Computer vision–based monitoring system
A regional waste management company responsible for municipal collection was facing growing challenges in accurately tracking bin pickups and route efficiency. Their existing system relied on RFID scanners mounted on collection trucks to log each waste bin as it was serviced — data that was used for planning, billing, and reporting.
However, gaps in the data started to raise concerns about both operational efficiency and accountability in the field.
The RFID system frequently missed pickups due to hardware inconsistencies, misaligned tags, or poor weather conditions. As a result:
Some bins were collected without being logged, leading to discrepancies between expected and actual waste volume.
Route plans became unreliable, with trucks carrying more waste than scheduled and teams forced to make unplanned stops.
The company lacked visibility into specific areas where these inconsistencies occurred, making it difficult to optimize or investigate.
There were also concerns that certain high-volume locations were contributing disproportionally to the issue, and that oversight was needed to ensure adherence to collection policies.
To address these challenges, we developed a computer vision–based monitoring system that worked alongside the existing RFID setup — providing an independent source of truth for bin pickup events.
Object Detection: Cameras mounted on collection trucks detected and tracked each bin as it was loaded, verifying the actual number of pickups per route.
Depth Estimation: The system estimated bin size and fill levels using onboard sensors, allowing comparisons between expected and actual waste volumes.
Route-Level Insights: By geotagging each event, the system created a detailed map of bin activity, highlighting areas where pickups occurred without RFID confirmation.
The AI system cross-referenced RFID records with visual evidence to detect discrepancies.
Data patterns revealed consistent hotspots where bin overloading and undocumented pickups were most frequent — giving the company the insights needed to investigate further and improve compliance without making assumptions.
The new system gave the firm much-needed transparency into what was actually happening during collection routes:
Data accuracy improved, with AI catching missed events the RFID system overlooked.
Operational efficiency increased, allowing better planning and right-sizing of collection routes.
Visibility into high-discrepancy areas empowered leadership to make informed adjustments — whether through retraining, equipment checks, or internal review processes.
This project demonstrates how AI can complement existing infrastructure to provide deeper insight, improve accuracy, and support fair, data-driven decision-making. In industries like waste management — where logistics, accountability, and trust all intersect — AI-powered monitoring helps teams stay aligned, efficient, and transparent.
