Hyperspectral imaging use in industrial machine vision systems grows
Hyperspectral imagining uses a greater number of possible wavelengths of light and gathering data for individual pixels of an image, and is able to identify a large number of distinct colors.
Hyperspectral imaging is similar to other spectral imaging techniques in that it uses information from across the electromagnetic spectrum. However, hyperspectral imaging differentiates itself from other spectral imaging techniques by using a far greater number of possible wavelengths of light and gathering data for individual pixels of an image.
So while another spectral imaging system may identify just a small number of colors in an object in an image (often red, green, blue, and near-infrared), hyperspectral imaging may be able to identify tens or hundreds of distinct colors. In hyperspectral machine vision systems, the increased imaging capabilities can make the process much more specific, detecting differences and impurities inside an object, not just on the surface.
Hyperspectral imaging in manufacturing
Hyperspectral machine vision is rapidly growing in popularity because it can provide faster and more accurate results than human sorting. This results in a sorting process that is more efficient, less expensive, and offers improved quality control, all of which adds up to increased revenue.
Hyperspectral imaging has a range of diverse applications in different industries, including the classification of substances that have no visual differences but have different chemical compositions (like plastics), or for identifying and gathering data about substances that are not transparent to visible light but may be to other wavelengths like infrared light.
Some industry-specific applications of hyperspectral machine vision include:
- Fault detection in the pharmaceutical industry: While all pills in a certain batch may look the same, there may be impurities and factory defects present beneath the surface. A system with hyperspectral imaging can detect these and mark them from removal, thereby improving quality control.
- Security in food production: Hyperspectral machine vision can detect contamination such as maggots in rice, as well as help identify non-food objects like rocks or branches in batches of harvested vegetables. There are applications for detecting impurities or contamination in factory-produced food like cheese or sausage.
- Quality control for construction materials: Hyperspectral machine vision can be used for quality control in non-consumable products as well. It is especially effective for detecting impurities, damp spots, knotholes, or resin pockets in wood.
- Medical imaging: Hyperspectral imaging can be used to detect hidden cancer. Read more here about how the process works.
Hyperspectral machine vision uses sophisticated imaging technology to do what our eyes can’t. Its applications are far-reaching and will only continue to grow as the technology becomes more streamlined and readily available.
This article originally appeared in Vision Online. AIA is a part of the Association for Advancing Automation (A3), a CFE Media content partner.