Hey everyone, we created StipraPOD (“Point of Disposal”), a network of AI powered smart bins able to identify products and generate consumption metrics so companies know their consumers better
It uses an Jetson to identify products, measuring their weight and temperature so we can see their level of consumption, and control the actuators to dispose the products trained
It was developed alongside London South Bank university and you can see a prototype here: https://www.youtube.com/watch?v=-L6kDVr5tfU&t=138s
We will deploying a full prototype (https://www.youtube.com/watch?v=n5fIKS_N6lc) at University of Bedforshire during 6 months to capture data and we are looking for people to help us developing it
Super cool, thanks for sharing Alvaro! Great application for a great purpose 👍
Looking forward to hearing and seeing more about this during the group meeting next week!
Glad you liked it Dustin, and look forward showing what’ve been concocting!
This is a really interesting approach. I like the idea of measuring both the weight and temperature instead of treating the bin as just an object counter. Having that information over time could make the system much more useful for understanding what is actually being thrown away.
I can also see the identification part becoming important when different types of products end up together. For example, something that looks like ordinary discarded material may need a completely different handling process if it contains electronics, batteries, or other components that shouldn’t go through the normal waste stream.
The actuator control is interesting too. If the system can identify the item and make the disposal decision at the same time, that could reduce a lot of manual sorting. I’d be curious to see how well it handles objects that are partially hidden or several different items appearing together.
It also makes me think that the identification step could eventually be useful beyond smart bins. Once an item has been identified, the system could potentially associate it with the proper disposal options instead of simply deciding which bin it belongs in. That would make the collected data useful for the next stage of the waste process as well.