Massachusetts Roads Dataset

该数据集包含了马萨诸塞州道路网络的像素化图像，旨在用于检测交叉点和急转弯处的奇异点。这些图像代表了图的平面嵌入，其中奇异点即是交叉点和角落。每张图像的分辨率为1500×1500像素，包含45,000至200,000个具有非零亮度值的像素。该数据集的任务是进行奇异点检测。

This dataset contains pixelated images of the road network in Massachusetts, designed for detecting singularities at intersections and sharp turns. These images represent planar embeddings of graphs, where singularities correspond to intersections and corners. Each image has a resolution of 1500×1500 pixels and contains between 45,000 and 200,000 pixels with non-zero luminance values. The task associated with this dataset is singularity detection.

搜集汇总

背景与挑战

背景概述

Massachusetts Roads Dataset是一个包含训练集、验证集和测试集的道路检测数据集，提供输入图像、目标地图及生成目标地图的shapefile向量数据，适用于机器学习研究。