[论文解读] HybridSolarNet: A Lightweight and Explainable EfficientNet-CBAM Architecture for Real-Time Solar Panel Fault Detection

Md. Asif Hossain, G M Mota-Tahrin Tayef+1 more|arXiv (Cornell University)|Jan 6, 2026

Photovoltaic System Optimization Techniques被引用 0

一句话总结

HybridSolarNet combines EfficientNet-B0 with CBAM to create a lightweight, explainable model for real-time solar panel fault detection on UAVs, achieving strong accuracy with low memory usage.

ABSTRACT

Manual inspections for solar panel systems are a tedious, costly, and error-prone task, making it desirable for Unmanned Aerial Vehicle (UAV) based monitoring. Though deep learning models have excellent fault detection capabilities, almost all methods either are too large and heavy for edge computing devices or involve biased estimation of accuracy due to ineffective learning techniques. We propose a new solar panel fault detection model called HybridSolarNet. It integrates EfficientNet-B0 with Convolutional Block Attention Module (CBAM). We implemented it on the Kaggle Solar Panel Images competition dataset with a tight split-before-augmentation protocol. It avoids leakage in accuracy estimation. We introduced focal loss and cosine annealing. Ablation analysis validates that accuracy boosts due to added benefits from CBAM (+1.53%) and that there are benefits from recognition of classes with imbalanced samples via focal loss. Overall average accuracy on 5-fold stratified cross-validation experiments on the given competition dataset topped 92.37% +/- 0.41 and an F1-score of 0.9226 +/- 0.39 compared to baselines like VGG19, requiring merely 16.3 MB storage, i.e., 32 times less. Its inference speed measured at 54.9 FPS with GPU support makes it a successful candidate for real-time UAV implementation. Moreover, visualization obtained from Grad-CAM illustrates that HybridSolarNet focuses on actual locations instead of irrelevant ones.

研究动机与目标

• 解决在边缘设备和无人机上实现高精度太阳能板故障检测的需求。
• 开发一种轻量级、带注意力机制的适用于实时检查的架构。
• 通过 split-before-augmentation 的严格评估防止数据泄漏，确保鲁棒评估。

提出的方法

• Backbone: EfficientNet-B0 pretrained on ImageNet to balance accuracy and efficiency.
• Attention: Insert CBAM after the backbone to refine feature maps via channel and spatial attention.
• Classifier head: Global average pooling, dropout (p=0.4), and a linear layer for six fault classes.
• Training strategies: Focal loss to handle class imbalance and cosine annealing scheduler over 25 epochs.
• Data handling: Split-before-augmentation with 70/15/15 train/val/test, preprocessing to 380x380 and ImageNet normalization.

实验结果

研究问题

• RQ1 能否用带注意力机制的轻量级主干在边缘设备上实现与太阳能板故障检测相关的有竞争力的准确率？
• RQ2 与仅使用 EfficientNet-B0 相比，CBAM 是否提升了缺陷定位和整体分类指标？
• RQ3 焦点损失和余弦退火是否有助于在不平衡的太阳能故障类别中获得更好性能？
• RQ4 在速度和内存占用方面，该模型是否适合在无人机上实现实时部署？

主要发现

模型	Acc.	F1-Score	FPS	Size (MB)	table_rows([["Hybrid (Ours)", "92.37%", "0.9226", "54.9", "16.3"], ["EfficientNet-B0", "90.84%", "0.9072", "57.8", "15.5"], ["VGG19", "87.79%", "0.8780", "39.9", "532.6"], ["MobileNetV3", "86.26%", "0.8593", "59.0", "16.2"], ["ResNet50", "83.97%", "0.8391", "43.6", "89.9"], ["Custom CNN", "78.63%", "0.7853", "56.5", "5.0"]]
Hybrid (Ours)	92.37%	0.9226	54.9	16.3
EfficientNet-B0	90.84%	0.9072	57.8	15.5
VGG19	87.79%	0.8780	39.9	532.6
MobileNetV3	86.26%	0.8593	59.0	16.2
ResNet50	83.97%	0.8391	43.6	89.9
Custom CNN	78.63%	0.7853	56.5	5.0

• HybridSolarNet 在留出测试集上达到 92.37% 的准确率和 0.9226 的 F1 分数。
• CBAM 相较 EfficientNet-B0 提升了 1.53% 的准确率和 1.54% 的 F1。
• 焦点损失提高了少数类的性能和整体 F1 得分。
• 余弦退火在收敛稳定性方面优于固定学习率。
• HybridSolarNet 比 VGG19 小 32 倍（16.3 MB vs. 532.6 MB），GPU 上的运行速度为 54.9 FPS。
• 交叉验证显示平均准确率为 92.37% ± 0.41，平均 F1 为 0.9226 ± 0.39。
• Grad-CAM 可视化表明模型关注的是实际缺陷而非伪影。