The Importance of Optical Filters for Machine Vision

Tradeshow Talks with Midwest Optical

Why are optical filters useful in machine vision?

The primary role of filters as they're used in machine vision can be broken down into three benefits. The first is contrast, filters help to increase the contrast of an imaging application. The second is that they help with resolution, improving the resolution of your lens in your camera. The last benefit is that it increases your chance of repeatability. Using a filter in your application typically ensures that the application is going to work, regardless of the environment.

With contrast, in a typical machine vision application what you're trying to do is create contrast, you're trying to pick something out of something else, or pick something out of the foreground or the background. For example, say it's red and blue bottle caps. In a color image you can easily pick out the red and blue bottle caps, but a more efficient way to do it would be to use a monochrome camera, and a band-pass filter. If you have a monochrome camera and a red band-pass filter, the red wavelengths are passed and in the system those red wavelengths appear to be highlighted, they appear to be white. Any of the blue wavelengths which are blocked, appear to be darker.  “Why would you use a monochrome camera as opposed to a color camera in this situation?” It’s a lot cheaper and more efficient use of resources. The filters are great at creating contrast.

For the second benefit in a monochrome system, filters are helpful in increasing your resolution. There's a type of distortion that happens in lenses called chromatic aberration. That happens when a lens is trying to focus several wavelengths at the same point. This is because different wavelengths have different refractive indexes, they're not able to fully focus the same point which leads to the distortion.

However, most of the time in machine vision applications, you're only looking for a single color or maybe two colors, thus by using a filter to limit the number of incoming wavelengths, you allow the lens not to work as hard to try to focus all the wavelengths

In the previous example, a red band-pass filter was used so if you only care about red, you can limit the other wavelengths and just pass red and you can increase the resolution in that way.

What industries does that make it particularly suited for?

Whenever you're using a vision system to look at something, so whether it be sorting, looking at fill levels, packaging, inspection, metrology, 3D metrology. When it’s applied to 3D metrology we found it is popular in the automotive manufacturing industry because they use lasers to get very high precision measurements. They then use those measurements to recreate a part repeatedly, whether it be a door or a body panel or a switch that goes inside of a car. They use these lasers to take precise measurements with really tight tolerances. The laser and the sensor are looking for this laser line and we use a filter to block everything out except what you want the laser to find.

What specific capabilities make you stand out in this field?

We have five key features that we talk about that kind of make us stand out against our competitors. You can buy a filter from a lot of different places but our focus is filters. The first key feature is something called wavelength control. When we talk about a band-pass filter, if light enters a filter at an angle that passband, that curved part of the transmission that is allowing certain wavelengths through can shift. Typical manufacturers make their filters by taking glass and coating it to create that passband, we do it a little bit differently. We use a hybrid design where we use filtered glass and a coating. What happens with our filters is that passband doesn't shift, it stays stable. We call that technology StablEDGE, it's unique to our filters, you're not going to find it elsewhere.

The other aspect is the performance of the actual passband. We designed our passband to have a bell-shaped curve which we call the Gaussian curve. We find that this is the most efficient way to capture the output of the LED that you're using for inspection because the LED is also being outputted with the same shape, the bell-shape. Other filter manufacturers, they're going to have a flat-topped design. It almost looks like a rectangle. There are some issues with that because it's just not an efficient way to capture it.

The fact that we use A/R coatings on almost all our filters. When light enters glass, there's an immediate refraction that happens meaning there's some light loss that happens. The A/R coating that we apply prevents a lot of that light from being lost.

Also, the way we inspect our filters, we're the only filter company that does 100% inspection of the filters. We inspect both their performance, so you know if we say that there's going to be this wavelength that gets passed, and this wavelength that gets blocked, you can be sure that's going to happen. We also inspect the actual substrate, the glass itself, for any scratches or dings to a very high tolerance, more so than any of our competitors.

We take care of things such as when a filter is inside of the mount, the retaining ring that holds that filter inside that mount. If that retaining ring is tightened too much or too much torque is applied, it can cause distortion in the filter. It can cause a distorted image in your result which is a level of inspection that you're not going to find with any other filter manufacturer.

Lastly, the different variety that we offer to mount the filters to your lens is vast, we mount different shapes and sizes with low leakage.

What customers have been able to take advantage of this?

An example of a customer we’ve helped is one who set up lights at a football stadium, they were color cameras and they were set up around the perimeter looking at the football field. The image had a weird pink hue over everything, it didn't look right, it didn't have the natural color that our human eyes are used to. It turned out though that the weird pinkish hue was caused by the stadium lights giving off a lot of infrared, which our eyes can't see but the camera was picking up as well as the infrared that was being given off by the sun. We used a shortpass filter that blocks the IR but passes the visible light. It produced a natural color image that our human eyes are used to.

What are you looking to develop further?

As you look at the industry, we're seeing the cost of infrared cameras, or InGas cameras going down. We're seeing more and more uses for infrared and because of this we want to create bandpass, shortpass, longpass filters that support those short-wave infrared wavelengths. We're seeing lots of benefits with those wavelengths like, absorption of moisture, lidar or autonomous vehicles, food inspection. We're seeing those variety of wavelengths being used in all those applications and we want to have filters that really support those type of applications.

Citations

Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Midwest Optical Systems, Inc.. (2019, January 30). The Importance of Optical Filters for Machine Vision. AZoOptics. Retrieved on November 23, 2024 from https://www.azooptics.com/Article.aspx?ArticleID=1487.

  • MLA

    Midwest Optical Systems, Inc.. "The Importance of Optical Filters for Machine Vision". AZoOptics. 23 November 2024. <https://www.azooptics.com/Article.aspx?ArticleID=1487>.

  • Chicago

    Midwest Optical Systems, Inc.. "The Importance of Optical Filters for Machine Vision". AZoOptics. https://www.azooptics.com/Article.aspx?ArticleID=1487. (accessed November 23, 2024).

  • Harvard

    Midwest Optical Systems, Inc.. 2019. The Importance of Optical Filters for Machine Vision. AZoOptics, viewed 23 November 2024, https://www.azooptics.com/Article.aspx?ArticleID=1487.

Tell Us What You Think

Do you have a review, update or anything you would like to add to this article?

Leave your feedback
Your comment type
Submit

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.