A System for Real-Time Detection of Abandoned Luggage

Abstract

Highlights: What are the main findings? The fine-tuned model based on YOLOv11-l architecture achieves better results in detecting people and luggage in surveillance camera footage than the YOLOv8 and DETR models fine-tuned on the same dataset. The fine-tuned YOLOv8 and YOLOv11 models in m and l versions significantly improve object detection accuracy in demanding surveillance scenes with many small and medium-sized objects, with mAP@0.5 from 3.34% to over 86%. The fine-tuned YOLOv11-l model shows excellent performance in object detection, with mAP@0.5 accuracy of 96% for medium-sized objects and 85% for small objects. An algorithm for detecting abandoned luggage in public areas in real-world scenes was designed, implemented in Python 3.10, and tested on different scenarios in airport scenes. Image datasets were created, with images collected from surveillance cameras in public areas of airports and walkways and prepared for machine learning of object detectors. What is the implication of the main finding? The accurate detection of people and luggage significantly contributes to increasing the functionality of the abandoned luggage detection algorithm and creating a system that helps in monitoring public spaces and increasing safety in crowded public areas. Including adjustable parameters in the algorithm (such as luggage dwell time, owner’s distance from luggage) reduces false alarms and improves the system efficiency. In this paper, we propose a system for the real-time automatic detection of abandoned luggage in an airport recorded by surveillance cameras. To do this, we use an adapted YOLOv11-s model and a proposed algorithm for detecting unattended luggage. The system uses the OpenCV library for the video processing of the recorded footage, a detector, and an algorithm that analyzes the movement of a person and their luggage and evaluates their spatial and temporal relationships to determine whether the luggage is truly abandoned. We used several popular deep convolutional neural network architectures for object detection, e.g., Yolov8, Yolov11, and DETR encoder–decoder transformer with a ResNet-50 deep convolutional backbone, we fine-tuned them on our dataset, and compared their performance in detecting people and luggage in surveillance scenes recorded by an airport surveillance camera. The fine-tuned model significantly improved the detection of people and luggage captured by the airport surveillance camera in our custom dataset. The fine-tuned YOLOv8 and YOLOv11 models achieved excellent real-time results on a challenging dataset consisting only of small and medium-sized objects. They achieved real-time precision (mAP) of over 88%, while their precision for medium-sized objects was over 96%. However, the YOLOv11-s model achieved the highest precision in detecting small objects, corresponding to 85.8%, which is why we selected it as a component of the abandoned luggage detection system. The abandoned luggage detection algorithm was tested in various scenarios where luggage may be left behind and in situations that may be potentially suspicious and showed promising results.