This paper presents a distributed implementation of a geometric-based multi-camera vehicle tracking system designed to address the challenges of inter-camera identity matching. The proposed method employs a novel geometric technique that maps the 3D perspective view of each camera into a unified 2D flat representation. This transformation allows for the precise spatial alignment of overlapping fields of view, enabling accurate detection of a vehicle’s exact position and maintaining its track ID across multiple cameras with consistency. By leveraging this geometric approach that integrates flattening and stitching of camera views, the system eliminates the need for computationally expensive feature extraction and re-identification tasks, making it highly efficient and suitable for deployment on low-power edge devices. To enhance scalability and efficiency, the framework adopts a distributed architecture where each camera operates as an independent node. These nodes locally perform vehicle detection, tracking, and transformation, transmitting lightweight 2D data to a central server for global map alignment and inter-camera association. Experimental evaluations demonstrate the system’s high matching accuracy, validating its effectiveness on CPU-based hardware. Future work will focus on deploying the framework on Raspberry Pi devices to assess its real-time performance and scalability in resource-constrained environments.