Construction Dataset
收藏arXiv2017-03-30 更新2024-08-06 收录
下载链接：
http://arxiv.org/abs/1703.10196v1
下载链接
链接失效反馈官方服务：
资源简介：
Construction Dataset是由Edward Boyda等人开发的一个小型基准数据集，用于检测加利福尼亚州2010年至2012年间的大规模住宅、工业和市政建设变化。该数据集包含100对手动选择和标记的图像，用于检测或确认变化。数据集的创建旨在通过KAZE特征和Poisson点过程模型，提高对人类干预景观变化的检测精度。该数据集主要用于评估算法在识别和区分不同类型建设活动中的性能，以及解决传统像素级变化检测方法的局限性。
The Construction Dataset is a small benchmark dataset developed by Edward Boyda et al. for detecting large-scale residential, industrial, and municipal construction changes in California between 2010 and 2012. This dataset contains 100 pairs of manually selected and annotated images used for detecting or validating changes. The creation of this dataset aims to improve the detection accuracy of landscape changes induced by human intervention via KAZE features and Poisson point process models. It is primarily used to evaluate algorithmic performance in identifying and distinguishing between different types of construction activities, as well as to address the limitations of traditional pixel-level change detection methods.
提供机构：
圣玛丽学院物理与天文学系
创建时间：
2017-03-30
