Military Aircraft Detection Algorithm Based on Improved YOLOv5
To address the problems of low accuracy and high rate of missed detection and false alarm of military aircraft target detection in remote sensing images,YOLO-Military Aircraft Recognition(YOLO-MAR),a lightweight remote sensing image military aircraft target detection algorithm based on YOLOv5s is proposed.Firstly,a new network structure is proposed,multi-scale sensing field weights adjustment is completed,and the feature extraction network and feature fusion network are redesigned to increase the weight of small target features and perform lightweight processing.Then,FPGM is used to prune the reconstructed model,which greatly reduces the number of parameters and volume of the model.Finally,SIoU Loss is used as the loss function of the model to accelerate the convergence speed of the model and improve the accuracy of detection.The results show that on the open military aircraft dataset MAR20,the model volume of YOLO-MAR is as low as 3.95 MB,which is reduced by 71.5％compared with the original YOLOv5s,and the minimum model volume after pruning can be reduced to 0.2 MB,and the average detection accuracy of the model can reach up to 91.7％,which is increased by 2.34％.And it is advanced in terms of detection effect,model volume,parameter quantity,and calculation amount,which is capable of high-quality real-time detection of military aircraft targets.
target detectionmilitary aircraftYOLOv5sFPGMSIoU Loss
万方数据
维普
