Artificial Intelligence for cameras on Industry


Nowadays, it is possible to use cameras to recognize people behavior (working, stealing, running, smoking), motion patterns, physical metrics (age, size, gender). Using AI on cameras, what different analytics and use-cases an industrial manufacter can take leverage for business intelligence (security, management, optimization of work)?

Industry 4.0
Data Analytics
Artificial Intelligence
Ricardo Santos
57 months ago

2 answers


Something that I think could really benefit from combining AI and cameras is the efficiency of elevators. Not only making elevators more efficient in operation for the user, but potentially reducing operational costs (e.g. electricity costs). Let me explain...
If you travel a lot like me, and stay in hotels, you may have shared this experience more than just a handful of times. Have you ever waited for an elevator for several minutes during the morning breakfast rush or check out time period. And did that elevator arrive to your floor, filled with so many people, that you could not get on? Or were you in that elevator, and got to enjoy stopping on every single floor, all the way down to the lobby, just to have the doors open and close and then smile with a grin because nobody was getting on the elevator. I've been in both situations, many times.
Well, here is my solution. There are cameras on most, if not all elevators these days. Through AI technology and image recognition, it should be fairly straight forward for technology to determine when the elevator is full (by analyzing available floor space. i.e. whether or not another person could stand in the elevator). By doing this, the elevator would know that it was full and rather than stop to pick up new passengers on each floor being called, it would just skip those floors and drop off people at the lobby or whatever floor selection they had set. It obviously makes no sense to stop on each floor, with a fully packed elevator, unless someone on that elevator pushes the button to get off on that floor as well. For customs satisification, the hotel or building operator could even set the capacity. Perhaps they want a comfortable experience for their clients, set max to say 90% full. This way the elevator would consider it full and give everyone an enjoyable ride down to the lobby. This does not only apply to hotels, but office buildings, schools, etc.
I believe by using the existing hardware (e.g. the installed camera) paired with software to perform the analysis and provide feedback to the elevator system, you could have smart AI enabled elevators which would make moving people between floors much more effeciently than they are doing today.
Just one use case, some what a selfish one, as I would just like to get out of the hotel faster. ;)

56 months ago
That's actually a great idea, thanks for your reply! - Ricardo 55 months ago

Cameras can play a major role in the Industry 4.0 journey. Especially at the beginning can they help to bring data to data collection systems which previously would be unable to connect or would require retrofitting projects. For example if you can use a camera to look at instruments which have been used no connection to my existing systems. And can use picture analysis to read the instrument iI can send this value to further systems like predictive Maintenance. This may even be necessary option if iI am unable or have other restrictions preventing me from introducing IoT enabled devices. Other options for picture analysis which cameras are especially good for can be for optical quality analysis.
Also for security in terms of prevention of unauthorized access either of the perimeter or sections of a plant it can be used by not just showing it on a screen but feeding it to an AI or machine learning system to create alarms or trigger responses. Also using it for other maintenance activities like leak detection or other problems usually detected by visual check. It also can be used to assist other initiatives like showing user specific information based on the location of an operator and his identity. Normally you would use GPS but if this does not work cameras could assist there. There may be more cases but right now this is what comes immediately into my mind.

Stefan Zippel
56 months ago
Thanks Stefan. You did a really good point, industries are leveraging more and more artificial vision analysis for quality control, since a camera can detect defects faster and more reliable than a person. - Ricardo 55 months ago

Have some input?