Q&A: Optical filter manufacturer expands into the machine vision world

May 19 2016
James Carroll, Senior Web Editor

The following is a conversation with Jeff Carmichael, Technical and Product Marketing Manager, Chroma Technology Corp.

I saw your booth at The Vision Show 2016 in Boston and you walked me through your product demonstration.  What prompted you to introduce a line of machine vision filters now?

Well, we’ve been looking at the machine vision and factory automation markets for a while and we feel that our expertise allows us to offer something that was lacking for those seeking improved performance.

Many machine vision applications already use filters though, so what sort of improved performance are you referencing?

Chroma has been manufacturing optical filters for 25 years.  Although we’ve supplied filters in many industrial markets, our primary markets have been in the life sciences, and the primary imaging modalities that we’ve served have typically involved fluorescence detection of some kind.  We continue to work with all of the major microscope manufacturers as well as manufacturers in markets like flow cytometry, DNA sequencing, High-content-analysis, Raman spectroscopy, surgical instrumentation, and more recently the medical point of care industry.

Most of these applications typically require properties not found in available filters marketed to machine vision applications.  Properties like extremely high and broad levels of out-of-band attenuation – called “blocking”, very steep transition slopes, and very high levels of transmission to detect low signal intensities in the presence of high noise levels.  You’ll typically see >95% transmission across the full width of a square-wave shaped passband.

We know that most of these filters are over-built in terms of what’s generally required in machine vision, but on the other hand we’re seeing more demanding tasks that require improved contrast and greater levels of blocking of unwanted light.  We think there’s also a need for interference filters that don’t suffer from spectral blue-shift or collapse of bandshape when working with larger viewing angles. 

So it looks like evolving vision applications will require ever-increasing levels of performance, and that’s right in our wheelhouse.

What is it about these filters that you think will make them suited for machine vision?

We’ve specified 17 single-band filters to start with which correspond to the most popular LED wavelengths and the more common laser diodes.  Most passbands are 40nm-wide, while the narrower filters are 20nm-wide.

The main thing that’s different from our other catalog filters is that these are designed to accept a approximately 20° half-cone angle.  You’ll often hear that interference filters are very angle-dependent, but they don’t have to be.  The stacks of thin film layers can be built in a way that accommodates a wide range of angles.

This means that when using lenses with angular fields of view of 40-50°, you don’t suffer any real loss in performance in terms of contrast from edge-to-edge.  The passbands hold pretty well over that range.  We thought that made sense to start with, and then we can always improve on that if there’s a need for even greater angle acceptance.

So the idea is that this now allows you to benefit from the other advantages that sputtered interference filters already provide, but while also ensuring dependable performance over a wide range of angles common to vision applications.

Those other advantages include the >20% of increased contrast that these highly transmitting square-wave interference filters provide when compared with more traditional absorption glass filters. Traditional passbands are Gaussian in shape.

Square-wave interference passbands transmit less unwanted light and more desired light.  Look at a graph of the transmission – it’s all about the area under the curve when calculating both the signal and the noise.  But you can’t forget to also look at the Optical Density plot, because that determines how well the filter blocks what you don’t want your camera to detect.

We’ve relaxed the blocking compared with our fluorescence filters, although to achieve the steepness that we think is important, higher blocking is sort of a by-product.  More layers in the filter coating can result in both properties at the same time, and also allows for other design considerations, like the increased cone angle acceptance.

Where do you see Chroma in the future of the vision industry?

That’s one of the things that I was hoping to get a better feel for by being at the Vision Show.  Regarding the life sciences side of it, although there is plenty of potential common ground between what I think of as machine vision and life science imaging applications, I don’t think that it’s evolved much yet.  There’s some overlap, but I think that market is in its early stages.

As an example, we exhibited at AIA’s Vision in Life Sciences meeting in San Diego last November.  I saw traditional machine vision applications that just happened to be applied to life science products, such as measuring fill levels in pharmaceutical products, or inspecting test tubes for use in clinical practice.  But to me, that’s basic machine vision, and not really life sciences.

And then I saw some really advanced life science imaging that to me didn’t seem like it had much to do with machine vision.  I didn’t really notice a lot of overlap.  I think that’s because the various industries are just finding out what the other is all about and investigating potential opportunities, just like we were doing.

Then, we were asked to give a talk about using optical filters in the life sciences at the Vision Show, and it seems that there’s already more of a synergy compared with last year.

So I see Chroma continuing to grow in areas where there is overlap, like point-of-care products and surgical instrumentation, and we’ll also continue to serve the scientific, clinical, and diagnostic biomedical markets that we’ve long served.

But I’m hoping that Chroma’s filters will also fill a need for better contrast in more traditional machine vision applications. I also see Chroma satisfying a demand for higher levels of blocking of unwanted light from things like lasers, high powered LEDs, and other very bright light sources that you don’t want your camera to see.

Article Gallery

Jeff Carmichael, Technical and Product Marketing ManagerChroma Red, Blue and Green Filters

Media Inquiries

If you are a member of the press and can’t find something you need—or want to speak with someone from Chroma please contact us at sales@chroma.com.

Keep In Touch

Sign up to receive updates by e-mail: