A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Concrete precast beam cross section edge data set generation method
The invention discloses a concrete precast beam cross section edge data set generation method. The method comprises the following steps: S1, stipulating nine control parameters of a concrete precast beam cross section: beam height H, top plate gradient m, top plate right side width w1, top plate left side width w2, top plate thickness t3, top plate chamfering distance d, side plate thickness t1, bottom plate thickness t2 and bottom plate width w3; s2, coordinates of 18 edge angular points are calculated according to the nine stipulated control parameters; s3, sequentially connecting the calculated angular points and increasing noise; and S4, outputting and storing the nine control parameters, the coordinates of the edge angular points and the coordinates of the noise points. Effective data set support can be provided for a concrete precast beam cross section parameter identification technology based on deep learning, the problem that an existing concrete precast beam cross section data set is insufficient in diversity due to limitation of site cost and the like is solved, and edge data sets of various models and sizes are rapidly generated; and the manufacturing efficiency of the section data set of the concrete precast beam is improved.
本发明公开了一种混凝土预制梁横截面边缘数据集生成方法，包括以下步骤：S1，规定混凝土预制梁横截面的9个控制参数：梁高H、顶板坡度m、顶板右侧宽度w1、顶板左侧宽度w2、顶板厚度t3、顶板倒角距离d、侧板厚度t1、底板厚度t2、底板宽度w3；S2，由所规定的9个控制参数计算18个边缘角点的坐标；S3，依次连接所计算出的角点并增加噪声；S4，将9个控制参数和边缘角点坐标以及噪点的坐标输出保存。本发明可为基于深度学习的混凝土预制梁横截面参数识别技术提供有效数据集支撑，解决现有混凝土预制梁横截面数据集因场地成本等限制导致的多样性不足的问题，快速生成各型号尺寸的边缘数据集，提高了混凝土预制梁截面数据集的制作效率。
Concrete precast beam cross section edge data set generation method
The invention discloses a concrete precast beam cross section edge data set generation method. The method comprises the following steps: S1, stipulating nine control parameters of a concrete precast beam cross section: beam height H, top plate gradient m, top plate right side width w1, top plate left side width w2, top plate thickness t3, top plate chamfering distance d, side plate thickness t1, bottom plate thickness t2 and bottom plate width w3; s2, coordinates of 18 edge angular points are calculated according to the nine stipulated control parameters; s3, sequentially connecting the calculated angular points and increasing noise; and S4, outputting and storing the nine control parameters, the coordinates of the edge angular points and the coordinates of the noise points. Effective data set support can be provided for a concrete precast beam cross section parameter identification technology based on deep learning, the problem that an existing concrete precast beam cross section data set is insufficient in diversity due to limitation of site cost and the like is solved, and edge data sets of various models and sizes are rapidly generated; and the manufacturing efficiency of the section data set of the concrete precast beam is improved.
本发明公开了一种混凝土预制梁横截面边缘数据集生成方法，包括以下步骤：S1，规定混凝土预制梁横截面的9个控制参数：梁高H、顶板坡度m、顶板右侧宽度w1、顶板左侧宽度w2、顶板厚度t3、顶板倒角距离d、侧板厚度t1、底板厚度t2、底板宽度w3；S2，由所规定的9个控制参数计算18个边缘角点的坐标；S3，依次连接所计算出的角点并增加噪声；S4，将9个控制参数和边缘角点坐标以及噪点的坐标输出保存。本发明可为基于深度学习的混凝土预制梁横截面参数识别技术提供有效数据集支撑，解决现有混凝土预制梁横截面数据集因场地成本等限制导致的多样性不足的问题，快速生成各型号尺寸的边缘数据集，提高了混凝土预制梁截面数据集的制作效率。
Concrete precast beam cross section edge data set generation method
一种混凝土预制梁横截面边缘数据集生成方法
WANG HAO (author) / NI YOUHAO (author) / XU YINFEI (author) / LIN YUXUAN (author) / TIAN JIYUAN (author) / MAO JIANXIAO (author) / CHEN GANG (author) / LYU MANQI (author)
2023-03-03
Patent
Electronic Resource
Chinese