Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Large-scale curtain wall construction method and system based on LeNet-5
The invention discloses a large-scale curtain wall construction method and system based on LeNet-5. A convolutional layer C1 is introduced, the size of a filter is 5 * 5, the depth is 6, and the step length is 1; the design of the layer aims to extract features of the curtain wall structure through convolution operation so as to capture spatial information in the structure. For the convolutional layer C1, after S is input, an output feature map O is O = ReLU (S * K); k is a convolution kernel, and * represents convolution operation. According to the technical scheme, the LeNet-5 can learn and capture the complex nonlinear relation between curtain wall structures in a deep learning mode. Therefore, the system can more comprehensively understand the difference between different structures, and the expression ability of the model is improved. According to the technical scheme provided by the invention, the LeNet-5 can gradually extract and abstract the characteristics of the curtain wall structure through the multi-layer convolution and pooling layer, so that the complex relationship between the structures is more effectively expressed. The accuracy and generalization ability of the system can be improved.
本发明公开了一种基于LeNet‑5的大规模幕墙施工方法及系统；引入卷积层C1，其中过滤器大小为5x5，深度为6，步长为1。该层的设计旨在通过卷积操作提取幕墙结构的特征，以捕获结构中的空间信息。对于所述卷积层C1，输入S后，输出特征图O为：O＝ReLU(S*K)；K是卷积核，*表示卷积运算。本发明的技术方案中的LeNet‑5通过深度学习的方式能够学习和捕捉幕墙结构之间的复杂非线性关系。这使得系统能够更全面地理解不同结构之间的差异，提高了模型的表达能力。本发明的技术方案通过多层卷积和池化层，LeNet‑5能够逐渐提取和抽象出幕墙结构的特征，从而更有效地表示结构间的复杂关系。这有助于提高系统的准确性和泛化能力。
Large-scale curtain wall construction method and system based on LeNet-5
The invention discloses a large-scale curtain wall construction method and system based on LeNet-5. A convolutional layer C1 is introduced, the size of a filter is 5 * 5, the depth is 6, and the step length is 1; the design of the layer aims to extract features of the curtain wall structure through convolution operation so as to capture spatial information in the structure. For the convolutional layer C1, after S is input, an output feature map O is O = ReLU (S * K); k is a convolution kernel, and * represents convolution operation. According to the technical scheme, the LeNet-5 can learn and capture the complex nonlinear relation between curtain wall structures in a deep learning mode. Therefore, the system can more comprehensively understand the difference between different structures, and the expression ability of the model is improved. According to the technical scheme provided by the invention, the LeNet-5 can gradually extract and abstract the characteristics of the curtain wall structure through the multi-layer convolution and pooling layer, so that the complex relationship between the structures is more effectively expressed. The accuracy and generalization ability of the system can be improved.
本发明公开了一种基于LeNet‑5的大规模幕墙施工方法及系统；引入卷积层C1，其中过滤器大小为5x5，深度为6，步长为1。该层的设计旨在通过卷积操作提取幕墙结构的特征，以捕获结构中的空间信息。对于所述卷积层C1，输入S后，输出特征图O为：O＝ReLU(S*K)；K是卷积核，*表示卷积运算。本发明的技术方案中的LeNet‑5通过深度学习的方式能够学习和捕捉幕墙结构之间的复杂非线性关系。这使得系统能够更全面地理解不同结构之间的差异，提高了模型的表达能力。本发明的技术方案通过多层卷积和池化层，LeNet‑5能够逐渐提取和抽象出幕墙结构的特征，从而更有效地表示结构间的复杂关系。这有助于提高系统的准确性和泛化能力。
Large-scale curtain wall construction method and system based on LeNet-5
一种基于LeNet-5的大规模幕墙施工方法及系统
JIANG WUHENG (Autor:in)
16.04.2024
Patent
Elektronische Ressource
Chinesisch
IPC:
G06F
ELECTRIC DIGITAL DATA PROCESSING
,
Elektrische digitale Datenverarbeitung
/
E04B
Allgemeine Baukonstruktionen
,
GENERAL BUILDING CONSTRUCTIONS
/
E04G
SCAFFOLDING
,
Baugerüste
/
G06N
COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
,
Rechnersysteme, basierend auf spezifischen Rechenmodellen
CONSTRUCTION METHOD FOR CURTAIN WALL AND CURTAIN WALL
| Europäisches Patentamt | 2018
Large-span composite curtain wall keel construction system
| Europäisches Patentamt | 2024
Curtain wall fixing device and curtain wall construction method
| Europäisches Patentamt | 2022
Curtain wall fixing device and curtain wall construction method
| Europäisches Patentamt | 2023