Efficient Vehicle Detection and Tracking using Blob Detection and Kernelized Filter

Smart conveyance techniques are crucial in modern society, enabling efficient traffic management and ensuring public safety. Therefore, this paper proposes a novel approach combining Blob detection and the kernelled correlation filter (KCF)-based visual tracking method for efficient vehicle detection, tracking, and trajectory generation. Numerous preprocessing techniques, including gamma correction, contrast reduction, bilateral filtering, and Mean Shift filtering are applied to improve the results of vehicle detection. Subsequently, the KCF algorithm is employed for robust vehicle tracking and trajectory generation. The trajectories provide valuable information about the flow of vehicles, contributing to better traffic monitoring systems. The suggested approach is validated on the publicly accessible KITTI benchmark dataset, and the results demonstrate that the integration of blob detection and KCF-based tracking, along with the preprocessing techniques, significantly improves detection accuracy. Our approach achieved an accuracy of S2% for vehicle detection and 86% for tracking vehicles. There is a lot of potential for this research for improving public safety, traffic flow management, and traffic monitoring systems. Our technique yields useful information regarding vehicle trajectories, which enables a thorough study of vehicle flow patterns.