Aiming at the problems of low accuracy and poor real-time performance in remote sensing image aircraft de-tection tasks,an algorithm to improve YOLOv4 remote sensing image detection is proposed in this paper.The algorithm is based on YOLOv4,and uses a lightweight network MobileNetv3 instead of the original feature extraction network of YOLOv4 to reduce the number of parameters and improve the detection speed while ensuring its feature extraction capability.Mean-while,it uses depth-separable convolution instead of traditional convolution in the road aggregation network(PANet),and introduces the BAM attention mechanism in the backbone network to improve the overall.The BAM attention mechanism is introduced in the backbone network to improve the overall generalization ability of the model.Then the NMS network is op-timized to improve the final recognition accuracy of the model.
万方数据
维普
