Unmanned aerial vehicles (UAVs) and infrared imaging technology have numerous applications in civilian fields. To address the issues of low accuracy resulting from complex ground backgrounds, small target size, and limited target features in UAV remote sensing infrared image target detection, we utilize the YOLOv9s model and the latest Retentive Networks Meet Vision Transformers technology, and propose the RMT-YOLOv9s model for infrared small target detection. Firstly, a CNN-RMT-based backbone is proposed by incorporating the RMT model into the backbone network of YOLOv9s, which extracts both local and global features for small target detection. Then, an improved neck multi-scale feature-fusion network RMTELAN-PANet is designed using the novel convolutional RMTELAN module proposed in this letter, which can better capture and utilize semantic information from feature maps. Finally, EMA attention module and upsampling Dysample module are integrated into RMTELAN-PANet to further improve the feature information of small targets. Experiments on the HIT-UAV dataset show that RMT-YOLOv9s outperforms other popular methods in infrared small target detection.

Low-altitude unmanned aerial vehicles (UAVs) and unmanned ground vehicles (UGVs), which boast high-resolution imaging and agile maneuvering capabilities, are widely utilized in military scenarios and generate a vast amount of image data that can be leveraged for textual intelligence generation to support military decision making. Military image captioning (MilitIC), as a visual-language learning task, provides innovative solutions for military image understanding and intelligence generation. However, the scarcity of military image datasets hinders the advancement of MilitIC methods, especially those based on deep learning. To overcome this limitation, we introduce an open-access benchmark dataset, which was termed the Military Objects in Real Combat (MOCO) dataset. It features real combat images captured from the perspective of low-altitude UAVs or UGVs, along with a comprehensive set of captions. Furthermore, we propose a novel encoder–augmentation–decoder image-captioning architecture with a map augmentation embedding (MAE) mechanism, MAE-MilitIC, which leverages both image and text modalities as a guiding prefix for caption generation and bridges the semantic gap between visual and textual data. The MAE mechanism maps both image and text embeddings onto a semantic subspace constructed by relevant military prompts, and augments the military semantics of the image embeddings with attribute-explicit text embeddings. Finally, we demonstrate through extensive experiments that MAE-MilitIC surpasses existing models in performance on two challenging datasets, which provides strong support for intelligence warfare based on military UAVs and UGVs.