Trusted forensics scheme based on digital watermark algorithm in intelligent VANET

Abstract

Trusted forensics is one of the most important problems in VANET, and it often needs continuous video monitoring, once break out emergent vehicle accidents, then specific staff members take steps for forensics to obtain facts and define responsibility. Traditional forensics exists problems of inaccurate information, unfair responsibility definition and risk of leakage of user’s privacy. To solve the above problem, in this paper, we proposed a trusted forensics scheme based on digital image watermark in intelligent VANET, in which we proposed technical and fair algorithms for trusted forensics, and the trusted forensics scheme includes basic forensics parameter data obtaining critical forensics data automatic generation and forensics data extraction. Once there vehicle accident occurred, the forensics system first obtains the location, timestamp, forensics device data as basic forensics parameter data, and then, it embeds the forensics parameter as watermark into the real-time vehicle accident photograph by the proposed digital watermark algorithm, and thus, the real-time and undeniable forensics data are automatic generated as evidence; when necessary, the forensics system can extract the evidence data and watermark data from the critical forensics data. The proposed scheme can detect the content integrity of image data and even find out tampering mark when the image data are tampered. Additionally, we used neural network algorithm for vehicle license plate recognition and rapid vehicle information gathering. Finally, experiments evaluations manifest the proposed is forensics scheme is secure, robust, and efficient in vehicle forensics.