Vehicle detection and counting are critical tasks in intelligent transportation systems, essential for traffic management, urban planning, and autonomous driving. This paper proposes an enhanced method leveraging YOLOv8 for accurate and efficient vehicle detection and counting. Our approach includes advanced preprocessing techniques, data augmentation, and a novel object grouping mechanism before category identification, aimed at reducing resource consumption and improving processing speed. Extensive experiments performed using the COCO dataset show that our proposed method achieves competitive performance with significant improvements in resource efficiency and inference time. However, this approach may lead to slight reductions in detection accuracy. The results indicate the potential of the proposed method for real-time applications, balancing accuracy, and efficiency effectively. This study advances the development of traffic management systems that are more effective and efficient, with potential applications in smart cities. Future work will focus on further optimizing the object grouping mechanism and exploring the integration of additional data sources to enhance detection accuracy.