Passenger Vehicle Driver Detection Based on YOLOv5+Stacking

The intelligent detection of driver distracted driving and mask wearing can strengthen the safety management of passenger vehicle operators, the correct detection of drivers in monitoring images by regulatory authorities is the basis for the implementation of these two tasks.Although the traditional object detection model can detect the driver and the passenger accurately, there are still many wrong classification between the driver and the passenger.In order to increase the detection rate, this paper uses CBAM attention mechanism and the idea of non-maximum suppression in YOLO v5, and proposes a combination model of YOLOv5+Stacking integrated learning, which can effectively reduce false detection between drivers and passengers and ultimately increase the detection rate of YOLO v5 model.In this paper, this model is used to verify and detect the collected data set, and its evaluation indexes are not only better than the original YOLO v5 model but also better than other similar detection models.