Object detection is a fundamental task in computer vision. Recently, deep-learning-based object detection has made significant progress. However, due to large variations in target scale, the predominance of small targets, and complex backgrounds in remote sensing imagery, remote sensing object detection still faces challenges, including low detection accuracy, poor real-time performance, high missed detection rates, and high false detection rates in practical applications. To enhance remote sensing target detection performance, this study proposes a new model, the remote sensing detection transformer (RS-DETR). First, we incorporate cascaded group attention (CGA) into the attention-driven feature interaction module. By capturing features at different levels, it enhances the interaction between features through cascading and improves computational efficiency. Additionally, we propose an enhanced bidirectional feature pyramid network (EBiFPN) to facilitate multi-scale feature fusion. By integrating features across multiple scales, it improves object detection accuracy and robustness. Finally, we propose a novel bounding box regression loss function, Focaler-GIoU, which makes the model focus more on difficult samples, improving detection performance for small and overlapping targets. Experimental results on the satellite imagery multi-vehicles dataset (SIMD) and the high-resolution remote sensing object detection (TGRS-HRRSD) dataset show that the improved algorithm achieved mean average precision (mAP) of 78.2% and 91.6% at an intersection over union threshold of 0.5, respectively, which is an improvement of 2.0% and 1.5% over the baseline model. This result demonstrates the effectiveness and robustness of our proposed method for remote sensing image object detection.

The main task of remote sensing image target detection is to locate and classify the targets of interest in remote sensing images, which plays an important role in intelligence investigation, disaster relief, industrial application, and other fields. However, the targets in the remote sensing image scene have special problems such as special perspective, scale diversity, multi-direction, small targets, and high background complexity. In this paper, the YOLOv5 target detection algorithm is improved according to the above characteristics. Aiming at the problem of large target size span in remote sensing images, this paper uses K-Means++ clustering algorithm to eliminate the problem in which the original clustering algorithm is sensitive to initial position, noise, and outliers, and optimizes the instability caused by K-Means clustering to obtain preset anchor frames. Aiming at the redundancy of background information around the location of remote sensing image targets, the large number of small targets, and the denseness of targets, a double IoU-aware decoupling head (DDH) is introduced at the output end to replace the coupled yolo head, which eliminates the interference caused by different task sharing parameters. At the same time, the correlation between positioning accuracy and classification accuracy is improved by the IoU-aware method. The attention mechanism is introduced into the backbone network to optimize the detection and small target detection in complex backgrounds. The mAP of the improved YOLOv5 algorithm is improved by 9%, and the detection effect of small targets and dense targets is significantly improved. At the same time, the paper has achieved good results through the verification of the DIOR remote sensing data set and has also achieved good performance advantages in comparison with other models.

With the widespread application of ship detection in civilian and military domains, ship detection in optical remote sensing images has become an important research direction. The core task is to determine the presence of ship targets in remote sensing images and to perform detection, classification, and localization. However, ship targets in remote sensing images exhibit multi-scale characteristics, complex backgrounds, and varying meteorological conditions, leading to challenges such as low detection accuracy, false positives, and missed detections, particularly for small-scale ships. To address these issues, we propose an effective ship detection algorithm—Ship Detection Transformer (Ship-DETR). First, we introduce the high-low frequency (HiLo) attention into the intra-scale feature interaction module to enhance the extraction of both high- and low-frequency features, reduce computational complexity, and improve detection performance. To strengthen the model’s multi-scale fusion capability for ship features of different scales, we incorporate the bidirectional feature pyramid network (BiFPN) to optimize cross-scale feature fusion and add the $S_{2}$ features to further enhance the representation of small-scale ship features. Additionally, we replace traditional downsampling operations with the Haar wavelet-based downsampling (HWD) module, which reduces the spatial resolution of feature maps while preserving edge and texture details as much as possible, thereby minimizing feature information loss. Experimental results on the LEVIR-Ship dataset and the HRSC2016-MS dataset demonstrate that Ship-DETR improves the ship detection accuracy while maintaining a small number of parameters and computational load. Ship-DETR achieved a mean average precision of 75.7% and 73.1% on the two datasets, respectively, representing an improvement of 3.5% and 1.9% compared to the baseline model. Compared with other mainstream object detection algorithms, our proposed model shows superior detection performance.

Pedestrian detection is a crucial task in computer vision, applicable in object tracking, video surveillance, and autonomous driving. Recent years have witnessed substantial advancements in pedestrian detection due to the fast evolution of deep learning in object detection. Nonetheless, obstacles such as inadequate detection accuracy persist, mostly because of varied pedestrian postures and intricate environments. This study proposes the RT-DETR-improved model to overcome these issues based on the real-time detection transformer (RT-DETR). First, we incorporate the high-low frequency (HiLo) attention into the encoder, therefore enhancing the model’s detection performance. Furthermore, we present a nonlinear feature fusion module that fuses information from various feature scales and contexts more successfully. We also introduce a novel loss function, InnerMPDIoU, to enhance detection efficacy in congested environments. To evaluate our model’s performance, extensive experiments are conducted on the CityPersons dataset. Compared to the baseline model, the RT-DETR-improved model attains a 4.2% enhancement in mAP50, a 2.0% improvement in mAP, a 2.2% rise in accuracy, and a 3.1% gain in recall. The results demonstrate that the proposed method exhibits superior detection accuracy and robustness.

With the rapid development of deep learning networks, the accuracy of generic object detection has consistently improved. Nonetheless, small object detection tasks still face a range of challenges. On one hand, the limited pixel size of small objects severely constrains their visual features in images, making them susceptible to distortion and difficult to distinguish from background noise. On the other hand, small objects often appear in complex scenarios with severe occlusion and dense arrangement, further increasing the complexity and difficulty of small object detection tasks. In this context, this study proposes a new model, CenterNet-Elite, to overcome these challenges. To address the issue of feature information loss resulting from multiple downsamplings during feature extraction for small objects, we introduce the spatial and channel reconstruction convolution (SCConv) into the bottleneck to reduce spatial and channel redundancy and enhance feature representation. In the meantime, we construct multiple short connections to integrate feature maps of the same scale during the downsampling and upsampling stages, thereby retaining critical shallow spatial information. We introduce a multi-scale pooling module, SPPCSPC, to address the challenge of significant variations in object scale. This module obtains receptive fields of different sizes through max-pooling layers of diverse sizes, adapting to changes in object scale on the feature maps. Furthermore, we introduce the content-aware reassembly of features (CARAFE) to replace deconvolution, refining the upsampling process to enhance the quality of feature maps. A series of comparative experiments and ablation studies demonstrate the effectiveness of our method in small object detection. The CenterNet-Elite achieves a 2.3% increase in the average precision on the large-scale small object detection dataset (SODA-D).