BadODD
收藏资源简介:
BadODD数据集是由孟加拉国的沙贾拉尔科技大学计算机科学与工程系创建,专注于孟加拉国9个不同地区的自动驾驶环境中的目标检测。该数据集包含9825张图像,涵盖了78,943个对象,涉及13个类别,旨在解决现有数据集在孟加拉国道路环境中适用性不足的问题。数据集通过智能手机摄像头收集,确保了数据的真实性和多样性,包括日夜不同光照条件。此外,数据集提出了一种基于车辆特征而非本地名称的新分类系统,以适应孟加拉国道路上的多样化车辆,为自动驾驶技术的发展提供了重要资源。
The BadODD dataset was created by the Department of Computer Science and Engineering, Shahjalal University of Science and Technology, Bangladesh, focusing on object detection in autonomous driving environments across 9 distinct regions of Bangladesh. This dataset contains 9,825 images covering 78,943 objects spanning 13 categories, aiming to address the insufficient applicability of existing datasets in Bangladesh's road environments. The dataset is collected using smartphone cameras, ensuring data authenticity and diversity, including varying lighting conditions between day and night. Additionally, the dataset proposes a novel classification system based on vehicle features rather than local names to accommodate the diverse vehicles on Bangladesh's roads, providing a critical resource for the development of autonomous driving technologies.

- 1BadODD: Bangladeshi Autonomous Driving Object Detection Dataset计算机科学与工程系,沙贾拉尔科技大学,锡尔赫特,孟加拉国 · 2024年



