矿山图像点云多模态检测数据集

该数据集为矿山图像检测分割数据集和点云多模态检测数据集，内含图像检测标注集1份，图像分割标注集 1 份，图像-点云多模态检测标注集 1 份。本数据集在鄂尔多斯黑岱沟露天煤矿、哈尔乌苏露天煤矿的矿卡真实作业条件下采集。采集设备为主控机Jetson Xavier NX、相机Basler工业相机、Ouster 雷达和 Hap 雷达等。采集过程中传感器数据帧率为10Hz，以 rosbag 的形式存储，采集后续进行数据使用时，以2Hz 关键帧抽取的形式建立数据集。This dataset is a multi-modal mine dataset covering image detection, segmentation and point cloud-based detection, which includes 1 set of image detection annotations, 1 set of image segmentation annotations, and 1 set of image-point cloud multi-modal detection annotations. This dataset was collected under the actual operating conditions of mining trucks at the Heidaigou Open-pit Coal Mine and Halwusu Open-pit Coal Mine in Ordos. The acquisition equipment consists of the main controller Jetson Xavier NX, Basler industrial cameras, Ouster LiDARs and Hap radars, among other devices. The sensor data was collected at a frame rate of 10 Hz and stored in rosbag format. When building the official dataset for subsequent data usage, key frames were extracted at 2 Hz to create the final dataset.

搜集汇总

背景与挑战

背景概述

该数据集为矿山图像与点云多模态检测数据集，包含图像检测、图像分割及图像-点云多模态检测三类标注集，数据采集自鄂尔多斯黑岱沟和哈尔乌苏露天煤矿的真实作业环境，使用工业相机、雷达等设备获取，并以关键帧抽取方式构建。数据集总量为9.14GB，涵盖51948个文件，适用于矿山机械工程领域的研究与应用。