基于YOLOv8实例分割的桥梁裂纹识别

Bridge Crack Recognition Based on YOLOv8 Instance Segmentation

摘要: 桥梁结构的安全性是工程领域的重要关注点，裂纹检测是保障桥梁安全的关键环节。本文提出了一种基于YOLOv8实例分割的桥梁裂纹识别方法，旨在解决传统检测方法在准确性上不足的问题。首先，我们改进了数据增强方法，增强了模型对不同环境下裂纹图像的鲁棒性和泛化能力，使得模型在复杂实际应用中表现更加稳定。其次，在C2F模块中引入了Coordinate Attention注意力机制。该机制通过对特征图的通道关系进行建模，提升了模型对关键信息的关注度，从而提高了裂纹检测的准确性。Coordinate Attention在空间和通道维度上融合信息，使模型能够更精准地捕捉到裂纹的细节特征。实验结果表明，本文提出的方法在本文构造的数据集上表现出色，在精度上提高了3.5个百分点，并且召回率和mAP均有提升，对实际工程应用具有重要指导意义。

Abstract: The safety of bridge structures is an important concern in the engineering field, and crack detection is a key link to ensure the safety of bridges. In this paper, we propose a bridge crack recognition method based on YOLOv8 instance segmentation, aiming at solving the problem of insufficient accuracy of traditional detection methods. First, we improve the data enhancement method to enhance the robustness and generalization ability of the model to crack images in different environments, which makes the model perform more stably in complex practical applications. Second, the Coordinate Attention attention mechanism is introduced in the C2F module. This mechanism improves the accuracy of crack detection by modeling the channel relationship of the feature map, which enhances the model’s attention to the key information. Coordinate Attention fuses the information in both spatial and channel dimensions, enabling the model to capture the detailed features of the cracks more accurately. The experimental results show that the method proposed in this paper performs well on several bridge crack datasets, with an improvement of 3.5 percentage points in precision and both recall and mAP, which is of great significance as a guide for practical engineering applications.