Amsterdam’s AI Waste Solution Underscores Growing Trend

every day the world generates about 2 billion tons of trash and cities are struggling to keep up traditionally cities manage waste with technology by using sensors mainly placed in waste bins to monitor when the bins are full in amsterdam the city has gone a step further equipping all of its waste pickup trucks with sensors to measure the weight of the trash but it's still not uncommon to see this. In Amsterdam, the littering is so bad that the city planted many gardens around some bins in 2020 to deter it. This after the city contracted two private companies to remove the extra bulky waste. It still wasn't enough. In some neighborhoods people do not know how to contact the city so there we wouldn't know like if there's an issue so we decided hey with this new computer phishing technology that's been advancing. quickly over the last few years. Maybe we can just drive around and scan the streets for issues. This is a look at Amsterdam's object detection kit. Using computer vision it underscores a growing trend within the smart waste market. So you can see here this is the model is detecting like furniture, wood, cardboard. It's an open source tool where we're attaching cameras to vehicles that are already driving around so it's not new vehicles that drive. around. And basically, they scan the streets and see, hey, there's there's something lying here. It's a garbage bag that needs to be picked up. There are some bulky waste here that needs to be picked up. The solution runs off smartphones embedded with computer vision. They collected data, then get sent to the server, where it's analyzed using the software YOLO. The processing of the images is being done in the cloud because we have to gather training data. But later on, you might imagine it's better for the privacy. concerns if something would be processed in the camera. So we have like devices like this, like the Jetson Nano. We can already make the models so small that we can process like two or four camera feeds with one of these. It's that kind of flexibility that will allow for computer vision to be more widely adopted in cities, especially with the rise of autonomous cars. With autonomous vehicles coming to the city, there will be so many cameras everywhere driving around. So what if this can be a new way of to see where something is broken, see dangerous situations, see garbage that needs to be picked up? It comes as a growing number of vendors use computer vision to manage waste and other concerns. Rubicon uses an app to direct waste haulers on their trash routes, collecting insights in the fleet's operations and recycling rates. The platform also uses machine learning and AI to detect other items throughout the city, like potholes, graffiti. spots and road conditions. Then you have a Swiss startup by the name of Cortexia. It uses an AI algorithm to measure a city's cleanliness based off its Clean City Index. The algorithm can detect and count waste as as small as a cigarette stub. But despite the rising applications with computer vision technology, cities still hesitate on how to use it. We still have the privacy discussions. Is it far enough to really detect the waste? We just don't want to do the pilot in such a way where we know it's going to fail. Amsterdam's neighboring city of Utrecht also tested the algorithm, training it to detect waste left behind in the street by its sweeper cars. But the city says that's as far as it got, not only because of privacy concerns, but also because of the expense, since the system requires the city to analyze and store mountains of data. And then there's a question on the practicality of this. All right, you have your algorithm. Now, how does that fit in with your day-to-day business? Because image recognition is not a day-to-day business. Our day-to-day business is cleaning up the streets to clean up the city. But how can we use those technologies to actually involve them in your day -to -day business, like the planning, the cleaning, and the way we deal with it? citizens. Now, the Netherlands is famous for being an early adopter of smart waste technology and the country as a whole stands out for its low recycling rates compared to the rest of the EU. So when cities in this country test new solutions, we take notice. Stick with current insights for more updates on how the object detection kit evolves. And if you have a topic that you would like for us to dig into, hit the button below.