智能网联汽车地图标注训练数据

产品包括2D图像和3D点云两个大类。其中2D图像的子类包括全车检测、车头车尾、车辆关键点、行人骑车人、车道线、骑车人关键点、全景语义分割、车轮、锥桶柱、交通灯、交通牌、车尾灯、地面标志；3D点云有3D目标检测和3D语义分割两个子类。

The dataset consists of two main categories: 2D images and 3D point clouds. The subcategories under 2D images include full-vehicle detection, front/rear vehicle detection, vehicle key points, pedestrians and cyclists, lane lines, cyclist key points, panoramic semantic segmentation, vehicle wheels, traffic cones and pylons, traffic lights, traffic signs, vehicle taillights, and ground markings. For 3D point clouds, there are two subcategories: 3D object detection and 3D semantic segmentation.

搜集汇总

背景与挑战

背景概述

该数据集包含智能网联汽车相关的2D图像和3D点云标注数据，涵盖车辆、行人、交通设施等多类目标的检测与分割任务。2D图像包含13个细分标注类别，3D点云包含目标检测和语义分割两类标注。