BadODD: Bangladeshi Autonomous Driving Object Detection Dataset
Mirza Nihal Baig, Rony Hajong, Mahdi Murshed Patwary, Mohammad Shahidur Rahman, Husne Ara Chowdhury
TL;DR
The paper addresses the lack of region-specific data for autonomous driving in Bangladesh by introducing BDOR, a 9,825-image dataset collected from smartphone cameras across 9 districts under day and night conditions. It defines a characteristic-based 13-class vehicle taxonomy to accommodate locally manufactured and unnamed vehicles, and provides a frame-sampling strategy to capture diverse traffic scenarios. Two YOLO variants, YOLOv5 and YOLOv8, are benchmarked with identical hyperparameters, yielding mAPs of 0.60 and 0.70 respectively, thereby highlighting the potential gains from more capable detectors in unstructured traffic. The dataset and benchmarks are presented as a resource to drive robust, locally relevant autonomous navigation research and to enable future expansion for new vehicle types and conditions.
Abstract
We propose a comprehensive dataset for object detection in diverse driving environments across 9 districts in Bangladesh. The dataset, collected exclusively from smartphone cameras, provided a realistic representation of real-world scenarios, including day and night conditions. Most existing datasets lack suitable classes for autonomous navigation on Bangladeshi roads, making it challenging for researchers to develop models that can handle the intricacies of road scenarios. To address this issue, the authors proposed a new set of classes based on characteristics rather than local vehicle names. The dataset aims to encourage the development of models that can handle the unique challenges of Bangladeshi road scenarios for the effective deployment of autonomous vehicles. The dataset did not consist of any online images to simulate real-world conditions faced by autonomous vehicles. The classification of vehicles is challenging because of the diverse range of vehicles on Bangladeshi roads, including those not found elsewhere in the world. The proposed classification system is scalable and can accommodate future vehicles, making it a valuable resource for researchers in the autonomous vehicle sector.