Please use this identifier to cite or link to this item:
|Title:||Development of a vision system for construction materials classification and detection||Authors:||Li, Tao||Keywords:||Engineering::Electrical and electronic engineering||Issue Date:||2020||Publisher:||Nanyang Technological University||Abstract:||Computer vision (CV) is a science that studies how machines "see". In addition, it refers to the use of cameras and computers rather than the human eye to identify, track, and measure machine vision, and perform further graphics processing in the computer. As the core of artificial intelligence, algorithms such as pattern recognition, machine learning, and deep learning give computers powerful recognition capabilities. Among them, Convolutional Neural Network (CNN) is a feedforward neural network with convolutional calculation and deep structure. It has been widely used following its recognition performance with excellent characteristics such as weight sharing, less trainable parameters, and strong robustness. This dissertation mainly uses the features and advantages of convolutional neural networks and various network structures to complete two main applications in: construction materials recognition and detection, to improve efficiency and safety in construction site. Specifically, InceptionNet is used to train four categories, brick, mesh, wood and cement, each with 200 image datasets to complete the building materials classification, so as to feedback real time on-site materials inventory status to supervisors. The You Only Look Once (YOLO) object detection model is used to train 400 data sets containing stacked bricks and scattered bricks to complete the detection of stacked and scattered bricks, which can be used to investigate hidden safety hazards in construction and can further reflect the order of the construction site. The completion of the above tasks illustrates the practical applications of CV and CNN in the construction area, which can effectively help the construction supervisor to remotely monitor the progress of the work and the safety of the construction site.||URI:||https://hdl.handle.net/10356/141309||Fulltext Permission:||restricted||Fulltext Availability:||With Fulltext|
|Appears in Collections:||EEE Theses|
Updated on May 15, 2022
Updated on May 15, 2022
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.