DFSR: Super‐Resolution of Transmission Line Defective Fittings Image Using Stable Diffusion

ABSTRACT The low resolution and blurred details of transmission line inspection images lead to suboptimal defect detection performance. To address this issue, this paper proposes a defective fittings super‐resolution (DFSR) method based on stable diffusion. First, DFSR introduces a multi‐fittings low‐rank adaptation module to incorporate various fittings concepts into the stable diffusion model and fine‐tune it, enabling it to learn different fittings concepts effectively. Then, the conditional constraints module is designed, including an edge‐guided structural constraint and a histogram‐based colour constraint, to optimize structural reconstruction and colour consistency during the super‐resolution process, thereby improving the overall image quality and visual performance. Experimental results demonstrate that DFSR achieves high‐quality super‐resolution for fittings and outperforms existing methods across multiple reference and no‐reference metrics. Specifically, it improves PSNR and MUSIQ by 3.46 dB and 12.29, respectively, over the baseline model. Furthermore, a localized super‐resolution enhancement strategy is proposed to enhance fittings defect detection by performing super‐resolution on defective regions in inspection images. Its effectiveness was validated on the YOLOv11 model, achieving a improvement in , which further proves its practical applicability in real‐world transmission line inspection.