Poribohon-BD

Poribohon-BD is a vehicle dataset of 15 native vehicles of Bangladesh. The vehicles are: i) Bicycle, ii) Boat, iii) Bus, iv) Car, v) CNG, vi) Easy-bike, vii) Horse-cart, viii) Launch, ix) Leguna, x) Motorbike, xi) Rickshaw, xii) Tractor, xiii) Truck, xiv) Van, xv) Wheelbarrow. The dataset contains a total of 9058 images with a high diversity of poses, angles, lighting conditions, weather conditions, backgrounds. All of the images are in JPG format. The dataset also contains 9058 image annotation files. These files state the exact positions of the objects with labels in the corresponding image. The annotation has been performed manually and the annotated values are stored in XML files. LabelImg tool by Tzuta Lin has been used to label the images. Moreover, data augmentation techniques have been applied to keep the number of images comparable to each type of vehicle. Human faces have also been blurred to maintain privacy and confidentiality. The data files are divided into 15 individual folders. Each folder contains images and annotation files of one vehicle type. The 16th folder titled ‘Multi-class Vehicles’ contains images and annotation files of different types of vehicles. Poribohon-BD is compatible with various CNN architectures such as YOLO, VGG-16, R-CNN, DPM.

Poribohon-BD乃一套源自孟加拉国的交通工具数据集，汇集了该国15种本土车辆，包括自行车、船只、公交车、汽车、压缩天然气车、简易自行车、马车、快艇、勒古纳、摩托车、人力车、拖拉机、卡车、面包车及手推车。该数据集包含总计9058张图像，涵盖了丰富的姿态、角度、光照条件、天气状况以及背景。所有图像均以JPG格式呈现。此外，数据集还附带了9058个图像标注文件，这些文件详细记录了对应图像中物体的精确位置及标签。标注过程系人工完成，标注数据以XML格式存储。图像标注过程中采用了Tzuta Lin开发的LabelImg工具。为保持各类车辆图像数量相当，数据集还应用了数据增强技术。同时，出于隐私保护，人脸图像已被模糊处理。数据文件被划分为15个独立文件夹，每个文件夹包含某一类型车辆的图像及标注文件。第16个文件夹，题为‘多类别车辆’，汇集了不同类型车辆的图像及标注文件。Poribohon-BD与多种卷积神经网络架构兼容，例如YOLO、VGG-16、R-CNN及DPM。