宁波木构文物红外无损检测机理与典型损伤特征综合数据集

本数据集主要面向宁波木构文物的红外无损检测机理研究与典型损伤智能识别需求建设，数据来源于宁波保国寺、天一阁等现场红外检测与巡检记录，结合课题组太阳能热激励红外检测原理样机实验资料形成。数据通过“现场红外热像/可见光对照采集—预处理与增强—文保专家交互标注—特征归纳整理”的流程产生，主要包含原始红外热像序列、现场照片与环境/测点信息，裂纹、空洞、腐朽等病害的红外特征样本及像素级真值标注，以及与检测机理相关的技术文档，用于支撑机理分析、算法训练与评测。数据总体规模约 3.8 GB，文件数约25000个。

This dataset is constructed to fulfill the requirements of researching the infrared non-destructive testing mechanism of wooden cultural relics in Ningbo and intelligent identification of their typical defects. The data is compiled from on-site infrared testing and inspection records of cultural relic sites including Ningbo Baoguo Temple and Tianyi Pavilion, together with experimental data from the solar thermal excitation infrared detection prototype developed by the research team. The dataset is generated via the standardized workflow: "on-site comparative collection of infrared thermography and visible light images, preprocessing and enhancement, interactive annotation by cultural relic protection experts, and feature induction and organization". It mainly contains original infrared thermography sequences, on-site photos, environmental and measuring point information, infrared feature samples of defects such as cracks, voids and decay with pixel-level ground truth annotations, as well as technical documents related to the detection mechanism. These data can be used to support mechanism analysis, algorithm training and evaluation. The total size of the dataset is approximately 3.8 GB, comprising about 25,000 files.

搜集汇总

背景与挑战

背景概述

该数据集面向宁波木构文物的红外无损检测机理研究与典型损伤智能识别，整合了保国寺、天一阁等现场红外检测记录与太阳能热激励实验资料。它包含红外热像序列、现场照片、病害特征样本及像素级标注，以及相关技术文档，总规模约3.8 GB，文件数约25000个，用于支撑机理分析和算法训练。