针对电力设备监测红外图像分辨率低的问题，提出一种基于生成对抗网络的感知增强型电力设备红外图 像超分辨率模型，在生成器端构建双路径融合注意力架构，在原始分辨率与下采样分支并行提取多尺度依赖，结合 双层残差结构（residual in residual dense block，RRDB）密集残差块，实现全局结构与局部细节的协同重建；在判别器 端构建门控细粒度通道-空间（gated fine-grained channel-spatial，GFCS）判别器，在U-Net跳跃连接中嵌入细粒度通 道-空间注意力（fine-grained channel-spatial attention，FCSA）模块与注意力门控模块，通过并行建模全局-局部通道 依赖关系与方向敏感空间上下文，强化对设备发热区域关键特征的判别能力；实验中采用符合人眼感知的NIQE和 LPIPS指标，在电力设备伪彩色红外图像上开展的多倍率超分辨率实验 。结果表明，所提模型的NIQE和LPIPS指 标分别达到5.568和0.054 1，消融实验进一步验证了各个模块的有效性。
[1] Ledig C，Theis L，Huszar F，et al.Photo-realistic single image super-resolution using a generative adversarial network[C]// Proceedings of the IEEE/CVF international conference on computer vision and pattern recognition. Honolulu，HI，USA. IEEE/CVF.2017∶4681-4690.
[2] Wang X，Yu K，Wu S，et al.ESRGAN ∶Enhanced super-reso ‐ lution generative adversarial networks[C]//Proceedings of the European conference on computer vision workshops.Mu ‐ nich，Germany.ECVA. 2018∶63-79.
[3] Rakotonirina N C，Rasoanaivo A. ESRGAN+ ∶Further im ‐ proving enhanced super-resolution generative adversarial net‐ work[C] //2020 IEEE international conference on acoustics， speech and signal processing（ ICASSP）. Barcelona，Spain. IEEE.2020∶3637-3641.
[4] Ma C，Rao Y，Cheng Y，et al.Structure-preserving super reso ‐ lution with gradient guidance[C] // Proceedings of the IEEE/ CVF conference on computer vision and pattern recognition. Seattle，Washington，USA.IEEE/CVF.2020∶7769-7778.
[5] Zhang K，Liang J，Van Gool L，et al. Designing a practical degradation model for deep blind image super-resolution[C]
//Proceedings of the IEEE/CVF international conference on computer vision. Montreal，QC，Canada. IEEE/CVF. 2021 ∶ 4791-4800.
[6] Chen Z，Lu S. Caf-YOLO ∶A robust framework for multi- scale lesion detection in biomedical imagery[C] //2025 IEEE international cconference on acoustics，speech and signal pro ‐ cessing（ICASSP）.Hyderabad，India.IEEE.2025 ∶1-5.
[7] Yang J，Liu S，Wu J，et al. Pinwheel-shaped convolution and scale-based dynamic loss for infrared small target detection [C]//Proceedings of the AAAI Conference on Artificial Intel‐ ligence. Philadelphia，Pennsylvania，USA. AAAI. 2025，39
（09）∶9202-9210.
[8] Sun H，Wen Y，Feng H，et al.Unsupervised bidirectional con ‐ trastive reconstruction and adaptive fine-grained channel at‐ tention networks for image dehazing[J]. Neural Networks， 2024，176 ∶106314.
[9] Zhang Y，Li K，Li K，et al. Image super-resolution using very deep residual channel attention networks [C]//Proceedings of the European conference on computer vision（ECCV）.Munich， Germany，ECVA.2018∶286-301.
[10] Zhou Y，Li Z，Guo C L，et al. SRFormer ∶Permuted self-at‐ tention for single image super-resolution[C]//Proceedings of the IEEE/CVF international conference on computer vision. Paris，France.IEEE/CVF.2023 ∶12780-12791.
[11] Hsu C C，Lee CM，ChouYS.DRCT ∶Saving image super-reso ‐ lution away from information bottleneck[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Seattle，WA，USA.IEEE/CVF.2024∶6133-6142.
[12] Xia B，Zhang Y，Wang S，et al. DiffIR ∶Efficient diffusion model for image restoration[C]//Proceedings of the IEEE/ CVF international conference on computer vision. Paris， France.IEEE/CVF.2023 ∶13095-13105.
[13] Wang Y，Yang W，Chen X，et al. SinSR ∶diffusion-based im ‐ age super-resolution in a single step[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern rec‐ ognition.seattle，WA，USA.IEEE/CVF.2024∶25796-25805.
[14] Wang X，Xie L，Dong C，et al.Real-ESRGAN ∶Training real- world blind super-resolution with pure synthetic data[C]// Proceedings of the IEEE/CVF international conference on computer vision. Montreal，QC，Canada. IEEE/CVF. 2021 ∶ 1905-1914.
