火灾爆炸事故典型痕迹数据集

针对“火灾爆炸痕迹智能识别技术及装备”课题的关键指标：来自于模拟实验与真实案件的建筑火灾、汽车火灾、森林火灾和爆炸痕迹图痕，共计不少于2万张，用于痕迹识别研究中的机器学习。2022年12月至2025年7月期间，搭建了建（构）筑物、汽车、森林、爆炸场景火灾爆炸模拟实验平台，完成了实验室条件下火灾爆炸事故中热蚀、变形、烧失等典型痕迹样品的制备与图痕数据采集，于此同时，整理近十年内，我国火灾爆炸事故中各类典型痕迹图痕照片，挑选痕迹分辨率高、色彩及白平衡准确的痕迹图片与实验室条件下采集特征痕迹图片共同组成火灾爆炸事故典型痕迹数据集，相关数据满足课题中火灾爆炸典型痕迹智能识别系统开发过程中模型训练的需求。

Key indicators for the project titled "Intelligent Recognition Technology and Equipment for Fire and Explosion Traces": The dataset contains no fewer than 20,000 trace images covering building fires, vehicle fires, forest fires and explosion traces from both simulated experiments and actual cases, which is dedicated to machine learning-based trace recognition research. From December 2022 to July 2025, a simulated fire and explosion experiment platform for buildings and structures, vehicles, forests and explosion scenarios was constructed. During this period, the preparation of typical trace samples including thermal erosion, deformation and burn loss from fire and explosion accidents under laboratory conditions and the collection of corresponding trace image data were completed. Meanwhile, photos of various typical traces from fire and explosion accidents in China over the past decade were sorted out. The trace images with high resolution, accurate color reproduction and proper white balance were selected, and combined with the characteristic trace images collected under laboratory conditions to form the typical trace dataset for fire and explosion accidents. The relevant data meets the requirements of model training during the development of the intelligent recognition system for typical fire and explosion traces in this project.

搜集汇总

背景与挑战

背景概述

该数据集包含超过2万张火灾爆炸事故的典型痕迹图痕，来源于模拟实验和真实案件，覆盖建筑、汽车、森林及爆炸场景。数据采集于2022年12月至2025年7月，整合了实验室制备痕迹和历史图片，用于机器学习研究，以支持火灾爆炸痕迹智能识别系统的开发。