In recent years, solar photovoltaic (PV) energy, as a clean energy source, has received widespread attention and experienced rapid growth worldwide. However, the rapid growth of PV power deployment also brings important challenges to the maintenance of PV panels, and in order to solve this problem, this paper proposes an innovative algorithm based on PA-YOLO. First, we propose to use PA-YOLO’s asymptotic feature pyramid network (AFPN) instead of YOLOv7’s backbone network to support direct interactions of nonadjacent layers and avoid large semantic gaps between nonadjacent layers. For the occlusion problem of dense targets in the dataset, we introduce a repulsive loss function, which successfully reduces the occurrence of false detection situations. Finally, we propose a customized convolutional block equipped with an EMA mechanism to enhance the perceptual and expressive capabilities of the model. Experimental results on the dataset show that our proposed model achieves excellent performance with an average accuracy (mAP) of 94.5%, which is 6.8% higher than YOLOv7. In addition, our algorithm also succeeds in drastically reducing the model size from 71.3 MB to 48.4 MB, which well demonstrates the effectiveness of the model.

During the maintenance and management of solar photovoltaic (PV) panels, how to efficiently solve the maintenance difficulties becomes a key challenge that restricts their performance and service life. Aiming at the multi-defect-recognition challenge in PV-panel image analysis, this study innovatively proposes a new algorithm for the defect detection of PV panels incorporating YOLOv7-GX technology. The algorithm first constructs an innovative GhostSlimFPN network architecture by introducing GSConv and depth-wise separable convolution technologies, optimizing the traditional neck network structure. Then, a customized 1 × 1 convolutional module incorporating the GAM (Global Attention Mechanism) attention mechanism is designed in this paper to improve the ELAN structure, aiming to enhance the network’s perception and representation capabilities while controlling the network complexity. In addition, the XIOU loss function is introduced in the study to replace the traditional CIOU loss function, which effectively improves the robustness and convergence efficiency of the model. In the training stage, the sample imbalance problem is effectively solved by implementing differentiated weight allocations for different images and categories, which promotes the balance of the training process. The experimental data show that the optimized model achieves 94.8% in the highest mAP value, which is 6.4% higher than the original YOLOv7 network, significantly better than other existing models, and provides solid theoretical and technical support for further research and application in the field of PV-panel defect detection.

In the operation and maintenance of solar photovoltaic panels, maintenance issues are one of the important factors that limit their performance and lifetime. In this paper, an optimization algorithm based on YOLOv7-GX for PV panel defect detection is proposed for the problem of multi-fault identification of PV panel images. First, a detection layer dedicated to detecting tiny targets is designed, and a prior frame suitable for the dataset of this paper is generated using k-means and genetic algorithms to improve the detection performance of very small targets and reduce the false detection rate. Second, a custom 1x1 convolutional block with GAM(Global Attention Mechanism) attention mechanism is proposed to enhance the perceptual and expressive capabilities of the model while avoiding increasing the complexity of the network structure. Then, the XIOU loss function is adopted to replace the original CIOU(Complete-IOU) loss function to enhance the robustness and convergence speed of the model. Finally, during the training process, different weights are assigned to different images and categories to solve the problem of unbalanced sample distribution, so as to achieve balanced training. Experiments show that the highest mAP of the improved method reaches 93.4% better than other existing models.

Against the backdrop of the global energy transition, solar photovoltaic (PV) technology has garnered significant attention due to its environmentally friendly attributes and has exhibited rapid growth trends. While the rapid deployment of PV systems facilitates installation, accompanying maintenance challenges cannot be overlooked. To address the issue of multi‐fault detection in PV panel imagery, this study proposes an innovative algorithm based on You Only Look Once (YOLO)v8‐DA. The proposed algorithm first employs the CSPNeXt backbone network to replace YOLOv8's original backbone network, striking a balance between parameter scale and computational cost, thereby creating a faster and more efficient object detection mechanism. Secondly, it introduces the dynamic anchor assignment strategy (Dynamic ATSS) and the dynamic detection head technologies. These innovations resolve the issues of imprecise label assignment and fixed detection heads inherent in traditional methods by incorporating dynamic label assignment and adjusting detection heads during the prediction process. Finally, the study proposes combining 1 × 1 convolutional blocks with CloAttention, enhancing the model's ability to recognise complex defects while maintaining efficiency and practicality. Experimental results demonstrate that the model performs exceptionally well on the dataset, achieving a mean average precision of 95.1%, which is 3.3% higher than the original YOLOv8, effectively validating the model's efficacy.