Bangladeshi Pothole Dataset
收藏资源简介:
Bangladeshi Pothole Dataset是由BRAC大学和联合国际大学的研究团队创建的一个专门用于坑洞检测的数据集。该数据集包含824张图像,主要采集自孟加拉国的达卡和博格拉街道。数据集经过预处理和增强,图像分辨率调整为128×128用于分类任务,256×256用于分割任务。数据增强方法包括翻转、随机亮度、对比度调整和旋转等,增强了数据的多样性和泛化能力。该数据集旨在通过深度学习模型提高坑洞检测的准确性,应用于道路安全和交通管理领域,以预防交通事故和减少经济损失。
The Bangladeshi Pothole Dataset is a specialized dataset for pothole detection, developed by a research team from BRAC University and United International University. This dataset comprises 824 images primarily collected from streets in Dhaka and Bogura, Bangladesh. The dataset has undergone preprocessing and augmentation, with image resolutions adjusted to 128×128 for classification tasks and 256×256 for segmentation tasks. Data augmentation methods include flipping, random brightness adjustment, contrast adjustment and rotation, which enhance the diversity and generalization capability of the dataset. This dataset aims to improve the accuracy of pothole detection via deep learning models, and is applied in the fields of road safety and traffic management to prevent traffic accidents and reduce economic losses.




