基于小数据的钢筋检测方法的研究

Research on Rebar Detection Method Based on Small Data

在工地建筑领域,钢筋已然成为不可或缺的基本材料之一,在我国的绝大数工地,现有的钢筋计数方法是工人逐一统计已计数和未计数的钢筋,不仅耗时、耗力,而且由于工人长时间计数钢筋,人眼和人脑容易疲劳,计数误差将增加[2]。由于工地钢筋图像数据不便于大量采集和标注,提出一种基于小数据的钢筋检测模型,将该模型应用于智慧工地中的钢筋计数应用,工人只需将拍摄的钢筋图像通过终端传入模型,便可以快速准确地得到图像中钢筋的数量。

In the field of construction site,rebar has become one of the indispensable basic materials.At present,in the vast number of construction sites in China,the existing rebar counting method is that workers count the counted and uncounted rebar one by one.This process is not only time-consuming and labor-consuming,but also worker counts rebars for a long time,worker’s eyes and brain are prone to fatigue,and the counting error will increase[2].Because the image data of rebar in construction site is not easy to collect and label,it proposes a rebar detection model based on small data.This model is applied to the rebar counting application in intelligent construction site.Workers only need to upload the captured rebar image into the model through the terminal,then they can quickly and accurately get the number of rebar in the image.