<p>In modern traffic management systems, efficient and accurate vehicle detection from video feeds plays a critical role. This paper introduces a novel approach utilizing Two-Tier Parallel Virtual Lattice Layers (TPL2) to significantly enhance the efficiency of vehicle detection in video-based traffic management applications. The TPL framework employs grid-like structures, processed in parallel to detect vehicles within traffic video streams. Multiple layers are dynamically configured based on the traffic scene’s perspective, employing parallel computation to improve both computational efficiency and detection accuracy. Experimental evaluations conducted on traffic video datasets demonstrate notable enhancements in efficiency and accuracy compared to traditional methods. TPL2 exhibits improved efficiency by approximately 26.55% which signifies a notable reduction in processing time compared to a full frame and sequential processing, indicating enhanced efficiency or optimization in the process. Also, it is important to acknowledge that this efficiency gain is accompanied by a moderate increase in memory usage, approximately 3.72% due to parallel processing technique. This novel framework ensures that it is highly robust, scalable and compatible to optimize diverse video traffic management systems.</p>