安检三维图像训练数据集

安检三维图像训练数据集主要面向民航智能安检领域的违禁品识别、模型轻量化及小样本持续学习等前沿研究 。该数据集是针对当前安检领域面临的三维数据计算复杂度高、新类别样本稀缺以及分布式节点数据分布不均等关键挑战而建设 。数据集基于工业级双能CT安检设备（同方威视Kylin系统）构建，数据集主要记录了分辨率为512×512×512的高维CT体素观测值（NIFTI格式），包含反映物体内部物理结构的密度信息与表征材料化学组成的原子序数信息 。内容覆盖了刀具、枪支、枪支部件、打火机、锂电池等5类指南规定违禁品，以及子弹、警棍、拳刺、斧子、金属喷雾等5类扩展违禁品，共计10个类别。数据量为5.8GB。该数据集旨在为推动三维违禁品识别算法在复杂安检场景中的实际部署提供全面、高质量的数据资源。

This 3D security inspection training dataset is primarily targeted at cutting-edge research in civil aviation intelligent security inspection, including prohibited item detection, model lightweighting, and few-shot continuous learning. This dataset was developed to address key challenges currently faced in the security inspection field, including high computational complexity of 3D data, scarcity of new category samples, and uneven data distribution across distributed nodes. Constructed based on industrial-grade dual-energy CT security inspection equipment (Tongfang Weishi Kylin system), the dataset mainly records high-dimensional CT voxel observations with a resolution of 512×512×512 in NIFTI format, which contain density information reflecting the internal physical structure of objects and atomic number information characterizing the chemical composition of materials. It covers a total of 10 categories: 5 types of prohibited items specified in civil aviation guidelines, including knives, firearms, firearm parts, lighters, and lithium batteries, plus 5 types of extended prohibited items such as bullets, batons, knuckle dusters, axes, and metal sprays. The total size of the dataset is 5.8 GB. This dataset aims to provide comprehensive and high-quality data resources to facilitate the practical deployment of 3D prohibited item detection algorithms in complex security inspection scenarios.

搜集汇总

背景与挑战

背景概述

安检三维图像训练数据集是一个面向民航智能安检领域的高维CT图像数据集，包含10类违禁品（如刀具、枪支等）的512×512×512分辨率体素观测值，数据量为5.8GB。该数据集旨在支持三维违禁品识别算法在复杂安检场景中的前沿研究与应用部署。