印度车辆数据集

This dataset consists of images of niche Indian Vehicle such as Autorikshaw, Tempo, trucks, etc.IntroductionThe dataset consists of niche Indian vehicle images for classification and object detection. It is observed that there is very little or no dataset available on these niche vehicles like autorickshaw, tempo, trucks, etc. The images have been taken in varied weather conditions in daylight, evening, and night. The dataset has a wide variety of variations of illumination, distances, viewpoints, etc. This dataset represents a very challenging set of images of vehicles of niche classes. This dataset can be used for imge recognition and object detection for driver assistance systems, autonomous driving, etc.Vehicle ClassesIndian AutoIndian TruckBusTruckTempo TravellerTractorCarTwo WheelersDataset FeaturesApprox. 10,000+ unique images for each classCaptured by 3000+ unique usersCaptured with 50+ cities across IndiaCaptured using mobile phonesHighly diverseVarious lighting conditions like day, nightOutdoor scene with a variety of viewpointsDataset FormatClassification and detection annotations availableMultiple category annotations possibleCOCO, PASCAL VOC and YOLO formats

本数据集收录了各类印度小众车型的图像，例如自动三轮车（Autorikshaw）、Tempo、卡车等。

引言
本数据集旨在为分类与目标检测任务提供印度小众车辆图像数据。目前，针对自动三轮车（Autorikshaw）、Tempo、卡车这类小众车型的可用数据集极为稀缺，甚至近乎空白。所收录图像均在不同天气条件下的日间、傍晚及夜间拍摄，涵盖多样化的光照条件、拍摄距离与拍摄视角等变量，是一组极具挑战性的小众车型图像集。本数据集可应用于驾驶员辅助系统、自动驾驶等场景的图像识别与目标检测任务。

车辆类别
印度自动三轮车（Indian Auto）、印度卡车（Indian Truck）、巴士（Bus）、普通卡车（Truck）、Tempo Traveller、拖拉机（Tractor）、轿车（Car）、两轮车（Two Wheelers）

数据集特征
每个类别约含10000+张独特图像；由3000+名不同用户拍摄完成；采集自印度境内50余座城市；均通过移动设备拍摄；具备高度多样性；涵盖多种光照条件（如日间、夜间）；包含多样视角的户外拍摄场景。

数据集格式
提供分类与目标检测标注；支持多类别标注；包含COCO、PASCAL VOC及YOLO三种标注格式。