Abstract
Object counting is an active research area that gained more attention in the past few years. In smart cities, vehicle counting plays a crucial role in urban planning and management of the Intelligent Transportation Systems (ITS). Several approaches have been proposed in the literature to address this problem. However, the resulting detection accuracy is still not adequate. This paper proposes an efficient approach that uses deep learning concepts and correlation filters for multi-object counting and tracking. The performance of the proposed system is evaluated using a dataset consisting of 16 videos with different features to examine the impact of object density, image quality, angle of view, and speed of motion towards system accuracy. Performance evaluation exhibits promising results in normal traffic scenarios and adverse weather conditions. Moreover, the proposed approach outperforms the performance of two recent approaches from the literature.
Highlights
Drones are widely used in many emerging fields such as military, disaster monitoring and recovery, outer space, transportation, wildlife and historical conservation, medicine, agriculture and photography [1,2,3,4]
We extend our previous work and propose a more efficient system for object tracking and counting which integrates the latest version of You Only Look Once (YOLO) (Yolo5) for object detection and the Channel and Spatial Reliability Tracker (CSRT) for object tracking and counting
Object tracking is done based on two methods, the first method is single tracking where the goal of the work is to estimate the condition of a target, which is indicated in the first frame, in real-time across frames
Summary
Drones are widely used in many emerging fields such as military, disaster monitoring and recovery, outer space, transportation, wildlife and historical conservation, medicine, agriculture and photography [1,2,3,4]. In a variety of fields, such as traffic data gathering, traffic monitoring, film and television shooting, visual object tracking in an unmanned aerial vehicle (UAV) recording plays a significant role. Various issues such as appearance variation, background clutter, and extreme occlusion make it difficult to track the target reliably in a UAV vision task. UAV-based low-altitude aerial photography technology has been widely deployed as a viable supplement to aircraft remote sensing and satellite remote sensing for traffic data collection. This device uses aerial high-resolution cameras to clearly capture ground objects, with imaging resolution reaching centimeters. Regression-based algorithms are best suited for real-time applications where we can compromise the accuracy a little bit to gain faster processing speed
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
