Abstract: Aiming at the problems of small defect area,diverse shapes,blurred boundaries and complex,backgrounds of surface defects of hot-rolled strip steel,a surface defect detection model for hot rolled strip,named SFSP-YOLOv7,is proposed.First,by improving the k-means++clustering algorithm,the prior frame dimension is adjusted,the Euclidean distance metric is replaced by intersection over union (IoU),and a more representative anchor dimension is obtained by introducing genetic algorithm (GA).Enhancing the model's the regression speed and the accuracy of small area defect detection.Secondly,for the defects with fuzzy boundary and complex backgrounds,a new object detection boundary frame loss function FocalSIoU is proposed to reduce the learning of unnecessary features in the model,accelerate the detection speed,and improve the regression effect of the prediction frame.Finally,a multi-scale feature fusion attention module (MFFM) is designed to enhance the model's feature extraction capabilities through multi-scale information fusion,improve the detection accuracy of small targets,and improve the misdetection rate of model detection.The space to depth (SPD) convolution module is introduced into the model Head structure to improve the model to avoid the loss of fine-grained information and reduce the target miss detection rate.Through the verification of NEU-DET data set,the mean average detection accuracy (mAP) value of the SFSP-YOLOv7 model is 78.25%,which is 4.92% higher than the original YOLOv7 model,indicating the effectiveness of the detection method proposed.