Object Detection Algorithm Based on YOLOv5s

Object detection is a kind of image segmentation based on geometric and statistical characteristics of the object. It is not only necessary to identify which classification the object belongs to, but more importantly, to get the specific position of the object in the image. Object detection based on deep learning is the current research hotspot and the future development direction. The YOLOv5s network is the network with the smallest depth and the smallest width of the feature map in the YOLOv5 series, achieving a better balance between speed and accuracy. Based on the analysis of relevant literatures at home and abroad, the structure of YOLOv5s network is studied in this paper. In order to solve the problem that the detection effect is not accurate enough when dealing with complex scenes and high-precision tasks, we improve it by integrating attention mechanism and replacement loss function. The fusion attention mechanism can amplify the important information features and suppress the non-important information features, and the replacement loss function can accelerate the convergence speed and improve the target detection effect. The effectiveness of these two improvements is verified by simulation experiments.