Vehicle recognition plays a vital role in intelligent transportation systems, law enforcement, access control, and security operations—domains that are becoming increasingly dynamic and complex. Despite advancements, most existing solutions remain siloed, addressing individual tasks such as vehicle make and model recognition (VMMR), automatic number plate recognition (ANPR), and color classification separately. This fragmented approach limits real-world efficiency, leading to slower processing, reduced accuracy, and increased operational costs, particularly in traffic monitoring and surveillance scenarios. To address these limitations, we present a unified framework that consolidates all three recognition tasks into a single, lightweight system. The framework utilizes MobileNetV2 for efficient VMMR, YOLO (You Only Look Once) for accurate license plate detection, and histogram-based clustering in the HSV color space for precise color identification. Rather than optimizing each module in isolation, our approach emphasizes tight integration, enabling improved performance and reliability. The system also features adaptive image calibration and robust algorithmic enhancements to ensure consistent results under varying environmental conditions. Experimental evaluations demonstrate that the proposed model achieves a combined accuracy of 93.3%, outperforming traditional methods and offering practical scalability for deployment in real-world transportation infrastructures.

Multiple Object Tracking (MOT) is essential for applications such as autonomous driving, surveillance, and analytics; However, challenges such as occlusion, low-resolution imaging, and identity switches remain persistent. We propose HAMOT, a hierarchical adaptive multi-object tracker that solves these challenges with a novel, unified framework. Unlike previous methods that rely on isolated components, HAMOT incorporates a Swin Transformer-based Adaptive Enhancement (STAE) module—comprising Scene-Adaptive Transformer Enhancement and Confidence-Adaptive Feature Refinement—to improve detection under low-visibility conditions. The hierarchical Dynamic Graph Neural Network with Temporal Attention (DGNN-TA) models both short- and long-term associations, and the Adaptive Unscented Kalman Filter with Gated Recurrent Unit (AUKF-GRU) ensures accurate motion prediction. The novel Graph-Based Density-Aware Clustering (GDAC) improves occlusion recovery by adapting to scene density, preserving identity integrity. This integrated approach enables adaptive responses to complex visual scenarios, Achieving exceptional performance across all evaluation metrics, including a Higher Order Tracking Accuracy (HOTA) of 67.05%, a Multiple Object Tracking Accuracy (MOTA) of 82.4%, an ID F1 Score (IDF1) of 83.1%, and a total of 1052 Identity Switches (IDSW) on the MOT17; 66.61% HOTA, 78.3% MOTA, 82.1% IDF1, and a total of 748 IDSW on MOT20; and 66.4% HOTA, 92.32% MOTA, and 68.96% IDF1 on DanceTrack. With fixed thresholds, the full HAMOT model (all six components) achieves real-time functionality at 24 FPS on MOT17 using RTX3090, ensuring robustness and scalability for real-world MOT applications.