The transformation of transport systems increasingly relies on connecting vehicles into cooperative networks, promising to reduce traffic congestion, lower emissions, and enhance road safety. Achieving these benefits requires efficient sharing of high definition (HD) map information across vehicles. Federated learning (FL) enables vehicles to collaboratively train shared models for HD maps, supporting accurate localization and dynamic environment understanding, without transferring raw data, thus preserving privacy. However, implementing FL in this context faces significant challenges: malicious participants can disrupt learning, and honest participants may disengage without sufficient incentives. To address these issues, we propose integrating a dual blockchain framework with FL, which evaluates participant reliability, weights contributions accordingly, and provides financial incentives to maintain engagement. This coordinated approach ensures that only trustworthy information contributes to HD map updates, preserving both data privacy and system integrity. By modeling and optimizing the system’s economic utility, and validating through simulations, we demonstrate that the framework is feasible under realistic settings when reward and subscription fees are appropriately calibrated, providing a robust foundation for secure, incentivized, and privacy preserving vehicular networks.