Will Hyperspectral Imaging Work for My Application?
Answer: Feasibility Study
By Dr. Rand Swanson, CEO - May 30, 2024
Nature presents many challenges to those looking to solve problems:
High-quality ore may contain less than a gram of metal per ton, but it needs to be sorted from waste rock.
Diseased plants in a crop may be indistinguishable by eye from healthy plants, but early identification can lead to substantial cost savings.
The age of a lobster’s shell affects its flavor, shelf life, and price, but visually distinguishing between young and old shells is difficult.
Resonon provides feasibility studies on customer samples – even lobsters!
In each of these examples, a chemical difference between the objects of interest may alter the way light reflects or transmits through that material. With the aid of modern machine vision algorithms, hyperspectral imaging often enables the detection of these changes in light interaction, even when imperceptible to the human eye.
However, “often” is not “always”. Sometimes the spectral signature is too weak and other times – as with lobster shells – the objects have too much variability and complexity. For some problems, even with our years of experience, we cannot determine whether hyperspectral imaging will meet your needs without trying it first.
Feasibility Studies
At
Resonon, we encourage our customers to send samples to us to perform a
feasibility study.
If you can
name it, we’ve likely scanned a sample of it or something similar. Examples
include:
Vegetables, fruits, nuts,
grains, baked goods, sweets, meats, and cheeses, from fresh or raw to packaged
Ore and other raw
materials
Manufactured goods
and packaging
Pharmaceuticals
Currency (detecting
genuine vs. counterfeit)
Golf balls amongst
potatoes (a real problem in the potato industry)
Tuna (finding the
outline of the ‘blood meat’ to be trimmed off prior to packaging)
Lobsters
Genuine and counterfeit currency (both images are false-color HSI data)
Looking for golf balls amongst potato flesh (visible and false-color HSI data)
Finding the precise outline of the “bloodline” meat near the spine for automated removal (visible and false-color HSI data)
What Does a Feasibility Study Involve?
With over 20 years of hyperspectral imaging experience, Resonon has
developed the expertise and tools to accurately assess whether your questions can
be answered using hyperspectral imaging. Our Real-time Vision System (RVS)
software and hyperspectral cameras enable us to collect high-quality data
and then rapidly test different machine learning algorithms to determine the best
solution. After testing customer samples, Resonon provides a report detailing
how well our hyperspectral system addresses the specific question you need
answered.
Some applications
(including lobster shell age prediction) do not work, and our report will explain
why.
However, many
applications are feasible. Our Application Engineers can discuss their findings
and help you plan the next steps.
Wet signatures vs. copied signatures (visible and classified via HSI data)
How to Get Your Samples Tested
Sometimes, we
can provide an answer without seeing samples. For instance, detecting an object
buried inside concrete will not work, and we’ll tell you as much to save you
the effort of sending samples.
Most of the
time, though, we will coordinate with you to obtain samples, store and handle them
properly, and complete a feasibility study. Turnaround times are often under
two weeks.
If you have a
challenging problem and wonder if hyperspectral imaging might “see” the answer,
reach out to a member of our Sales Team.
Dr. Rand Swanson, CEO
Dr. Rand Swanson, CEO of Resonon
Rand Swanson is the Chief Executive Officer and co-founder of Resonon. Along with overseeing company management, Rand applies his decades of expertise in optics to optical design and radiometric modeling for commercial and R&D projects.
He considers himself fortunate to collaborate with the talented and dynamic Resonon team and to focus his efforts on hyperspectral imaging—a technology that is simple in concept, complex in execution, and applicable to a vast array of uses.
