A study of airport and runway detection based on improved RT-DETR

Aiming at the detection problems of airports and runways in remote sensing images, such as large scale differences and insignificant target features, this paper proposes a target detection method based on improved RT-DETR. First, based on the original architecture of RT-DETR, the AKConv module is introduced to replace the original Depthwise Convolution to enhance the feature extraction capability and context modeling capability. Secondly, the CBAM (Channel and Spatial Attention Mechanism) is added in the encoder stage to further enhance the model’s attention to key area features. The experiments are conducted on the remote sensing image dataset containing two categories, “airport” and “runway”. The results show that the improved model outperforms the original RT-DETR model in terms of detection accuracy, recall rate, and inference speed, which verifies the effectiveness and practicality of the proposed method in the remote sensing target detection task.