Target detection method and system for automobile aided driving

The invention discloses a target detection method and system for automobile aided driving, and relates to the field of automobile aided driving, a target detection model is obtained through a yolov5 model, a target object on a traffic road image captured in the automobile driving process is detected through the target detection model, and due to the fact that the yolov5 model belongs to a one-stage algorithm, the target object can be detected through the yolov5 model. Compared with R-CNN and Fast-RCNN, the method has the advantages that a candidate frame screening process is omitted, the position coordinates of the target frame and the classification probability of the target are directly regressed, and compared with R-CNN and Fast-RCNN, a two-stage algorithm of firstly screening the candidate frame, then judging whether the to-be-detected target is in the candidate frame and correcting the position of the target is needed, so that higher recognition efficiency is achieved; the method is suitable for the driv