When most manufacturing companies, indeed most companies of any type, think about “video surveillance”, they automatically think about security, theft and personnel monitoring. Then, they buy security systems and video surveillance systems to collect massive quantities of video like the following examples. Feel free to watch an entire video. You can see that absolutely nothing of interest occurs. … 🙂
Nothing happening in the Lobby
Nothing Happening at Back Door
These are not a particularly good video recordings. The native resolution is only 352×240 at 7 frames per second. They were captured with one of those systems … the kind you see at Best Buy … you know, the package deal with a base station and 8 cameras for just $449.00 ! Nonetheless, I wager that virtually every serious factory, trucking depot, or business has a system at least this good … and often far, far better. Serious systems can run into the 10s and even 100s of thousands of dollars. Of course, if nothing happens, the limited quality is probably not that evident.
So what? … Well compare the next two videos with the first two.
An Emergent Bottleneck in Production
An Additional, Temporary In-Process Step
Both of these videos were recorded with the same surveillance system. They are the same lousy recordings, but now the cameras are pointed at key points in the production process. The one on the left shows final manual sorting for a pretzel bun line. The sudden buildup of buns is surprising. How did it happen? The video on the right shows a “pop-up” manual processing step. It isn’t a permanent operation because it is only applies to a few items that are run sporadically. As a result, the staff might have to re-learn things if it has been a while. How do you improve occasional processes like this?
For any operations professional worth their salt, these are almost fun to watch. What happened? How did the bottleneck emerge? What was the root cause? How fragile is this process? Wow, that line blew up in a hurry!
Fess up. If you had videos like this for all of your processes … especially if they had far better quality … wouldn’t you be sneaking a peek at the footage from time to time? Like maybe after that day when throughput was so horrible? Or maybe after that morning where those weird defects appeared at final inspection? Or maybe just to see if you could spot a way to tweak a bit more output or yield from that problem-child line?
Taking the idea a bit further, isn’t it odd that companies automatically invest in security surveillance, but they don’t think to invest in process surveillance? That leads to a question: Where are you losing the most money?
- To theft and extra-long smoke breaks?
- To waste, muda, downtime, delays, shutdowns, mistakes, line imbalances and other operational losses?
I would be shocked if any competent operations professional would not see process problems as their number one avoidable cost. Yet, for the vast majority of companies, security merits a video surveillance system and operations do not.
At DeeperPoint, we think there is a world of opportunity to use mundane, ordinary security system technology to deeply empower process managers. The technology is relatively cheap and risk-free. The real challenge is to develop new analysis, methods and techniques to efficiently mine the huge quantities of video that will inevitably result. That’s what we’ve been working on and continue to explore.
So far, we don’t see this as being about flashy new technologies. If something in your process can benefit from fully automated viewing, you are probably looking for a machine vision solution. Those can be very productive, but they usually have a tight focus and are almost always pretty expensive.
We are talking about monitoring your general operations continuously, from many overlapping viewpoints. This is a video record that would allow you to “go back in time” to see your production process as it actually ran at any moment in the recent past. This type of system won’t see every problem, but it would give a comprehensive view of the general activity at the time. Then, with experience, you should be able to point more cameras at specific, critical or problem operations. Any decent HD camera (which is about everything that is sold these days), will generate videos with much greater detail and clarity. To see the full quality, use the control to open these videos full screen.
Closeup of a Manual Operation
Visible Process Variation without SPC charts
The video on the left shows a manual task and the video on the right shows how a camera can be placed on a purely machine operation. The extra marks (scales and numbering) on the dough ball conveyor were added after the fact. In addition, the fuzzy images are the result of taking several segments of video and overlaying them with transparency. You can clearly visualize the process variation that is happening. This is not a view that would be available from a standard surveillance system, but DeeperPoint can build these pretty mechanically. With video recordings like this from a variety of angles, on many key operations, there will be opportunities to apply many of the techniques that we previously described in this Blog. For example:
- Visualizing Process Variation
- Extracting Process Statistics from Video
- Analyzing Machine Operation
- and anything else we can imagine
These are the areas where DeeperPoint can help. We can help you find the opportunities for process improvement. We can help you find other uses for the video that you record:
- What about offering a video process tour to engineers in peer plants elsewhere in your organization? Problem-solving on the phone has to be more productive if both parties are viewing a video of the process in question. Perhaps even viewing a video of the process at the time the problem occurred.
- Can you use videos like the “Emergent Bottleneck” example in new hire training? Show some clips from recent production activity. Use them to explain concepts like bottlenecks and buffers and …
- Can you create job instructions from videos of well-functioning work stations? If you point a camera at a task being performed well (like the label application example), you shouldn’t have to add a lot of other explanation to make the requirements clear.
- What else can you/we imagine?
Think about it. It wouldn’t cost much to try. With a bit of help, you might find some really useful nuggets of value. Or put another way:
Sports teams routinely “watch the films” after a game. That is their main tool for problem-solving and continuous improvement. Why couldn’t it be the same for you?
Finally, I almost labeled the first Emergent Bottleneck video: “Lucy in the Chocolate Factory” after this iconic comedy sketch.