DEVELOPMENT OF TECHNOLOGY FOR DETECTING TRAFFIC HAZARDS FOR BUSINESS VEHICLES USING SMARTPHONES AND CLOUD-BASED AI

Abstract

While the number of traffic accidents and the number of casualties in Japan are on a downward trend, the number of accidents in which business vehicles are the first party is on a slight upward trend. Against this backdrop, further traffic safety measures are also required for commercial vehicles. In recent years, research in the field of traffic safety on various IoT sensors and traffic big data has progressed, but there is no technology to detect traffic accident dangerous events by content without statisticizing GNSS information and moving images that individual vehicles always acquire, and technologies for detecting dangerous events without using devices such as advanced sensors. We developed a dangerous event detection system that processes GNSS information and forward-moving images obtained from general smartphones on the cloud using image analysis and AI technology, and introduced it as a demonstration experiment to the Ministry of the Environment's transportation work project for removing soil, etc.