Railway foreign matter invasion small target detection method and system based on improved YOLO-v5s model

The invention discloses a railway foreign matter invasion small target detection method and system based on an improved YOLO-v5s model. The method comprises the following steps: acquiring and constructing a data set of railway foreign matters with labels; a CIOU loss function in the YOLOv5s model is modified into a Focal EIOU loss function; the YOLOv5s model is improved, and an improved non-parametric attention mechanism M-SimAM module is added behind the SPPF module; replacing a detection head part of the YOLOv5s model with a dynamic detection head DyHead; and detecting the attempt image to be detected, and evaluating the detection result. According to the method, the detection precision of the model can be improved, the feature extraction capability is enhanced, and the safety and monitoring capability of a railway system are improved. 本发明公开了一种基于改进YOLO-v5s模型的铁路异物入侵小目标检测方法及系统，方法包括：获取并构建带有标注的铁路异物的数据集；将YOLOv5s模型中的CIOU损失函数修改为Focal EIOU损失函数；改进所述YOLOv5s模型，在SPPF模块后添加改进后的非参数注意力机制M-SimAM模块；将YOLOv5s模型的检测头部替换为动态检测头DyH