EdgeVision: User Defined Object-Counting using Raspberry Pi
Image 1: Single Object Detection and Tracking
Description: The image highlights a single individual detected by the system, with bounding boxes indicating precise detection and tracking. The individual is interacting with defined lines for object counting.
Image 2: Multiple Object Detection and Tracking
Description: This image demonstrates the detection of multiple individuals using bounding boxes. The system tracks interactions with multiple lines for real-time counting as the objects approach the lines.
Image 3: Crossing Lines and Incremental Counting
Description: Captures multiple detected objects crossing predefined lines. The system accurately tracks and increments counts, demonstrating the effectiveness of the object counting mechanism.
EdgeVision is a real-time object detection and counting system designed for resource-constrained embedded platforms, leveraging the Raspberry Pi 4 paired with its high-resolution Camera Module V2. The system integrates the Faster R-CNN ResNet-50 deep learning model, enabling accurate object detection and tracking directly at the edge. EdgeVision is tailored for practical applications such as public safety monitoring, traffic analysis, and retail analytics, where real-time insights are critical.
The architecture incorporates a CentroidTracker module, which maintains consistent object identification across video frames using a centroid-based nearest neighbor tracking algorithm. This ensures reliable performance even under challenging scenarios involving temporary occlusions or overlapping objects.
One of EdgeVision’s core innovations is its interactive, user-friendly interface. The system allows users to draw virtual line segments onto live video feeds using mouse interactions. These line segments act as dynamic counters, automatically tracking and counting objects as they cross the defined boundaries. This approach empowers users to configure counting zones flexibly without modifying the underlying code.
Through efficient model optimization and real-time video streaming, EdgeVision demonstrates how deep learning models can be effectively deployed on low-power embedded systems, bridging the gap between high-performance object detection and the practical constraints of edge computing environments.