High-resolution (HR) crack images provide detailed structural assessments crucial for maintenance planning. However, the discrete nature of feature extraction in mainstream deep learning algorithms and computational limitations hinder refined segmentation. This study introduces a rendering-based lightweight crack segmentation network (RLCSN) designed to efficiently predict refined masks for HR crack images. The RLCSN combines a deep semantic feature extraction architecture—merging Transformer with a super-resolution boundary-guided branch—to reduce environmental noise and preserve crack edge details. It also incorporates customized point-wise refined rendering for training and inference, focusing computational resources on critical areas, and an efficient sparse training method to ensure efficient inference on commercial mobile computing platforms. Each RLCSN's components are validated through ablation studies and field tests, demonstrating its capability to enable unmanned aerial vehicle-based inspections to detect cracks as narrow as 0.15 mm from a distance of 3 m, thereby enhancing inspection safety and efficiency.

中文翻译：

---

用于分割高分辨率裂纹图像的基于渲染的轻量级网络


高分辨率 (HR) 裂纹图像提供对维护计划至关重要的详细结构评估。然而，主流深度学习算法中特征提取的离散性和计算限制阻碍了精细分割。本研究引入了一种基于渲染的轻量级裂纹分割网络 (RLCSN)，旨在有效预测 HR 裂纹图像的精细掩模。 RLCSN 结合了深度语义特征提取架构（将 Transformer 与超分辨率边界引导分支合并），以减少环境噪声并保留裂纹边缘细节。它还结合了用于训练和推理的定制逐点精细渲染，将计算资源集中在关键区域，以及高效的稀疏训练方法，以确保商业移动计算平台上的高效推理。每个 RLCSN 的组件均通过烧蚀研究和现场测试进行验证，证明其能够实现无人机检查，在 3 m 距离内检测窄至 0.15 mm 的裂纹，从而提高检查安全性和效率。