CSRM-MIM: A Self-Supervised Pretraining Method for Detecting Catenary Support Components in Electrified Railways

The core task of intelligent detection in electrified railway catenaries is to detect their supporting components. However, due to the large number and variety of components in catenary inspection images, labeled catenary data is often limited, and few studies have focused on leveraging large amounts of unlabeled datasets in this field. This paper proposes a novel self-supervised pre-training model (CSRM-MIM) for detecting catenary support components, which effectively utilizes the valuable information in unlabeled catenary data. Specifically, a new semantic masking strategy (CSRA masking) based on the catenary support rod area features is proposed to guide the model in learning meaningful catenary semantic information during the pre-training process. Additionally, a new Siamese pre-trained network framework is designed, incorporating an information interaction enhancement module (IIEM) and a dual reconstruction network (DIR) to perform dual reconstruction on the semantic mask, thereby extracting global features in the catenary domain. Finally, a multi-scale knowledge distillation strategy (MSKD) optimized by the cross-layer fusion module (CLFD) is introduced to assist the self-supervised model in acquiring specific-general representations for the catenary field and transferring multi-scale information, which benefits subsequent detection tasks. Experimental results demonstrate that this method significantly enhances the detector's performance in identifying catenary support components, owing to the retained self-supervised learning strategy.