The detection of infrared vehicle targets by UAVs poses significant challenges in the presence of complex ground backgrounds, high target density, and a large proportion of small targets, which result in high false alarm rates. To alleviate these deficiencies, a novel YOLOv7-based, multi-scale target detection method for infrared vehicle targets is proposed, which is termed YOLO-ViT. Firstly, within the YOLOV7-based framework, the lightweight MobileViT network is incorporated as the feature extraction backbone network to fully extract the local and global features of the object and reduce the complexity of the model. Secondly, an innovative C3-PANet neural network structure is delicately designed, which adopts the CARAFE upsampling method to utilize the semantic information in the feature map and improve the model’s recognition accuracy of the target region. In conjunction with the C3 structure, the receptive field will be increased to enhance the network’s accuracy in recognizing small targets and model generalization ability. Finally, the K-means++ clustering method is utilized to optimize the anchor box size, leading to the design of anchor boxes better suited for detecting small infrared targets from UAVs, thereby improving detection efficiency. The present article showcases experimental findings attained through the use of the HIT-UAV public dataset. The results demonstrate that the enhanced YOLO-ViT approach, in comparison to the original method, achieves a reduction in the number of parameters by 49.9% and floating-point operations by 67.9%. Furthermore, the mean average precision (mAP) exhibits an improvement of 0.9% over the existing algorithm, reaching a value of 94.5%, which validates the effectiveness of the method for UAV infrared vehicle target detection.

A UAV infrared target detection model ITD-YOLOv8 based on YOLOv8 is proposed to address the issues of model missed and false detections caused by complex ground background and uneven target scale in UAV aerial infrared image target detection, as well as high computational complexity. Firstly, an improved YOLOv8 backbone feature extraction network is designed based on the lightweight network GhostHGNetV2. It can effectively capture target feature information at different scales, improving target detection accuracy in complex environments while remaining lightweight. Secondly, the VoVGSCSP improves model perceptual abilities by referencing global contextual information and multiscale features to enhance neck structure. At the same time, a lightweight convolutional operation called AXConv is introduced to replace the regular convolutional module. Replacing traditional fixed-size convolution kernels with convolution kernels of different sizes effectively reduces the complexity of the model. Then, to further optimize the model and reduce missed and false detections during object detection, the CoordAtt attention mechanism is introduced in the neck of the model to weight the channel dimensions of the feature map, allowing the network to pay more attention to the important feature information, thereby improving the accuracy and robustness of object detection. Finally, the implementation of XIoU as a loss function for boundary boxes enhances the precision of target localization. The experimental findings demonstrate that ITD-YOLOv8, in comparison to YOLOv8n, effectively reduces the rate of missed and false detections for detecting multi-scale small targets in complex backgrounds. Additionally, it achieves a 41.9% reduction in model parameters and a 25.9% decrease in floating-point operations. Moreover, the mean accuracy (mAP) attains an impressive 93.5%, thereby confirming the model’s applicability for infrared target detection on unmanned aerial vehicles (UAVs).

A lightweight infrared target detection model, G-YOLO, based on an unmanned aerial vehicle (UAV) is proposed to address the issues of low accuracy in target detection of UAV aerial images in complex ground scenarios and large network models that are difficult to apply to mobile or embedded platforms. Firstly, the YOLOv8 backbone feature extraction network is improved and designed based on the lightweight network, GhostBottleneckV2, and the remaining part of the backbone network adopts the depth-separable convolution, DWConv, to replace part of the standard convolution, which effectively retains the detection effect of the model while greatly reducing the number of model parameters and calculations. Secondly, the neck structure is improved by the ODConv module, which adopts an adaptive convolutional structure to adaptively adjust the convolutional kernel size and step size, which allows for more effective feature extraction and detection based on targets at different scales. At the same time, the neck structure is further optimized using the attention mechanism, SEAttention, to improve the model’s ability to learn global information of input feature maps, which is then applied to each channel of each feature map to enhance the useful information in a specific channel and improve the model’s detection performance. Finally, the introduction of the SlideLoss loss function enables the model to calculate the differences between predicted and actual truth bounding boxes during the training process, and adjust the model parameters based on these differences to improve the accuracy and efficiency of object detection. The experimental results show that compared with YOLOv8n, the G-YOLO reduces the missed and false detection rates of infrared small target detection in complex backgrounds. The number of model parameters is reduced by 74.2%, the number of computational floats is reduced by 54.3%, the FPS is improved by 71, which improves the detection efficiency of the model, and the average accuracy (mAP) reaches 91.4%, which verifies the validity of the model for UAV-based infrared small target detection. Furthermore, the FPS of the model reaches 556, and it will be suitable for wider and more complex detection task such as small targets, long-distance targets, and other complex scenes.

During UAV aerial photography tasks, influenced by flight altitude and imaging mechanisms, the target in images often exhibits characteristics such as small size, complex backgrounds, and small inter-class differences. Under single optical modality, the weak and less discriminative feature representation of targets in drone-captured images makes them easily overwhelmed by complex background noise, leading to low detection accuracy, high missed-detection and false-detection rates in current object detection networks. Moreover, such methods struggle to meet all-weather and all-scenario application requirements. To address these issues, this paper proposes DVIF-Net, a visible-infrared fusion network for small-target detection in UAV aerial images, which leverages the complementary characteristics of visible and infrared images to enhance detection capability in complex environments. Firstly, a dual-branch feature extraction structure is designed based on YOLO architecture to separately extract features from visible and infrared images. Secondly, a P4-level cross-modal fusion strategy is proposed to effectively integrate features from both modalities while reducing computational complexity. Meanwhile, we design a novel dual context-guided fusion module to capture complementary features through channel attention of visible and infrared images during fusion and enhance interaction between modalities via element-wise multiplication. Finally, an edge information enhancement module based on cross stage partial structure is developed to improve sensitivity to small-target edges. Experimental results on two cross-modal datasets, DroneVehicle and VEDAI, demonstrate that DVIF-Net achieves detection accuracies of 85.8% and 62%, respectively. Compared with YOLOv10n, it has improved by 21.7% and 10.5% in visible modality, and by 7.4% and 30.5% in infrared modality, while maintaining a model parameter count of only 2.49 M. Furthermore, compared with 15 other algorithms, the proposed DVIF-Net attains SOTA performance. These results indicate that the method significantly enhances the detection capability for small targets in UAV aerial images, offering a high-precision and lightweight solution for real-time applications in complex aerial scenarios.