面向地下结构的渗漏病害标记数据

该数据记录了地下轨道交通的隧道渗漏水图像，是用于针对渗漏性态辨识模型的测评数据，记录了地铁11号线的东体三林下行、东体龙华上行、东体三林上行、龙华东体下行等路段的疑似渗漏数据。文件格式为图像，每张图像记录了一处隧道的局部状态信息，可通过对图像的分析来判断该隧道局部是否发生渗漏水异常。通过对原始图像进行人工标注，可以得到标注图像数据。具体来说，标注过程是针对图像上的每个像素点，人工进行判断该像素点是否发生渗漏水异常，如果发生，则将该像素点的值标注为1，否则标注为0，从而得到一个由0和1组成的标注图像。标注图像和原始图像具有一一对应的关系，在模型训练时标注图像作为原始图像的标签参与训练。数据量878．00 MB。

This dataset collects water leakage images of metro tunnels, serving as evaluation data for leakage behavior identification models. It gathers suspected leakage data from sections of Metro Line 11, namely Dongti-Sanlin Downbound, Dongti-Longhua Upbound, Dongti-Sanlin Upbound, and Longhua-Dongti Downbound sections. All data are stored in image format, where each image records the local state of a specific tunnel section, and whether there is water leakage abnormality at the corresponding local section can be determined via image analysis. Annotated image data can be obtained by performing manual annotation on the original images. Specifically, during the annotation process, each pixel in the image is manually inspected to judge whether water leakage abnormality occurs at that pixel: if abnormality exists, the pixel value is labeled as 1; otherwise, it is labeled as 0, thus generating an annotated image composed of 0s and 1s. The annotated images and the original images have a one-to-one correspondence, and the annotated images are used as ground-truth labels for the original images during model training. The total size of this dataset is 878.00 MB.

搜集汇总

背景与挑战

背景概述

该数据集是一个用于地下结构渗漏病害检测的图像标注数据集，包含地铁隧道渗漏水图像及其像素级人工标注，标注以二值图像形式标识渗漏异常区域。数据源自地铁11号线多个路段，专门用于训练和评估渗漏性态辨识模型，支持地下基础设施安全监控研究。