Toward the Secure Transmission of Large-Scale Images in Autonomous Driving

The rapid advancement of autonomous driving technology and the increasing complexity of in-vehicle systems present novel challenges to system safety. The substantial volume of data collected by in-vehicle sensors greatly supports the intelligent enhancement of autonomous driving systems. However, traditional public key cryptographic systems, which rely on computational complexity, may fail to meet the requirements for secure transmission and real-time response of the vast image information. Resource-intensive AI algorithms often necessitate computing and storage resources offered by cloud platforms for training and predicting. Securely integrating AI into autonomous driving systems has emerged as an imminent problem that demands attention during its development process. To address the aforementioned issues, this paper focuses on analyzing secure image transmission technologies and trusted AI algorithms for autonomous driving. To tackle the challenge of traditional public key-based data protection methods for the real-time secure transmission of massive data in autonomous driving systems, this paper introduces a secure and efficient image partitioning scheme based on visual cryptography (VC), which resolves the conflict between data protection in autonomous driving and real-time responsiveness.