Abstract: Due to limitations such as interference among various apparent diseases in concrete bridges, there is a problem of insufficient exploration of deep contextual information in the network, this paper proposes a Concrete Bridge Surface Multi-Defect Semantic Segmentation Network based on Three-Branch Architecture (TBNet). TBNet comprises three-branch encoder, skip connection and decoder. The three-branch encoder decouples feature information, consisting of progressively hierarchical context branch, "wide and large" detail branche, and "narrow and small" semantic branche. Each branch is responsible for extracting contextual information, detailed information, and semantic information, respectively. The Guidance Aggregation Module (GAM) guides the fusion of contextual information through semantic and detailed information, enhancing the deep contextual information mining capability of the network. Simultaneously, the Biformer attention mechanism module reduces redundancy in the encoder's output feature information, further improving segmentation performance. Experimental results demonstrate that TBNet outperforms the baseline UNet on the concrete bridge surface multi-defect semantic segmentation dataset, with an average pixel accuracy (mPA) improvement of 5.92% and an average intersection over union (mIoU) improvement of 6.65%. Additionally, on the public dataset GAPs384, the intersection over union (IoU) of TBNet is 3.46% higher than that of the baseline UNet.

Key words: Bridge Disease Segmentation, Three-Branch Encoder, Feature Information Decoupling, Context Information