OTTB-IQA: A No-Reference Quality Assessment Method for Overhead Transmission Tower Bolt Images

Le résumé fourni par la source

Regular inspection of the safety of the bolt components on overhead transmission towers is fundamental to ensuring the safe and stable operation of power systems. However, during unmanned aerial vehicle (UAV) inspection, factors such as posture changes and lens defocus often lead to distortion in the captured image quality, affecting image recognizability and the accuracy of downstream detection tasks. To address this problem, this paper proposes a no-reference quality assessment method for overhead transmission tower bolt images (OTTB-IQA). Distortion features at multiple scales are extracted from the image, using the Swin Transformer as the encoder. The multi-scale feature multiplicative fusion (MSF-MF) module is employed to integrate features of different scales, thereby enhancing the completeness and robustness of the quality-aware features. The channel-spatial attention hybrid (CSAH) module is incorporated to enable the network to focus on critical information. Experimental results demonstrate that the proposed method achieves high accuracy on both the self-constructed bolt dataset and publicly available datasets, showing a high correlation with human visual perception scores.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.