Computer Vision and Machine Learning Based Traffic Scene Perception and Car Driver Injury Assessment

Traffic scene perception and risk assessment for car drivers is critical to improving road safety. This study aimed to develop a deep supervised learning detector for automotive traffic scenes based on computer vision and machine learning techniques to detect the effects of weather factors and road factors on car drivers and to assess the degree of injury in traffic accidents. The goal of this research was to assess the extent to which motorists are injured in traffic accidents in each traffic scenario by analyzing and evaluating weather factors and road factors. This approach used deeply supervised learning detectors to detect and classify weather and road factors in motor vehicle traffic scenarios using computer vision and machine learning techniques. By processing and analyzing the data, it can determine the relationship between weather and road factors and the extent of injuries sustained by motorists. In addition, the research analyzed the relationship between weather and road factors and the degree of driver injury using a linear regression model. By calculating the regression coefficients, the research can quantify the effect of weather and road factors on the level of driver injury. The experimental results showed that weather and road factors were positively related to the level of injury of motorists. This study provided important insights into car driver risk assessment in traffic scenarios through the evaluation and discussion of deep supervised learning detectors and linear regression models.