vertical-signage
收藏资源简介:
Vertical Signage 是一个用于垂直交通标志检测与质量分类的公开数据集。数据集包含两个独立任务:一是基于边界框的检测任务,采用YOLO格式,分为训练、验证和测试集,检测类别包括四种葡萄牙语标识类型:regulamentacao(禁令标志)、advertencia(警告标志)、indicacao(指示标志)和percurso(路线标志);二是质量分类任务,将标志图像分为“boa”(好)和“ruim”(差)两类,数据同样按训练、验证和测试集划分。此外,数据集中还提供了原始标注文件(JSON格式),存放于annotations/json_abril/目录下。该数据集适用于交通标志检测、分类及质量评估等计算机视觉任务。
Vertical Signage is a publicly available dataset for vertical traffic sign detection and quality classification. The dataset comprises two independent tasks: the first is the bounding box-based detection task, which adopts the YOLO format and is split into training, validation and test sets, with detection categories covering four Portuguese sign types: regulamentacao (prohibition signs), advertencia (warning signs), indicacao (indicative signs) and percurso (route signs); the second is the quality classification task, which categorizes sign images into two classes: "boa" (good) and "ruim" (bad), and the data is also divided into training, validation and test sets. In addition, the raw annotation files in JSON format are provided in the dataset and stored under the directory annotations/json_abril/. This dataset is applicable to computer vision tasks such as traffic sign detection, classification and quality assessment.
数据集概述
Vertical Signage 是一个专注于垂直标牌(señalización vertical)的数据集,包含两个独立的任务:目标检测(bounding boxes)和质量分类(calidad)。
目录结构
数据集的根目录下包含以下主要部分:
detection/:用于目标检测的YOLO格式数据集,包含图像、标签和数据集划分文件。quality/:质量分类数据,包含boa(好)和ruim(差)两个类别,并按train、val、test进行划分。annotations/json_abril/:原始标注文件,以JSON格式存储(名为json_abril)。
任务一:边界框检测
检测数据位于detection/目录下,具体结构如下:
detection/images/:图像文件。detection/labels/:标签文件(YOLO格式)。detection/splits/train.txt、detection/splits/val.txt、detection/splits/test.txt:训练、验证和测试集的划分文件。detection/dataset.yaml:数据集配置文件。
检测类别包括4类:
regulamentacao(法规)advertencia(警告)indicacao(指示)percurso(路径/路线)
任务二:质量分类
质量分类数据位于quality/目录下,分为两个类别:
boa(好)ruim(差)
数据同样按train、val、test进行了划分。
原始标注
原始标注文件存放在annotations/json_abril/目录下,以JSON格式提供。





