanning process.
  2. Limited spatial coverage.
  3. Challenging to capture wide-area scenes efficiently.
Push Broom Imagers: Push broom imagers employ linear detectors that capture hyperspectral data along a single line or a linear array of spatial locations. The imager is moved across the scene to capture continuous spectral information.
1. Advantages:.
  1. Medium Spectral resolution (up to 200 wavelengths).
  2. Fast data acquisition due to continuous scanning.
  3. Efficient for capturing wide-area scenes.
  4. Allows for high spatial resolution.
2. Disadvantages:
  1. Requires mechanical movement, which may introduce motion artifacts.
  2. Sensitive to sensor misalignment.
  3. Prone to spectral distortions caused by sensor non-uniformity.
Camera Snapshot Imagers: Camera snapshot imagers use area detectors, such as CCD or CMOS sensors, to capture the entire hyperspectral image in a single snapshot. Each pixel on the sensor records the spectral information at a specific wavelength.
1. Advantages:
  1. Medium to low spatial resolution.
  2. Fast data acquisition as the full image is captured simultaneously.
  3. Suitable for capturing dynamic scenes.
  4. Enables high spatial and spectral resolution.
2. Disadvantages:
  1. Limited spatial coverage depending on the sensor size.
  2. May require sophisticated calibration to correct for sensor non-uniformity.
  3. Generally more expensive compared to other imaging techniques.
Monochrome Camera with Color Filter Wheel: This approach uses a monochrome camera combined with a color filter wheel to sequentially capture images at different wavelengths. The filter wheel rotates to allow the camera to capture the hyperspectral data at various spectral bands.
1. Advantages:
  1. Relatively low-cost setup compared to other hyperspectral imagers.
  2. Can achieve good spatial resolution.
  3. Allows for flexibility in selecting specific wavelengths of interest.
2. Disadvantages:
  1. Limited spectral resolution compared to dedicated hyperspectral imagers.
  2. Slower data acquisition due to sequential image capture.
  3. Prone to motion artifacts if the scene or object is not stationary.

It's worth noting that these are just a few examples of hyperspectral imaging techniques, and there are other variations and hybrid approaches available. The choice of hyperspectral imager depends on factors such as the specific application, desired spatial and spectral resolutions, data acquisition speed, and budget constraints.

What We Do

A typical 8MP digital camera produces 24,000,000 bytes of information. The same Hyperspectral image with 2048 spectra per pixel and two bytes per spectra, will produce 32,768,000,000 bytes of information. That's 1300 times more information per image. With our Spectra Insight Software, this data can be acquired, stored, processed, visualized, and analyzed in near-real-time. In addition any spectra within an image can be compared with local and/or 3rd party spectral databases for substance identification. See our Gallery for examples.

Spectra Insight supports all Hyperspectral Imager Types as follows:

Point Scan Systems

We incorporate a spectrometer of your choice with a fiber probe into our area scanners. Our software then automatically scans and acquires spectral data over the area-of-interest.
Our Spectra Insight Software interfaces directly to the hardware acquiring data at high speeds for near-real time analysis.

Area size is dependent on the scanner. Contact us for area scans larger than our standard products.

Line Scan Systems

We incorporate a spectrometer of your choice with a conveyer platform. Our software then automatically scans and acquires spectral data over the area-of-interest.
Our Spectra Insight Software interfaces directly to the hardware acquiring data at high speeds for near-real time analysis.

Wavelength Scan and Snapshot Systems

Our Spectra Insight Software interfaces directly to the hardware acquiring data at high speeds for near-real time analysis.

Summary

Virtually any camera hyperspectral imager/spectrometer can be supported.
We provide complete integrated turnkey systems with scanner, spectrometer/imager, software, and computer.
See our list of recommended packaged hyperspectral imaging systems.