Aircraft Target Detection Based on Attention Mechanism and Faster R-CNN

Aircraft, as a special strategic target, has high value in both civil and military use, and it is especially important to achieve intelligent scientific aircraft target detection compared to traditional methods. Therefore, an aircraft detection method based on attention mechanism and Faster R-CNN target detection algorithm is proposed in this paper. The RSOD dataset is selected, and the VOC format dataset is used for pre-processing and data enhancement. The Faster R-CNN model is constructed, which uses ResNet50 as the backbone network for feature extraction and also adds the attention mechanism in the backbone network; the region proposal is obtained after convolving the feature extraction results using the RPN(Region Proposal Network); the prediction results are finally obtained by pooling using ROI Pooling. The experimental results show that the model can have good recognition effect and provide help for the intelligent recognition of aircraft.