Steelcrack
收藏资源简介:
Steelcrack数据集是由东南大学土木工程学院创建,用于识别钢结构裂缝的标准化基准数据集。该数据集包含4355张512x512像素的钢结构裂缝图像,来源于多种实际工程项目,具有不同的背景和裂缝特征。数据集分为训练集、验证集和测试集,分别包含3300张、525张和530张图像。创建过程包括从原始图像中裁剪小片,并进行翻转和旋转等数据增强操作。该数据集旨在解决钢结构裂缝自动检测的问题,提高结构安全性和维护效率。
The Steelcrack dataset is a standardized benchmark dataset created by the School of Civil Engineering, Southeast University, for steel structural crack recognition. It comprises 4355 512×512 pixel steel structural crack images sourced from a range of real-world engineering projects, with varied backgrounds and crack features. The dataset is split into training, validation, and test sets, holding 3300, 525, and 530 images respectively. Its development workflow includes cropping small patches from raw images and performing data augmentation operations such as flipping and rotation. This dataset is intended to resolve the challenge of automated steel structural crack detection, thereby enhancing structural safety and maintenance efficiency.

- 1Infrastructure Crack Segmentation: Boundary Guidance Method and Benchmark Dataset东南大学土木工程学院 · 2023年



