基于深度学习法的建筑结构裂缝检测  

Crack detection in building structures based on deep learning method

作　　者：周旭文 Zhou Xuwen(Shaanxi Ancient Architecture and Landscape Construction Group Co.,Ltd.,Xi’an Shaanxi 710026,China)

机构地区：[1]陕西古建园林建设集团有限公司,陕西西安710026

出　　处：《山西建筑》2025年第8期53-57,共5页Shanxi Architecture

摘　　要：文中首先对长短期记忆神经网络(Long Short-Term Memory,LSTM)深度学习的原理以及计算流程进行了介绍。而后基于建筑结构裂缝监测结果,采用长短期记忆神经网络建立了建筑结构的裂缝监测模型,且与BP神经网络模型的结果进行了比较。结果表明:LSTM算法在建筑物结构裂缝长度预测检测中具有显著的优越性,为建筑物结构裂缝检测预测提供了一种新思路。This article first introduces the principle and calculation process of Long Short-Term Memory(LSTM)deep learning.Then,based on the monitoring results of building structural cracks,a Long Short-Term Memory neural network was used to establish a crack monitoring model for building structures,and the results were compared with those of the BP neural network model.The results showed that the LSTM algorithm in predicting and detecting the lenqth of cracks in building structures,providing a new approach for predicting and detecting cracks in building structures.

关 键 词：深度学习 LSTM神经网络 循环神经网络 建筑物裂缝 

分 类 号：TU312[建筑科学—结构工程]