Data files for Sheehan et al. 2023 'City Scale Traffic Monitoring Using WorldView Satellite Imagery and Deep Learning: A Case Study of Barcelona' DOI: https://doi.org/10.3390/rs15245709

Data files for Sheehan et al. (2023) City Scale Traffic Monitoring Using WorldView Satellite Imagery and Deep Learning: A Case Study of Barcelona. Remote Sensing. 15(24) DOI: https://doi.org/10.3390/rs15245709

Description of contents: 

xView-YOLOv3_Model6_Barcelona_weights.pt

This file contains the pre-trained weights for the xView-YOLOv3 model (model code available here: https://github.com/ultralytics/xview-yolov3). These weights were trained on a manually created training data set of vehicles present in WorldView 2/3 imagery covering the city of Barcelona. The weights relate to Model 6 set up: a single vehicle class (parked, static and moving), RGB imagery, Barcelona training data set derived anchor boxes, 1500 x 1500 pixel sized images and to 1000 epochs.

本数据集对应Sheehan等人2023年发表于《遥感》(Remote Sensing)期刊的研究《基于WorldView卫星影像与深度学习的城市级交通监测：以巴塞罗那为例》，论文卷期为15(24)，DOI：https://doi.org/10.3390/rs15245709

数据集内容说明：
xView-YOLOv3_Model6_Barcelona_weights.pt
该文件为xView-YOLOv3模型的预训练权重文件，模型代码可通过以下链接获取：https://github.com/ultralytics/xview-yolov3。该权重基于人工构建的训练数据集训练得到，数据集涵盖覆盖巴塞罗那市的WorldView 2/3卫星影像中的车辆目标。该权重对应第6号模型配置：仅包含单一车辆类别（含停放、静止及行驶状态的车辆），输入为RGB影像，使用基于巴塞罗那训练数据集生成的锚框，训练图像尺寸为1500×1500像素，训练轮次为1000轮。