Original Network Model Research Articles

Research on intelligent classification of weld defects based on improved ResNeXt network

Welding technology is widely used in the connection of metal workpieces, and then due to improper operation and other reasons, small defects may occur in the welding, which will lead to welding failure and hazardous accidents. Therefore, it is very important to accurately identify welding defects. The automatic defect identification method based on deep learning learns features directly from the full-focus mapping of weld defects and automatically classifies them; the ResNeXt network is selected for improvement, and the SK convolution unit is introduced into the original ResNeXt-50 network model to enhance the adaptive adjustment of sensory field; in order to alleviate the problem of the point-by-point convolution in the ResNeXt-50 which consumes a large amount of In order to alleviate the problem that point-by-point convolution in ResNeXt-50 consumes a large amount of computation, 1×1 group convolution is used instead of the original point-by-point convolution, and channel rearrangement is introduced to alleviate the problem of weakened information exchange of groups caused by group convolution. After improvement, the ResNeXt-50 network model can achieve a classification accuracy of 98.2%, which is 5.5% more than that of the original ResNeXt-50 on this paper's dataset, and the classification accuracy after T-sne visualization also has obvious separation boundary and clustering effects.

Discrete element model for effective electrical conductivity of spark plasma sintered porous materials

AbstractThis paper aims to analyse electrical conduction in partially sintered porous materials using an original resistor network model within discrete element framework. The model is based on sintering geometry, where two particles are connected via neck. Particle-to-particle conductance depends on neck size in sintered materials. Therefore, accurate evaluation of neck size is essential to determine conductance. The neck size was determined using volume preservation criterion. Additionally, grain boundary correction factor was introduced to compensate for any non-physical overlaps between particles, particularly at higher densification. Furthermore, grain boundary resistance was added to account for the porosity within necks. For numerical analysis, the DEM sample was generated using real particle size distribution, ensuring a heterogeneous and realistic microstructure characterized by a maximum-to-minimum particle diameter ratio of 15. The DEM sample was subjected to hot press simulation to obtain geometries with different porosity levels. These representative geometries were used to simulate current flow and determine effective electrical conductivity as a function of porosity. The discrete element model (DEM) was validated using experimentally measured electrical conductivities of porous NiAl samples manufactured using spark plasma sintering (SPS). The numerical results were in close agreement with the experimental results, hence proving the accuracy of the model. The model can be used for microscopic analysis and can also be coupled with sintering models to evaluate effective properties during the sintering process.

Insulating Glove Wearing State Detection for Substation Personnel Based on Faster‐YOLOv8

Electrical workers must wear insulating gloves during daily maintenance. Detecting the condition of these gloves is significant for power safety. Due to the limitations of existing detection algorithms in accuracy and lightweight on power construction sites, this paper proposes a Faster‐YOLOv8 algorithm to detect the wearing condition of insulating gloves. First, it replaces the Bottleneck in the C2f module with the Faster Block and introduces a new C2Faster module. This reduces model parameters while improving performance. Second, the feature pyramid is improved to enhance the combination of deep and shallow semantic information. This improvement aims to improve the network's focus on detecting small targets. Finally, the GSConv is introduced into the neck network to reduce the model parameters and calculation. Additionally, GSConv enhances feature information exchange through shuffle operation. The effectiveness of the Faster‐YOLOv8 algorithm is verified through comparisons with mainstream algorithms such as the Faster‐RCNN algorithm, SSD algorithm, YOLOv5 algorithm, and YOLOv7‐tiny algorithm. The results demonstrate that the Faster‐YOLOv8 algorithm outperforms the above algorithms in terms of detection accuracy. Compared with the original YOLOv8 network model, the mAP is improved by 3.8%, the mAP of the wrong glove increased by 9.8%, the model size is reduced by 33.3%, and the detection speed can reach 81 FPS. © 2024 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.

Vehicle Multi-Object Detection and Tracking Algorithm Based on Improved You Only Look Once 5s Version and DeepSORT

The increasing popularity of vehicles has led to traffic congestion and frequent traffic accidents. Intelligent transportation technology is an effective solution to this problem. In order to improve the accuracy and effectiveness of vehicle detection and tracking, this paper combined the improved YOLOv5s model with the optimized DeepSORT tracking algorithm to detect and track vehicles on traffic roads. Firstly, in the detection model of YOLOv5s, the Attention-based Intra-scale Feature Interaction (AIFI) module is introduced to detect vehicles more quickly and accurately. Secondly, the Kalman filtering (KF) algorithm of DeepSORT is optimized to improve the accuracy of predictions of the vehicle state by using the width to replace the length-to-width ratio of the vehicle prediction box in the original KF algorithm. Finally, in the re-recognition network of DeepSORT, the original Convolutional Neural Network (CNN) model is replaced by an improved ResNet36 as the backbone network for feature extraction. The experimental results show that, compared with the original algorithm, when examining the performance of the improved algorithm in terms of target detection, the recall rate, average accuracy (mAP), and detection speed, are increased by 7.7%, 15.5%, and 14.2%, respectively; in terms of multi-object tracking performance, such as multi-object tracking precision (MOTP) and multi-object tracking accuracy (MOTA), improvements of 14.84% and 9.62%, respectively, are obtained and the total number of times a trajectory is fragmented (Frag) is reduced by 32.52%.These results indicate that the proposed algorithm can meet the requirements of accuracy, real-time detection, and stable vehicle detection and tracking on traffic roads.

An efficient and accurate surface defect detection method for quality supervision of wood panels

The wood panel processing sector is integral to the landscape of industrial manufacturing, and automated detection of wood panel surface defects has become an important guarantee for improving the efficiency and quality of processing production. However, due to the diverse scales and shapes of wood panel surface defects, as well as their complex and varied colors and texture characteristics, the efforts to efficiently and accurately detect surface defects in wood panels through existing methods have fallen short. Therefore, the paper proposes an enhanced YOLOx-tiny deep learning network for wood panel surface defect detection. We introduce new modules multi-pooling feature fusion module and comprehensive feature extraction module, instead of the original SPP and Bottleneck modules to enhance key feature extraction and reduce the number of computational parameters. The experimental results conducted on the self-constructed wood panel surface defects dataset show that the mAP of our proposed method is 95.01%, which is 9.58% higher than the original YOLOx-tiny network model, and the defects recall is 91.46%, which is 13.21% higher compared to the original network. Meanwhile, the method is able to reduce 12.22% of computational parameters, which effectively improves the efficiency of the detection of surface defects on wood panels. In summary, the proposed intelligent surface defect detection approach for wood panels, which utilizes an enhanced YOLOx-tiny deep learning network, has yielded notable outcomes in enhancing both accuracy and efficiency. This method holds significant practical relevance for the wood panel manufacturing sector, offering the potential to enhance both production efficiency and quality. It also explores the automation and intelligent technology in the process of man-made board processing, which provides a valuable reference for the research in related fields.

Detection of Small Lesions on Grape Leaves Based on Improved YOLOv7

The precise detection of small lesions on grape leaves is beneficial for early detection of diseases. In response to the high missed detection rate of small target diseases on grape leaves, this paper adds a new prediction branch and combines an improved channel attention mechanism and an improved E-ELAN (Extended-Efficient Long-range Attention Network) to propose an improved algorithm for the YOLOv7 (You Only Look Once version 7) model. Firstly, to address the issue of low resolution for small targets, a new detection head is added to detect smaller targets. Secondly, in order to increase the feature extraction ability of E-ELAN components in YOLOv7 for small targets, the asymmetric convolution is introduced into E-ELAN to replace the original 3 × 3 convolution in E-ELAN network to achieve multi-scale feature extraction. Then, to address the issue of insufficient extraction of information from small targets in YOLOv7, a channel attention mechanism was introduced and improved to enhance the network’s sensitivity to small-scale targets. Finally, the CIoU (Complete Intersection over Union) in the original YOLOv7 network model was replaced with SIoU (Structured Intersection over Union) to optimize the loss function and enhance the network’s localization ability. In order to verify the effectiveness of the improved YOLOv7 algorithm, three common grape leaf diseases were selected as detection objects to create a dataset for experiments. The results show that the average accuracy of the algorithm proposed in this paper is 2.7% higher than the original YOLOv7 algorithm, reaching 93.5%.

Detection of the Corn Kernel Breakage Rate Based on an Improved Mask Region-Based Convolutional Neural Network

Real-time knowledge of kernel breakage during corn harvesting plays a significant role in the adjustment of operational parameters of corn kernel harvesters. (1) Transfer learning by initializing the DenseNet121 network with pre-trained weights for training and validating a dataset of corn kernels was adopted. Additionally, the feature extraction capability of DenseNet121 was improved by incorporating the attention mechanism of a Convolutional Block Attention Module (CBAM) and a Feature Pyramid Network (FPN) structure. (2) The quality of intact and broken corn kernels and their pixels were found to be coupled, and a linear regression model was established using the least squares method. The results of the test showed that: (1) The MAPb50 and MAPm50 of the improved Mask Region-based Convolutional Neural Network (RCNN) model were 97.62% and 98.70%, in comparison to the original Mask Region-based Convolutional Neural Network (RCNN) model, which were improved by 0.34% and 0.37%, respectively; the backbone FLOPs and Params were 3.09 GMac and 9.31 M, and the feature extraction time was 206 ms; compared to the original backbone, these were reduced by 3.87 GMac and 17.32 M, respectively. The training of the obtained prediction weights for the detection of a picture of the corn kernel took 76 ms, so compared to the Mask RCNN model, it was reduced by 375 ms; based on the concept of transfer learning, the improved Mask RCNN model converged twice as quickly with the loss function using pre-training weights than the loss function without pre-training weights during training. (2) The coefficients of determination R2 of the two models, when the regression models of the pixels and the quality of intact and broken corn kernels were analyzed, were 0.958 and 0.992, respectively. These findings indicate a strong correlation between the pixel characteristics and the quality of corn kernels. The improved Mask RCNN model was used to segment mask pixels to calculate the corn kernel breakage rate. The verified error between the machine vision and the real breakage rate ranged from −0.72% to 0.65%, and the detection time of the corn kernel breakage rate was only 76 ms, which could meet the requirements for real-time detection. According to the test results, the improved Mask RCNN method had the advantages of a fast detection speed and high accuracy, and can be used as a data basis for adjusting the operation parameters of corn kernel harvesters.

Object Detection Based on an Improved YOLOv7 Model for Unmanned Aerial-Vehicle Patrol Tasks in Controlled Areas

When working with objects on a smaller scale, higher detection accuracy and faster detection speed are desirable features. Researchers aim to endow drones with these attributes in order to improve performance when patrolling in controlled areas for object detection. In this paper, we propose an improved YOLOv7 model. By incorporating the variability attention module into the backbone network of the original model, the association between distant pixels is increased, resulting in more effective feature extraction and, thus, improved model detection accuracy. By improving the original network model with deformable convolution modules and depthwise separable convolution modules, the model enhances the semantic information extraction of small objects and reduces the number of model parameters to a certain extent. Pretraining and fine-tuning techniques are used for training, and the model is retrained on the VisDrone2019 dataset. Using the VisDrone2019 dataset, the improved model achieves an mAP50 of 52.3% on the validation set. Through the visual comparative analysis of the detection results in our validation set, we find that the model shows a significant improvement in detecting small objects compared with previous iterations.

Automatic Recognition of Blood Cell Images with Dense Distributions Based on a Faster Region-Based Convolutional Neural Network

In modern clinical medicine, the important information of red blood cells, such as shape and number, is applied to detect blood diseases. However, the automatic recognition problem of single cells and adherent cells always exists in a densely distributed medical scene, which is difficult to solve for both the traditional detection algorithms with lower recognition rates and the conventional networks with weaker feature extraction capabilities. In this paper, an automatic recognition method of adherent blood cells with dense distribution is proposed. Based on the Faster R-CNN, the balanced feature pyramid structure, deformable convolution network, and efficient pyramid split attention mechanism are adopted to automatically recognize the blood cells under the conditions of dense distribution, extrusion deformation, adhesion and overlap. In addition, the Align algorithm for region of interest also contributes to improving the accuracy of recognition results. The experimental results show that the mean average precision of cell detection is 0.895, which is 24.5% higher than that of the original network model. Compared with the one-stage mainstream networks, the presented network has a stronger feature extraction capability. The proposed method is suitable for identifying single cells and adherent cells with dense distribution in the actual medical scene.

Detection of Pine Wood Nematode Infestation Based on Improved YOLOv5

Pine wilt disease is a destructive forest disease caused by the parasitism of pine wood nematodes inside pine trees. Once infected, pine trees quickly wither and die due to the inability to drain water, hence it is also known as pine tree withering disease. It is mainly spread by pine sawyer beetles, with the characteristics of fast dissemination, rapid onset, and high mortality rate, making it a significant global plant disease. Failure to effectively control the pine wilt disease in a timely manner will result in the massive death of pine trees in a short period of time in forest areas. With the continuous development and popularization of artificial intelligence and UAV (Unmanned Aerial Vehicle) technology, combining various methods can timely detect diseased trees, and performing timely treatment and protection can greatly save the time and personnel costs of biological pest control. While it is also conducive to promoting the work of multiple related disciplines, such as bioengineering and greening engineering. In this paper, we first conducted a large number of aerial surveys on the forest area through UAV, collected relevant data, and preprocessed the obtained data, including data augmentation, cleaning, filtering, deduplication, formatting, etc., to ensure data quality and accuracy. Then, we labeled the data, and made certain improvements to the original YOLOv5. We added a new RRAM (Recurrent Residual Attention Module) to the original network model, which enables the network to timely focus on important information in redundant data, thus improving the network’s performance. Compared with the original YOLOv5, our network has a stronger performance.

Bird Detection on Power Transmission Lines Based on Improved YOLOv7

The safety of transmission lines is essential for ensuring the secure and dependable operation of the power grid. However, the harm caused by birds to transmission lines poses a direct threat to their safe operation. The main challenges in detecting birds on lines is that the detected targets are small, densely packed, and susceptible to environmental interference. We introduce a novel dynamic convolutional kernel specifically designed for detecting small and densely packed targets, the ODconv in the backbone of YOLOv7, to capture richer contextual information and improve performance. The substitution of Alpha_GIoU for CIoU in the original YOLOv7 network model serves to refine the loss function, decrease its parameters, and bolster the network’s resilience. The results confirmed that the proposed YOLOv7 with ODConv reached mAP0.5, mAP0.5:0.95, and precision of up to 78.42%, 46.14%, and 73.56% respectively. In contrast to the base model, the enhanced model demonstrated a 2.58% rise in mAP0.5, a 0.72% improvement in mAP0.5:0.95, and an increased precision of 2.34%.

A Forest Wildlife Detection Algorithm Based on Improved YOLOv5s.

A forest wildlife detection algorithm based on an improved YOLOv5s network model is proposed to advance forest wildlife monitoring and improve detection accuracy in complex forest environments. This research utilizes a data set from the Hunan Hupingshan National Nature Reserve in China, to which data augmentation and expansion methods are applied to extensively train the proposed model. To enhance the feature extraction ability of the proposed model, a weighted channel stitching method based on channel attention is introduced. The Swin Transformer module is combined with a CNN network to add a Self-Attention mechanism, thus improving the perceptual field for feature extraction. Furthermore, a new loss function (DIOU_Loss) and an adaptive class suppression loss (L_BCE) are adopted to accelerate the model's convergence speed, reduce false detections in confusing categories, and increase its accuracy. When comparing our improved algorithm with the original YOLOv5s network model under the same experimental conditions and data set, significant improvements are observed, in particular, the mean average precision (mAP) is increased from 72.6% to 89.4%, comprising an accuracy improvement of 16.8%. Our improved algorithm also outperforms popular target detection algorithms, including YOLOv5s, YOLOv3, RetinaNet, and Faster-RCNN. Our proposed improvement measures can well address the challenges posed by the low contrast between background and targets, as well as occlusion and overlap, in forest wildlife images captured by trap cameras. These measures provide practical solutions for enhanced forest wildlife protection and facilitate efficient data acquisition.

YOLOv7 Optimization Model Based on Attention Mechanism Applied in Dense Scenes

With object detection technology, real-time detection of dense scenes has become an important application requirement in various industries, which is of great significance for improving production efficiency and ensuring public safety. However, the current mainstream target detection algorithms have problems such as insufficient accuracy or inability to achieve real-time detection when detecting dense scenes, and to address this problem this paper improves the YOLOv7 model using attention mechanisms that can enhance critical information. Based on the original YOLOv7 network model, part of the traditional convolutional layers are replaced with the standard convolution combined with the attention mechanism. After comparing the optimization results of three different attention mechanisms, CBAM, CA, and SimAM, the YOLOv7B-CBAM model is proposed, which effectively improves the accuracy of object detection in dense scenes. The results on VOC datasets show that the YOLOv7B-CBAM model has the highest accuracy, reaching 87.8%, 1.5% higher than that of the original model, and outperforms the original model as well as other models with improved attention mechanisms in the subsequent results of two other different dense scene practical application scenarios. This model can be applied to public safety detection, agricultural detection, and other fields, saving labor costs, improving public health, reducing the spread and loss of plant diseases, and realizing high-precision, real-time target detection.

Dense Papaya Target Detection in Natural Environment Based on Improved YOLOv5s

Due to the fact that the green features of papaya skin are the same colour as the leaves, the dense growth of fruits causes serious overlapping occlusion phenomenon between them, which increases the difficulty of target detection by the robot during the picking process. This study proposes an improved YOLOv5s-Papaya deep convolutional neural network for achieving dense multitarget papaya detection in natural orchard environments. The model is based on the YOLOv5s network architecture and incorporates the Ghost module to enhance its lightweight characteristics. The Ghost module employs a strategy of grouped convolutional layers and weighted fusion, allowing for more efficient feature representation and improved model performance. A coordinate attention module is introduced to improve the accuracy of identifying dense multitarget papayas. The fusion of bidirectional weighted feature pyramid networks in the PANet structure of the feature fusion layer enhances the performance of papaya detection at different scales. Moreover, the scaled intersection over union bounding box regression loss function is used rather than the complete intersection over union bounding box regression loss function to enhance the localisation accuracy of dense targets and expedite the convergence of the network model training. Experimental results show that the YOLOv5s-Papaya model achieves detection average precision, precision, and recall rates of 92.3%, 90.4%, and 83.4%, respectively. The model’s size, number of parameters, and floating-point operations are 11.5 MB, 6.2 M, and 12.8 G, respectively. Compared to the original YOLOv5s network model, the model detection average precision is improved by 3.6 percentage points, the precision is improved by 4.3 percentage points, the number of parameters is reduced by 11.4%, and the floating-point operations are decreased by 18.9%. The improved model has a lighter structure and better detection performance. This study provides the theoretical basis and technical support for intelligent picking recognition of overlapping and occluded dense papayas in natural environments.

Airport Perimeter Damage Detection Method Based on Alexnet Optimization

With the continuous development and expansion of airports, people pay more and more attention to the security of airport boundary. The damage of the boundary is one of the main factors leading to the security of the boundary, so it is very important to develop an automatic detection method of the boundary damage. Based on AlexNet deep learning model, this paper proposes an airport boundary damage detection method. By processing the collected video of the airport boundary, making the boundary damage data set, and then optimizing AlexNet, the experimental results show that compared with the original AlexNet network model, the improved AlexNet network model has shorter training time and significantly improved recognition accuracy, which can effectively identify the boundary damage.

Research on Safety Helmet Detection Algorithm Based on Improved YOLOv5s.

Safety helmets are essential in various indoor and outdoor workplaces, such as metallurgical high-temperature operations and high-rise building construction, to avoid injuries and ensure safety in production. However, manual supervision is costly and prone to lack of enforcement and interference from other human factors. Moreover, small target object detection frequently lacks precision. Improving safety helmets based on the helmet detection algorithm can address these issues and is a promising approach. In this study, we proposed a modified version of the YOLOv5s network, a lightweight deep learning-based object identification network model. The proposed model extends the YOLOv5s network model and enhances its performance by recalculating the prediction frames, utilizing the IoU metric for clustering, and modifying the anchor frames with the K-means++ method. The global attention mechanism (GAM) and the convolutional block attention module (CBAM) were added to the YOLOv5s network to improve its backbone and neck networks. By minimizing information feature loss and enhancing the representation of global interactions, these attention processes enhance deep learning neural networks' capacity for feature extraction. Furthermore, the CBAM is integrated into the CSP module to improve target feature extraction while minimizing computation for model operation. In order to significantly increase the efficiency and precision of the prediction box regression, the proposed model additionally makes use of the most recent SIoU (SCYLLA-IoU LOSS) as the bounding box loss function. Based on the improved YOLOv5s model, knowledge distillation technology is leveraged to realize the light weight of the network model, thereby reducing the computational workload of the model and improving the detection speed to meet the needs of real-time monitoring. The experimental results demonstrate that the proposed model outperforms the original YOLOv5s network model in terms of accuracy (Precision), recall rate (Recall), and mean average precision (mAP). The proposed model may more effectively identify helmet use in low-light situations and at a variety of distances.

Improved YOLOv7 Network Model for Gangue Selection Robot for Gangue and Foreign Matter Detection in Coal.

Coal production often involves a substantial presence of gangue and foreign matter, which not only impacts the thermal properties of coal and but also leads to damage to transportation equipment. Selection robots for gangue removal have garnered attention in research. However, existing methods suffer from limitations, including slow selection speed and low recognition accuracy. To address these issues, this study proposes an improved method for detecting gangue and foreign matter in coal, utilizing a gangue selection robot with an enhanced YOLOv7 network model. The proposed approach entails the collection of coal, gangue, and foreign matter images using an industrial camera, which are then utilized to create an image dataset. The method involves reducing the number of convolution layers of the backbone, adding a small size detection layer to the head to enhance the small target detection, introducing a contextual transformer networks (COTN) module, employing a distance intersection over union (DIoU) loss border regression loss function to calculate the overlap between predicted and real frames, and incorporating a dual path attention mechanism. These enhancements culminate in the development of a novel YOLOv71 + COTN network model. Subsequently, the YOLOv71 + COTN network model was trained and evaluated using the prepared dataset. Experimental results demonstrated the superior performance of the proposed method compared to the original YOLOv7 network model. Specifically, the method exhibits a 3.97% increase in precision, a 4.4% increase in recall, and a 4.5% increase in mAP0.5. Additionally, the method reduced GPU memory consumption during runtime, enabling fast and accurate detection of gangue and foreign matter.

Fast and Accurate ROI Extraction for Non-Contact Dorsal Hand Vein Detection in Complex Backgrounds Based on Improved U-Net.

In response to the difficulty of traditional image processing methods to quickly and accurately extract regions of interest from non-contact dorsal hand vein images in complex backgrounds, this study proposes a model based on an improved U-Net for dorsal hand keypoint detection. The residual module was added to the downsampling path of the U-Net network to solve the model degradation problem and improve the feature information extraction ability of the network; the Jensen-Shannon (JS) divergence loss function was used to supervise the final feature map distribution so that the output feature map tended to Gaussian distribution and improved the feature map multi-peak problem; and Soft-argmax is used to calculate the keypoint coordinates of the final feature map to realize end-to-end training. The experimental results showed that the accuracy of the improved U-Net network model reached 98.6%, which was 1% better than the original U-Net network model; the improved U-Net network model file was only 1.16 M, which achieved a higher accuracy than the original U-Net network model with significantly reduced model parameters. Therefore, the improved U-Net model in this study can realize dorsal hand keypoint detection (for region of interest extraction) for non-contact dorsal hand vein images and is suitable for practical deployment in low-resource platforms such as edge-embedded systems.

Intelligent Fault Diagnosis of Rolling Bearings Using Efficient and Lightweight ResNet Networks Based on an Attention Mechanism (September 2022)

Focusing on the problems of complex structure and low feature extraction efficiency that exist in some traditional neural network algorithms, an improved convolutional neural network, combined with a convolutional attention mechanism, is proposed to im- prove the network feature extraction efficiency and reduce the network complexity. First, the convolutional block attention module (CBAM) was introduced, which can strengthen the extraction of useful features from the data, thereby improving the network feature extraction efficiency. Second, the residual structure was optimized, and the original multi-layer convolution was simplified to a single-layer convolution, which reduces the complexity of the model and improves the diagnostic efficiency. Then, the Rectified Linear Unit (ReLU) activation function in the model was replaced with the better-performing Exponential Linear Unit (ELU) activation function. Finally, the overall structure of the ResNet model was optimized. The CBAM-ResNet network model was based on the traditional ResNet18 model and consists of a CBAM module and the optimized ResNet18 model. Finally, experiments were performed using sensor data collected from rolling bearing test benches. The experimental results show that the feature extraction capability and efficiency of the new model are better than those of several other classic deep network models and the original network model, and this reduces the time complexity and model size while also maintaining a high accuracy rate.

Estimating thermohaline structures in the tropical Indian Ocean from surface parameters using an improved CNN model

Accurately estimating the ocean’s subsurface thermohaline structure is essential for advancing our understanding of regional and global ocean dynamics. In this study, we propose a novel neural network model based on Convolutional Block Attention Module-Convolutional Neural Network (CBAM-CNN) to simultaneously estimate the ocean subsurface thermal structure (OSTS) and ocean subsurface salinity structure (OSSS) in the tropical Indian Ocean using satellite observations. The input variables include sea surface temperature (SST), sea surface salinity (SSS), sea surface height anomaly (SSHA), eastward component of sea surface wind (ESSW), northward component of sea surface wind (NSSW), longitude (LON), and latitude (LAT). We train and validate the model using Argo data, and compare its accuracy with that of the original Convolutional Neural Network (CNN) model using root mean square error (RMSE), normalized root mean square error (NRMSE), and determination coefficient (R²). Our results show that the CBAM-CNN model outperforms the CNN model, exhibiting superior performance in estimating thermohaline structures in the tropical Indian Ocean. Furthermore, we evaluate the model’s accuracy by comparing its estimated OSTS and OSSS at different depths with Argo-derived data, demonstrating that the model effectively captures most observed features using sea surface data. Additionally, the CBAM-CNN model demonstrates good seasonal applicability for OSTS and OSSS estimation. Our study highlights the benefits of using CBAM-CNN for estimating thermohaline structure and offers an efficient and effective method for estimating thermohaline structure in the tropical Indian Ocean.