Transformer networks have excellent performance in various different vision tasks, especially object detection. However, in practical applications, Transformer is difficult to use on-board due to its large computational complexity. In this paper, we propose a new approach for reducing the computation of self-attention, which is called conv-attention. Different from the work of self-attention, conv-attention is inspired by NetVLAD and adopts the probability obtained by soft classification to replace the similarity calculation between query and key. Moreover, we combine the three convolution operations for computing the attention matrix in order to reduce the computational effort. Using the Swin Transformer as a comparison, experiments show that the parameters and FLOPs are reduced by 15% and 16%. Meanwhile, MAP is improved in vehicle object detection, including both FILR and Nuscenses datasets.
In response to the limited accuracy of current three-dimensional (3D) object detection algorithms for small objects, this paper presents a multi-sensor 3D small object detection method based on LiDAR and a camera. Firstly, the LiDAR point cloud is projected onto the image plane to obtain a depth image. Subsequently, we propose a cascaded image fusion module comprising multi-level pooling layers and multi-level convolution layers. This module extracts features from both the camera image and the depth image, addressing the issue of insufficient depth information in the image feature. Considering the non-uniform distribution characteristics of the LiDAR point cloud, we introduce a multi-scale voxel fusion module composed of three sets of VFE (voxel feature encoder) layers. This module partitions the point cloud into grids of different sizes to improve detection ability for small objects. Finally, the multi-level fused point features are associated with the corresponding scale’s initial voxel features to obtain the fused multi-scale voxel features, and the final detection results are obtained based on this feature. To evaluate the effectiveness of this method, experiments are conducted on the KITTI dataset, achieving a 3D AP (average precision) of 73.81% for the hard level of cars and 48.03% for the hard level of persons. The experimental results demonstrate that this method can effectively achieve 3D detection of small objects.
With the rapid advancement of remote sensing technology, remote sensing images are increasingly being used in applications such as geographical monitoring, disaster warning, and urban planning. However, detecting small objects—such as vehicles and small buildings—in such imagery remains challenging due to complex backgrounds, weak features, and interference from factors like terrain, clouds, and lighting, leading to high rates of missed detections and false alarms. To tackle these issues, this paper proposes an improved YOLOv8-based framework for small object detection in remote sensing images. The enhancements include a multi-scale feature fusion mechanism, optimized data augmentation strategies incorporating super-resolution techniques, and a redesigned loss function that emphasizes small objects. These refinements significantly improve the model’s ability to extract discriminative features and detect small targets against cluttered backgrounds. Experimental results demonstrate superior performance across multiple metrics, including precision, recall, mAP50, and mAP50-95, particularly for challenging categories like small vehicles and buildings. This research not only provides an effective solution to the key technical bottleneck in small object detection, advancing the progress of related algorithms, but also offers important theoretical and practical experience for subsequent work.
This paper addresses the issues of insufficient detection accuracy for small targets, occlusion misjudgments, and network lightweighting in drone aerial image recognition, by proposing the MFZD-YOLOv10 algorithm for small target detection. To tackle difficulties caused by aerial photography including tiny target features, dense distribution, and severe occlusions, this paper introduces four primary innovations to improve the YOLOv10 network structure: 1) A novel multi-perception fusion SEC2f module is designed to replace the original C2f module, enhancing the ability of texture feature capturing for tiny targets; 2) Design a new module (GSPPF) of the reception field to enrich model in the process of feature extraction of multi-scale spatial information perception and integration ability, to enhance general characterization of the performance; 3)A novel ZModule feature fusion architecture is introduced to substitute traditional Concat operations, enabling the network to efficiently interact with cross-layer feature and integrate more feature information; 4)The ELA attention mechanism is incorporated to enhance the network's sensitivity to key features and optimize the detection head structure. The high-resolution P2 detection head is used in place of the P4 and P5 detection heads to improve the localization accuracy of small targets. Compared to the original YOLOv10 model, experiments on the VisDrone dataset demonstrate that the improved network achieves 3.4% and 3.2% increases in detection precision and mAP50 respectively, while reducing parameters by 25%, effectively balancing detection performance and model efficiency. This provides a practical solution for UAV small target detection. The improved network's generalization ability is further demonstrated on the DOTA dataset.