This paper addresses the challenge of small object detection in remote sensing image recognition by proposing an improved YOLOv8-based lightweight attention cross-scale feature fusion model named LACF-YOLO. Prior to the backbone network outputting feature maps, this model introduces a lightweight attention module, Triplet Attention, and replaces the Concatenation with Fusion (C2f) with a more convenient and higher-performing dilated inverted convolution layer to acquire richer contextual information during the feature extraction phase. Additionally, it employs convolutional blocks composed of partial convolution and pointwise convolution as the main body of the cross-scale feature fusion network to integrate feature information from different levels. The model also utilizes the faster-converging Focal EIOU loss function to enhance accuracy and efficiency. Experimental results on the DOTA and VisDrone2019 datasets demonstrate the effectiveness of the improved model. Compared to the original YOLOv8 model, LACF-YOLO achieves a 2.9% increase in mAP and a 4.6% increase in mAP_S on the DOTA dataset and a 3.5% increase in mAP and a 3.8% increase in mAP_S on the VisDrone2019 dataset, with a 34.9% reduction in the number of parameters and a 26.2% decrease in floating-point operations. The model exhibits superior performance in aerial object detection.

Traditional fault diagnosis methods often require extracting features from raw vibration signals based on prior knowledge, which are then input into intelligent classifiers for pattern recognition. This process is prone to information loss and can be inaccurate when relying on human experience for fault identification. To address this issue, this paper proposes an intelligent fault classification and diagnosis model for rolling bearings based on Fast Fourier Transform (FFT) combined with a time convolutional network (SE-TCN) incorporating an attention mechanism, with a Support Vector Machine (SVM) used as the classifier. First, the FFT is applied to transform the collected raw time-domain data of bearing faults into the frequency domain, obtaining the sequence information in the frequency domain. Second, the frequency–domain sequence data are fed into the SE-TCN model, which uses multiple convolutional layers and a channel attention mechanism to extract deep fault features. Finally, the extracted feature vectors are input into the SVM classifier, and the Particle Swarm Optimization (PSO) algorithm is used to optimize the SVM parameters. The optimal separating hyperplane is obtained through training to classify the fault types of the rolling bearings. To verify the effectiveness and diagnostic performance of the proposed method, experiments are conducted using bearing fault datasets from Case Western Reserve University (CWRU) and a laboratory self-built fault diagnosis experimental platform. The experimental results show that the classification accuracy of the proposed method exceeds 99% on the CWRU test dataset, and it also demonstrates advantages in handling small sample data, with an accuracy of over 90%. Additionally, it exhibits good diagnostic performance on the bearing fault data collected from the laboratory self-built platform. The results validate the effectiveness of the proposed classification model in bearing a fault diagnosis.

This study addresses the challenges of detecting small targets and targets with significant scale variations in UAV aerial images. We propose an improved YOLOv5 model, named LCM‐YOLO, to tackle these challenges. Initially, a local fusion mechanism is introduced into the C3 module, forming the C3‐LFM module to enhance feature information acquisition during feature extraction. Subsequently, the CCFM is employed as the neck structure of the network, leveraging its lightweight convolution and cross‐scale feature fusion characteristics to effectively improve the model's ability to integrate target features at different levels, thereby enhancing its adaptability to scale variations and detection performance for small targets. Additionally, a multi‐head attention mechanism is integrated at the front end of the detection head, allowing the model to focus more on the detailed information of small targets through weight distribution. Experiments on the VisDrone2019 dataset show that LCM‐YOLO has excellent detection capabilities. Compared to the original YOLOv5 model, its mAP50 and mAP50‐95 metrics are improved by 7.2% and 5.1%, respectively, reaching 40.7% and 22.5%. This validates the effectiveness of the LCM‐YOLO model for detecting small and multi‐scale targets in complex backgrounds.

The fusion of infrared and visible light images is a significant image enhancement technique that leverages the strengths of both infrared and visible light images to achieve high-quality results. However, during the fusion process, information loss is inevitable. To minimize this loss and enhance contrast and detail information, this paper proposes a novel image fusion algorithm based on attention mechanisms, termed Dual-Branch Squared Quadratic (DBSQ)Fusion. This method fully integrates the characteristics of different source images and processes them through specifically designed channels to maximize the retention of important information from the original images. Additionally, Feature Contrast Enhancement Fusion Network(FCEFN) is designed to exploit the differences between infrared and visible light features, enabling information complementarity by separating these distinct features. Furthermore, a Sparse Multi-scale Loss function (SMLoss) is employed to enhance multi-scale information and edge textures, thereby further improving the quality of the fused image. Extensive experiments were conducted on multiple datasets, and various evaluation metrics were utilized to demonstrate the superiority of our proposed model.

Maritime remote sensing ship detection has long been plagued by two major issues: the failure of geometric priors due to the extreme length-to-width ratio of ships; and the sharp drop in edge signal-to-noise ratio caused by the overlapping chromaticity domain between ships and seawater, which leads to unsatisfactory accuracy of existing detectors in such scenarios. Therefore, this paper proposes an optical remote sensing ship detection model combining channel shuffling and bilinear interpolation, named CSBI-YOLO. The core innovations include three aspects: First, a group shuffling feature enhancement module is designed, embedding parallel group bottlenecks and channel shuffling mechanisms into the interface between the YOLOv8 backbone and neck to achieve multi-scale semantic information coupling with a small number of parameters. Second, an edge-gated upsampling unit is constructed, using separable Sobel magnitude as structural prior and a learnable gating mechanism to suppress low-contrast noise on the sea surface. Third, an R-IoU-Focal loss function is proposed, introducing logarithmic curvature penalty and adaptive weights to achieve joint optimization in three dimensions: location, shape, and scale. Dual validation was conducted on the self-built SlewSea-RS dataset and the public DOTA-ship dataset. The results show that on the SlewSea-RS dataset, the mAP₅₀ and mAP_50–95 values of the CSBI-YOLO model increased by 6% and 5.4%, respectively. On the DOTA-ship dataset, comparisons with various models demonstrate that the proposed model outperforms others, proving the excellent performance of the CSBI-YOLO model in detecting maritime ship targets.