Results with Machine Vision Filters
Comparison to Photographic Filters
Comparison to Narrow Band Filters
Recommended Filters for Machine Vision
Kits for testing Filters and Lighting

What is the difference between a machine vision filter and a photographic filter?

When it comes to machine vision, the ideal filter should be an immediate solution that provides greater contrast, improved transmission and resolution, and long term control over the variability of ambient light.

For more than 100 years, photographers have been using filters to reduce reflections, balance the color of a scene and bring out contrast in black-and-white photos. With similar results in mind, integrators all over the world have attempted to use photographic filters in industrial vision systems. In most cases, some improvement can be seen when using these types of filters, however the problem with photographic filters is not just in adapting their (usually) larger sizes to smaller CCTV lenses but in the type of filtering created. These filters were originally intended for use with film cameras and they have not changed.

The spectral response (sensitivity) of film is from 400-700nm, i.e., the visible spectrum. Most all CCD/CMOS cameras are sensitive in the ultraviolet (UV), visible, and near-infrared (NIR) portions of the spectrum. In order to take control of lighting conditions and image quality, filters are required that take this into account and perform well over the entire sensitivity range.

Comparison of human eye, film and CCD photopic response as measured in Quantum Efficiency.

We have long recognized that most photographic filters are far from ideal for use in digital imaging and so designed a full line of filters specifically for most common machine vision applications.

As shown by the examples on the right, the improvement in contrast can be significant when a filter designed for industrial vision is used instead of a traditional photographic filter.

The graphs illustrate the spectral response of a typical CCD/CMOS sensor, output from common machine vision LED lighting and performance characteristics of photographic filters currently offered elsewhere as "Industrial Filters" v/s MidOpt's machine vision filters.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
This material is covered in our printed Resource Book but with much better quality and resolution. Call to request your copy today.

Filter comparisons:

Original unfiltered image of printed circuit board in white light.



(upper chart) The photographic filter reduces light intensity from green LEDs by 33% or more and passes more of the unwanted shorter and longer wavelengths, decreasing overall contrast.
(lower chart) The MidOpt BP525 filter blocks all unwanted wavelengths while passing 95% of the LED's spectral output, creating maximum contrast and control of ambient light conditions.

Original unfiltered image of date and batch code on coffee can in white light.



(upper chart) The photographic (photographic) filter blocks only the lower wavelengths; the unwanted longer wavelengths greatly reduce contrast and overall image quality.
(lowerchart) Our BP660 filter is designed specifically for 660nm dark red LEDs that actually peak at 660nm +/-10nm, providing maximum contrast for system speed and accuracy regardless of the LED's specific output.

UV fluorescence of date and batch code on motor oil can. Original unfiltered image of in white light.



(upper chart) The photographic filter commonly sold for fluorescence applications does not block the overpowering UV LED light source and passes more than just the blue, greatly reducing the ability to detect the typically weak blue "glow" from the subject.
(lower chart) The BP470 filter blocks the powerful UV source and passes only the required blue portion of the spectrum, insuring maximum contrast when imaging any blue fluorescence.