Chao Fang

The field of underwater robotics is experiencing rapid growth, wherein accurate object detection constitutes a fundamental component. Given the prevalence of false alarms and omission errors caused by intricate subaquatic conditions and substantial image noise, this study introduces an enhanced detection framework that combines the YOLOv8 architecture with multimodal visual fusion methodology. To solve the problem of degraded detection performance of the model in complex environments like those with low illumination, features from Visible Light Image are fused with the Thermal Distribution Features exhibited by Infrared Image, thereby yielding more comprehensive image information. Furthermore, to precisely focus on crucial target regions and information, a Multi-Scale Cross-Axis Attention Mechanism (MSCA) is introduced, which significantly enhances Detection Accuracy. Finally, to meet the lightweight requirement of the model, an Efficient Shared Convolution Head (ESC_Head) is designed. The experimental findings reveal that the YOLOv8-FUSED framework attains a mean average precision (mAP) of 82.1%, marking an 8.7% enhancement compared to the baseline YOLOv8 architecture. The proposed approach also exhibits superior detection capabilities relative to existing techniques while simultaneously satisfying the critical requirement for real-time underwater object detection. Moreover, the proposed system successfully meets the essential criteria for real-time detection of underwater objects.

This study introduces a sandwich structure model and employs the Monte Carlo (MC) technique to investigate the electromagnetic interference shielding effectiveness of carbon nanotubes (CNT) nanocomposite foams. Firstly, a two-dimensional Voronoi diagram to simulate the composite foam structure is established. Then, by assuming that the edges of the polygon have a sandwich structure of insulator-conductor-insulator type, we evaluate the optical transmission matrix of the sandwich structure and calculate the transmission and reflection probabilities. Finally, the MC method is used to model the path of numerous light rays entering the composite foams. By statistically analyzing the electromagnetic fields of waves that are transmitted and reflected, we obtained the electromagnetic shielding properties of composite foams in the X-band frequency range. The results indicate that the foams amplify the multiple reflections of electromagnetic waves, thereby increasing its capacity to absorb electromagnetic energy. It is demonstrated that the variation of SE with CNT concentration matches well with the experimental findings. A detailed study is conducted on the effects of foam size, sandwich structure thickness and conductivity, and electromagnetic wave frequency on SE.

Underwater object detection technology is widely used in fields such as ocean exploration. However, due to the complex underwater environment, issues like light attenuation and scattering lead to low detection accuracy, which fails to meet the requirements. To address these issues, we propose an improved YOLOv8n-based model called YOLOv8-UC. This model incorporates a modified Dilation-wise Residual (DWR) C2f module to enhance the ability to extract features from the network’s high-level expandable receptive fields. It also integrates the Large Separable Kernel Attention (LSKA) module with the SPPF of YOLOv8 to enhance multi-scale feature extraction capabilities, reducing the loss of details. To solve the problem of redundant parameters and computational load in the detection head, the original detection head is replaced with a shared parameter structure, and RepConv is introduced. Additionally, the Inner-SIoU loss function is improved by using auxiliary boundaries at different scales to accelerate bounding box regression and improve detection accuracy. Experimental results show that the designed YOLOv8-UC achieves an mAP@0.5 of 79.3%, with a 6.9% increase in detection accuracy (P) and a 5.9% increase in precision (mAP@0.5) compared to YOLOv8n, demonstrating the effectiveness and application prospects of this method.

This study introduces an improved YOLOv8 model tailored for detecting objects in road scenes. To overcome the limitations of standard convolution operations in adapting to varying targets, we introduce Adaptive Kernel Convolution (AKconv). AKconv dynamically adjusts the convolution kernel’s shape and size, enhancing the backbone network’s feature extraction capabilities and improving feature representation across different scales. Additionally, we employ a Multi-Scale Dilated Attention (MSDA) mechanism to focus on key target features, further enhancing feature representation. To address the challenge posed by YOLOv8’s large down sampling factor, which limits the learning of small target features in deeper feature maps, we add a small target detection layer. Finally, to improve model training efficiency, we introduce a regression loss function with a Wise-IoU dynamic non-monotonic focusing mechanism. With these enhancements, our improved YOLOv8 model excels in road scene object detection tasks, achieving a 5.6 percentage point improvement in average precision over the original YOLOv8n on real road datasets.

In order to monitor the growth and development of tomatoes, and improve the efficiency of flower and fruit thinning and tomato picking, we propose a model TF-YOLOv5s for the detection of tomato flowers and fruits in natural environments. Based on the YOLOv5s model, a C3Faster module is introduced to reduce the number of parameters and calculations while maintaining detection accuracy. The regular convolution is replaced by depth-wise separable convolution (DWConv) to avoid parameter redundancy. To improve the convergence and accuracy of the model, we replace Complete Intersection over Union (CIoU) loss with Efficient Intersection over Union (EIOU) loss. The Squeeze and Excitation (SE) attention mechanism is added to improve the model’s sensitivity to the features of the tomato flowers and fruits. Compared with the baseline model, the number of parameters is reduced by 54.5%, the weight file is reduced by 52.8%, the Floating-point Operation Per second (FLOPs) is reduced by 48.7%, and the detection accuracy of tomato flowers and fruits is increased by 1.4%. Furthermore, the improved algorithm is deployed on two edge computing devices to verify its effectiveness. Experimental results show that the algorithm in this paper can achieve high detection with less computational resources. This algorithm has the potential value of application in practical tomato production.

Skin cancer is one of the most common malignant tumors. In order to completely remove tumors and accurately identify and segment lesion areas, this paper proposes a skin lesion segmentation method based on U2Net. Firstly, deep separable convolution is used to replace conventional convolution for feature extraction of RSU blocks. Embedded branch structures are added at skip connections, and Receptive Field Block(RFB) modules are used to integrate semantic information at different scales. Then, Add Attention Gates (AG) and Normalization-based Attention Modules (NAM), which together form the NAM-Gate Attention mechanism can redistribute image feature weights to reduce the loss of spatial feature information. This article selects Dice, Accuracy, and F1 Score as evaluation indicators and uses the ISIC2018 Task 1 dataset to evaluate our model. The experimental results show that the methods Dice, Accuracy, and F1 Score in this paper have improved by 2.06%, 1.01%, and 1.50%, respectively.