文献类型：期刊文献

英文题名：Deep learning-based in-situ identification of coniferous wood components in heritage architectures of China

作者：Zheng, Chang[1,2,3] Jiao, Lichao[1,2,3] He, Tuo[1,2,3,4] Lu, Yang[1,2,3] Liu, Shoujia[1,2,3] Li, Tianxiao[1,2,3] Ye, Ruochen[1,5] Zhou, Haibin[1] Yin, Yafang[1,2,3]

第一作者：Zheng, Chang

通信作者：Jiao, LC[1];Yin, YF[1];Jiao, LC[2];Yin, YF[2];Jiao, LC[3];Yin, YF[3]

机构：[1]Chinese Acad Forestry, Res Inst Wood Ind, Beijing, Peoples R China;[2]Chinese Acad Forestry, Wood Collect, Beijing, Peoples R China;[3]Wood Specimen Resource Ctr Natl Forestry & Grassla, Beijing, Peoples R China;[4]Natl Forestry & Grassland Adm, Wildlife Conservat Monitoring Ctr, Beijing, Peoples R China;[5]Shanxi Prov Inst Ancient Architecture, Painted Sculpture & Mural Preservat, Taiyuan, Shanxi, Peoples R China

年份：2025

卷号：13

期号：1

外文期刊名：NPJ HERITAGE SCIENCE

收录：;Scopus(收录号：2-s2.0-105020935322);WOS:【SCI-EXPANDED(收录号:WOS:001604505900001)，A&HCI(收录号:WOS:001604505900001)】；

基金：This study was supported financially by the National Key Research and Development Program of China (Grant No. 2023YFF0906301), the National Science & Technology Fundamental Resources Investigation Program (Grant No. 2023FY101400), and the China Scholarship Council (CSC) (Grant No. 202403270010). The authors would like to express our gratitude to Professor Xiaomei Jiang of the Research Institute of wood Industry, the Chinese Academy of Forestry for her valuable academic advice, and Professor Yongping Chen, Professor Juan Guo, Mrs. Mingkun Xu, Mr. Yonggang Zhang and Mr. Yu Sun of the Research Institute of wood Industry, the Chinese Academy of Forestry for their technical supports. The academic advice from Professor Jianjun Mei, Professor Geoffrey Lloyd, Dr. Sally Church and Mr. John Moffett of the Needham Research Institute, Cambridge of the United Kingdom is gratefully acknowledged.

语种：英文

摘要：Identification of structural wooden components is crucial for heritage architecture conservation as it elucidates the utilization patterns of forest resources and evolution of human civilization. This paper proposes a computer vision-based in situ identification method for wooden components of Chinese heritage architectures, using a dataset comprising 4050 images from 63 components of nine buildings. The optimal algorithm, RepLKNet, trained on coniferous xylarium specimens, achieves 96.67% identification accuracy, with 98.33%, 93.33%, and 90% precision at 50%, 70%, and 90% confidence thresholds, respectively. A minimum sample size of 25 species and 1500 images per genus ensures test accuracy >90%. Impact of structural deterioration (decay and cracks) on accuracy is also evaluated. Cracks significantly affect the wood recognition accuracy of historical components. Performance degrades significantly when cracks span >30% of the image. Latewood integrity is also critical to identification. The proposed method advances structural preservation strategies and preventive maintenance practices in heritage architecture.