Vision-Based Waste Enclosure Monitoring for Fill and Sorting Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing waste collection monitoring systems are ineffective for open collection areas, fail to detect waste type, and are costly when applied to leased bins, leading to inefficient transportation and sorting errors.
Innovation Solution
A system using an image acquisition camera and machine learning models to detect collection enclosures, waste type, and compute filling levels without additional sensors, allowing for automated monitoring and error detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If ultrasonic sensors are installed in all containers to monitor filling levels, then monitoring accuracy is improved, but installation cost and complexity increase significantly
Solution Approach 1:
The patent replaces mechanical ultrasonic sensors with an optical vision system using cameras and image processing algorithms. This substitution eliminates the need for physical sensor installation in each container, reducing complexity while maintaining monitoring capability through external visual observation and computational analysis
Solution Approach 2:
Instead of installing physical sensors in each container, the system creates visual copies (images) of the containers and their contents, then processes these copies computationally to determine filling levels. This approach avoids physical installation complexity while achieving the same monitoring function
2Measurement precision
If traditional monitoring systems are used, then filling level can be detected, but waste type classification and sorting error detection are not possible
Solution Approach 1:
The vision-based system performs multiple functions simultaneously: it detects filling levels, classifies waste types, and identifies sorting errors all through the same image processing pipeline. This multi-functional approach eliminates the need for separate detection systems for each parameter
Solution Approach 2:
The system analyzes visual characteristics including color information from images to classify different waste types. By processing color and visual data, the system can distinguish between different material types and detect contamination without requiring additional specialized sensors
3Productivity
If bins are transported when under-loaded, then collection frequency increases, but transportation costs and carbon footprint increase unnecessarily
Solution Approach 1:
The system continuously monitors filling levels and provides real-time feedback about container status. This feedback enables dynamic scheduling of collections based on actual need rather than fixed schedules, optimizing transportation routes and timing to avoid unnecessary trips while ensuring timely collection when containers are actually full
4Device complexity
If manual monitoring of waste collection areas is used, then system complexity is reduced, but labor costs increase and sorting errors are not detected
Solution Approach 1:
The system enables self-monitoring of waste collection areas through automated image capture and processing. The computational system automatically analyzes images to detect sorting errors and classify waste types without requiring human intervention, providing intelligent monitoring that surpasses manual capabilities while remaining relatively simple to deploy
Data Source
AI summary
The invention relates to a system for monitoring a loose waste collection enclosure (30) comprising: an image acquisition camera; an image processing unit (100); characterized in that said processing unit comprises at least one module for detecting each collection enclosure present in said acquired images based on a first trained machine learning model; a module for determining the type of waste present in each image portion detected as being a collection enclosure based on a second trained machine learning model; a module for computing a filling rate of each collection enclosure by analyzing said enclosure images detected by said enclosure detection module.