A Semantically Annotated 15-Class Ground Truth Dataset for Substation Equipment

This dataset contains 1660 images of electric substations with 50705 annotated objects. The images were obtained using different cameras, including cameras mounted on Autonomous Guided Vehicles (AGVs), fixed location cameras and those captured by humans using a variety of cameras. A total of 15 classes of objects were identified in this dataset. These are the classes and their number of instances:<br> <strong>Open blade disconnect switch:</strong> 310 <strong>Closed blade disconnect switch:</strong> 5243 <strong>Open tandem disconnect switch:</strong> 1599 <strong>Closed tandem disconnect switch: </strong>966 <strong>Breaker: </strong>980 <strong>Fuse disconnect switch:</strong> 355 <strong>Glass disc insulator:</strong> 3185 <strong>Porcelain pin insulator:</strong> 26499 <strong>Muffle:</strong> 1354 <strong>Lightning arrester:</strong> 1976 <strong>Recloser: </strong>2331 <strong>Power transformer:</strong> 768 <strong>Current transformer: </strong>2136 <strong>Potential transformer:</strong> 654 <strong>Tripolar disconnect switch: </strong>2349 <br> All images in this dataset were collected from a single electrical distribution substation in Brazil over a period of two years. The images were captured at various times of the day and under different weather and seasonal conditions, ensuring a diverse range of lighting conditions for the depicted objects. A team of experts in Electrical Engineering curated all the images to ensure that the angles and distances depicted in the images are suitable for automating inspections in an electrical substation. <br> The file structure of this dataset contains the following directories and files:<br> <strong>images:</strong> This directory contains 1660 electrical substation images in JPEG format. <strong>labels_json:</strong> This directory contains JSON files annotated in the VOC-style polygonal format. Each file shares the same filename as its respective image in the images directory. <strong>15_masks: </strong>This directory contains PNG segmentation masks for all 15 classes, including the porcelain pin insulator class. Each file shares the same name as its corresponding image in the images directory. <strong>14_masks:</strong> This directory contains PNG segmentation masks for all classes except the porcelain pin insulator. Each file shares the same name as its corresponding image in the images directory. <strong>porcelain_masks:</strong> This directory contains PNG segmentation masks for the porcelain pin insulator class. Each file shares the same name as its corresponding image in the images directory. <strong>classes.txt: </strong>This text file lists the 15 classes plus the background class used in LabelMe. <strong>json2png.py:</strong> This Python script can be used to generate segmentation masks using the VOC-style polygonal JSON annotations. <br> The dataset aims to support the development of computer vision techniques and deep learning algorithms for automating the inspection process of electrical substations. We expect it to be useful for researchers, practitioners, and engineers interested in developing and testing object detection and segmentation models for automating inspection and maintenance activities in electrical substations.

本数据集包含1660张变电站图像，共计标注有50705个目标对象。这些图像通过多种设备采集，包括搭载于自动导引车（Autonomous Guided Vehicles, AGVs）上的相机、固定位置相机，以及人类使用各类相机拍摄的画面。本数据集共识别出15类目标对象，各类别及其实例数量如下：
**分闸式闸刀隔离开关（Open blade disconnect switch）：** 310
**合闸式闸刀隔离开关（Closed blade disconnect switch）：** 5243
**分闸式双联隔离开关（Open tandem disconnect switch）：** 1599
**合闸式双联隔离开关（Closed tandem disconnect switch）：** 966
**断路器（Breaker）：** 980
**熔断器式隔离开关（Fuse disconnect switch）：** 355
**玻璃盘形绝缘子（Glass disc insulator）：** 3185
**瓷针式绝缘子（Porcelain pin insulator）：** 26499
**消声罩（Muffle）：** 1354
**避雷器（Lightning arrester）：** 1976
**重合器（Recloser）：** 2331
**电力变压器（Power transformer）：** 768
**电流互感器（Current transformer）：** 2136
**电压互感器（Potential transformer）：** 654
**三极隔离开关（Tripolar disconnect switch）：** 2349

本数据集所有图像均采集自巴西某一座配电变电站，采集周期长达两年。图像拍摄于一天中的不同时段，涵盖不同天气与季节条件，确保了目标对象的光照环境具备丰富多样性。电气工程领域的专家团队对所有图像进行了审核标注，以确保图像中的拍摄角度与拍摄距离适配变电站巡检自动化的实际应用需求。

本数据集的文件结构包含以下目录与文件：
**images：** 该目录包含1660张JPEG格式的变电站图像。
**labels_json：** 该目录包含采用VOC风格多边形标注格式的JSON标注文件，每个文件与images目录中对应的图像文件同名。
**15_masks：** 该目录包含全部15类目标的PNG格式分割掩码，其中包含瓷针式绝缘子类别。每个文件与images目录中对应的图像文件同名。
**14_masks：** 该目录包含除瓷针式绝缘子外其余14类目标的PNG格式分割掩码，每个文件与images目录中对应的图像文件同名。
**porcelain_masks：** 该目录包含瓷针式绝缘子类别的PNG格式分割掩码，每个文件与images目录中对应的图像文件同名。
**classes.txt：** 该文本文件列出了LabelMe标注工具中使用的15个目标类别以及背景类别。
**json2png.py：** 该Python脚本可用于基于VOC风格多边形JSON标注文件生成分割掩码。

本数据集旨在支撑面向变电站巡检自动化的计算机视觉技术与深度学习算法研发。我们期望该数据集能够为致力于开发并测试用于变电站巡检与运维自动化的目标检测、分割模型的研究人员、从业者与工程师提供切实帮助。