When should you choose a bandpass filter for a machine vision application?

|K WONG

Introduction: What is a Bandpass Filter?

In machine vision, the camera needs a clear, high-contrast image to do its job correctly. However, cameras often pick up extra light that confuses the system. A bandpass filter is a special piece of glass or plastic placed over the camera lens that solves this problem. Think of it like a strict bouncer at a door: it only lets a very specific "band" (or color) of light pass through to the camera sensor, while completely blocking out all other light.

How Does a Bandpass Filter Work?

Light travels in waves, and different colors have different wavelengths. A bandpass filter is designed to be completely transparent to one specific range of wavelengths, and acting like a brick wall to everything else.

If you have a machine vision setup using a red LED light to illuminate a part, you would place a red bandpass filter on the camera. The filter allows the red LED light bouncing off the part to enter the camera, but blocks the blue light, green light, and UV light from the surrounding environment.

Top Reasons to Choose a Bandpass Filter for Machine Vision

You should generally choose a bandpass filter when your system is struggling with lighting consistency or contrast. Here are the specific scenarios where they are essential:

  • Blocking Unwanted Ambient Light: This is the most common reason to use a bandpass filter. Factory floors have overhead lights, sunlight coming through windows, and even sparks from nearby welding. This ambient light changes throughout the day and can wash out a machine vision image. By using a specific colored light (like a blue LED) and a matching blue bandpass filter, the camera only sees the light you are providing. The sun and factory lights become practically invisible to the camera.
  • Maximizing Image Contrast: Machine vision is all about contrast—making the feature you want to inspect stand out from the background. If you want to highlight a red object on a green background, you can shine a red light on it and use a red bandpass filter. The red object will appear bright white to a monochrome camera, and the green background will appear pitch black, creating perfect contrast.
  • Inspecting Specific Colors or Features: Sometimes, you need to check if a specific colored label is present, or if a liquid has a certain chemical hue. A bandpass filter tuned to that exact color ensures the camera is only analyzing the specific feature you care about, ignoring the rest of the product.
  • Using Lasers or Highly Specific Light Sources: Lasers emit a very narrow, pure band of light. If your machine vision system uses a laser to measure 3D profiles or scan barcodes, a very narrow bandpass filter is practically mandatory. It ensures the camera only sees the laser line and nothing else.

Key Terms to Know When Choosing a Filter

When you are ready to pick out a filter, you will run into two main terms:

  • Center Wavelength (CWL): This is the exact center of the light "window" you want to let through. If you are using an 850nm infrared light, you want a filter with an 850nm CWL.
  • Bandwidth / Full Width at Half Maximum (FWHM): This simply means how wide the "window" is. A wide bandwidth (like 100nm) lets in a broader range of light, which is good for general LED lighting. A narrow bandwidth (like 10nm) only lets in a very tight slice of light, which is perfect for lasers.

Summary

You should choose a bandpass filter anytime you need to control what the camera sees. They are the easiest and most effective way to eliminate confusing background light, boost image contrast, and make your machine vision system much more reliable.

 

0 comments

Leave a comment

Please note, comments need to be approved before they are published.