Problem Statement

You have production equipment that perform key functions. The actions aren’t hidden, but they aren’t easy to see. They may reside behind shrouds, they may be so repetitive that no one can stand to watch them for long, or they may sit outside the normal line of sight of busy workers. The equipment sometimes fail or behave strangely but no one sees it happen. If someone were watching, they might have picked up valuable clues.


You installed HD cameras to permanently watch these important points of interest. The cameras are well-positioned and the camera output is continuously recorded on a server with enough resolution and frame rate to see what is happening.


This system will record everything that happens at the location … and keep it for days, weeks or months. If something unusual happens, you can review the recorded video for clues to explain the problem. As security camera technology rapidly improves in speed and resolution, even inexpensive cameras can capture surprising levels of detail. With suitable software (e.g., good video editors or Dartfish), you can enhance the video to extract even more insight and information. For example:

  • Add a visible marker to a given production unit. Time-shift the video clips so the marker is visible in all frames simultaneously. You can now see a given production unit in ALL stages of production simultaneously.
    Whether this will help solve a given production problem will depend on the process. Regardless, it is a perspective on production that you have probably never seen before. It can’t hurt! Note that this video example is just a mockup.
  • Overlay video to see physical variation. Take video from different time periods and blend and overlay them with images that are partly transparent. Process variation is evident and visible. If you run the video slowly, the detail jumps out.

We can investigate more deeply into this phenomenon if we slow down the video and watch it in slow motion. We might see, for example, if one deposit point is timed differently than the others. We might see where excess material buildup is causing problems. Or we might see it running just fine and that would be OK.

  • Measure things directly from the video. We synchronized and blended four video clips and added a measuring scale. We can stop the video as it passes the scale and measure the variation. There is a blog article with details on the FluidProjects site.

This video and the companion analysis shows how production video can open up new measurements and new opportunities for detailed process understanding.

These examples are possible with standard (HD) cameras that watch and record continuously. If something unexpected happens, you have a place to look for clues to understand the problem and prevent it from recurring. Surely, that is the essence of continuous improvement.