Highway vehicle analytics

Abstract

Highway Vehicle Analytics plays an important role in intelligent transportation system, especially for traffic management. This project can also be used in areas where high security is needed or just for the purpose of gathering valuable data. Our project proposes the technique for detection and counting of vehicles, which is video-based. In Pakistan, the security situation has been critical for the past few years. Notwithstanding the circumstances, majority institutions of our country employ manual security y personnel. This apart system from being ultra-expensive is also vulnerable. This project aims to provide a real- time vehicle security and analytics mechanism that will detect all the details of the vehicle that are entering through the gate and then comparing the data with the already y authorized database. Lf the details match, the vehicle will be let through. If not, the vehicle will not be allowed to pass through the security system. Current services do not work in Pakistan, because all those services available globally are trained on general dataset which has no resemblances with our vehicles. Pakistan has a quite a unique set of vehicles on its roads as compared to rest of the globe. We have used manual data gathering technique of recording the details of all vehicles entering or leaving manually from various CCTV sources. After gathering the data, we then applied various techniques to filter the most efficient form of information that would be intern used to train custom models for max efficiency y of the system.