Unmanned aerial vehicle (UAV) image object detection has great application value in the military and civilian fields. However, the objects in the captured images from UAVs have problems of large-scale variation, complex backgrounds, and a large proportion of small objects. To resolve these problems, a multi-scale object detector based on coordinate and global information aggregation is proposed, named CGMDet. Firstly, a Coordinate and Global Information Aggregation Module (CGAM) is designed by aggregating local, coordinate, and global information, which can obtain features with richer context information. Secondly, a Feature Fusion Module (FFM) is proposed, which can better fuse features by learning the importance of different scale features and improve the representation ability of multi-scale features by reusing feature maps to help models better detect multi-scale objects. Moreover, more location information of low-level feature maps is integrated to improve the detection results of small targets. Furthermore, we modified the bounding box regression loss of the model to make the model more accurately regress the bounding box and faster convergence. Finally, we tested the CGMDet on VisDrone and UAVDT datasets. The proposed CGMDet improves mAP0.5 by 1.9% on the VisDrone dataset and 3.0% on the UAVDT dataset.

Unmanned Aerial Vehicles (UAVs) image object detection has great application value in military and civilian fields. However, the objects in the captured images from UAVs have problems of large scale variation, complex backgrounds, and a large proportion of small objects. To resolve these problems, a multi-scale object detector based on coordinate and global information aggregation is proposed, named CGMDet. Firstly, a Coordinate and Global Information Aggregation Module (CGAM) is designed by aggregating local, coordinate, and global information, which can obtain features with richer context information. Secondly, a Multi-Feature Fusion Pyramid Network (MF-FPN) is proposed, which can better fuse features of different scales and obtain features containing more context information through repeated use of feature maps, to better detect multi-scale targets. Moreover, more location information of low-level feature maps is integrated to improve the detection results of small targets. Furthermore, we modified the bounding box regression loss of the model to make the model more accurately regress the bounding box and faster convergence. Finally, the proposed CGMDet was tested on VisDrone and UAVDT datasets and mAP0.5 of 50.9% and 48% was obtained, respectively. At the same time, our detector achieved the best results compared to other detectors.

Object detection in unmanned aerial vehicle (UAV) imagery is highly valuable in both military and civilian domains. However, deploying conventional detectors is challenging due to limited computing resources on UAV platforms. In addition, UAV images often contain densely arranged objects, demanding higher accuracy from detectors. In response to these challenges, we proposed the lightweight aerial object detection method based on mask feature enhancement, named BFDet. First, to enhance the feature extraction capability of the model under limited parameters, we designed the balanced channel attention layer (BCA Layer). The BCA Layer improves the extraction of high-quality channel information during feature processing. Second, to enhance model performance while minimizing parameters, we designed a feature extraction network utilizing the efficient feature extraction module (EFEM) and the downsampling module (DM) as its backbone. This design reduces the number of parameters while improving the extraction of discriminative features. Third, to enrich the semantic information of features, we proposed a progressive space pyramid pooling (PSPP) module, obtaining features with rich semantic information to enhance model performance. Finally, to tackle the challenge of detecting densely packed objects, we proposed a mask information enhancement module (MIEM) to obtain the fine-grained features, enhancing the model’s discriminative ability in dense scenes. The proposed BFDet underwent extensive experimentation on both the VisDrone and UAVDT datasets. The mAP0.5 and FPS values of BFDet reach 51.4% and 33, respectively, on the VisDrone dataset, and the parameter number is only 5.6 M.

Object detection in remote sensing images (RSIs) has become crucial in recent years. However, researchers often prioritize detecting small objects, neglecting medium- to large-sized ones. Moreover, detecting objects hidden in shadows is challenging. Additionally, most detectors have extensive parameters, leading to higher hardware costs. To address these issues, this paper proposes a multi-scale and high-precision lightweight object detector named MHLDet. Firstly, we integrated the SimAM attention mechanism into the backbone and constructed a new feature-extraction module called validity-neat feature extract (VNFE). This module captures more feature information while simultaneously reducing the number of parameters. Secondly, we propose an improved spatial pyramid pooling model, named SPPE, to integrate multi-scale feature information better, enhancing the model to detect multi-scale objects. Finally, this paper introduces the convolution aggregation crosslayer (CACL) into the network. This module can reduce the size of the feature map and enhance the ability to fuse context information, thereby obtaining a feature map with more semantic information. We performed evaluation experiments on both the SIMD dataset and the UCAS-AOD dataset. Compared to other methods, our approach achieved the highest detection accuracy. Furthermore, it reduced the number of parameters by 12.7% compared to YOLOv7-Tiny. The experimental results illustrated that our proposed method is more lightweight and exhibits superior detection accuracy compared to other lightweight models.

Deep Q-Network (DQN) has demonstrated significant application value in path planning. However, several challenges remain in UAV path planning tasks, such as difficulty in perceiving dependencies among the components of the input state, imbalanced experience sampling, and the lack of constraints on action selection. To address these issues, this paper proposes a DQN framework based on a semantic-group self-attention mechanism (SGSA-DQN). First, a semantic-group self-attention Q-network (SGSAQ-network) is designed to improve the accuracy of the network’s Q-value output by capturing the dependencies among semantic component groups in the state vector. Second, a guidance- and perception-based action selection algorithm (GPAS) is proposed to reduce unnecessary exploration in the early training stage and accelerate model convergence. Then, a coverage-constrained experience replay mechanism (CCER) is proposed to balance experience sampling probabilities across different state spaces and enhance model robustness. Finally, a Lyapunov safety constraint (LSC) is introduced to mask unsafe actions and improve the task success rate. Experimental results demonstrate that the proposed method achieves an average success rate of 79.17% in UAV path planning tasks.