Passenger Detection for Preventing Overheating Accidents in Vehicles Based on Deep Learning

The monitoring of life left in the car, as a technical means of intelligent cockpit safety assurance, has been widely applied in the automotive industry in recent years. However, incidents of injury or even death caused by vulnerable groups such as children or pets being left behind in closed cars continue to occur, making in car biological monitoring gradually a key issue in car passenger monitoring systems. In response to the unique and poor visual recognition performance of vision based biological legacy monitoring scenes in vehicles, this paper uses Efficientnet as the backbone of SSD object detection framework and constructs an E-SSD network combined with ResFPN multi-scale feature fusion module. In addition, this article will apply transfer learning to the monitoring system. The effectiveness of our method was verified by comparing it with different networks on the test dataset.