Development of a robot vision system for installation progress monitoring

Abstract

This dissertation researches on the implementation of an industrial solution for construction site progress monitoring by deploying a robot with vision sensors and artificial intelligence techniques. Construction progress monitoring is an important topic for efficient construction project management with satisfactory quality. This research aims to use a robot for progress monitoring in order to reduce manpower involved and to ensure well-organized supervision on the construction progress. The application scope of this dissertation focuses on the constructions using Prefabricated Prefinished Volumetric Construction (PPVC) method. Thus, this dissertation will research and propose a program including the pre-processing of the images captured in a PPVC construction site, the development of the machine learning and convolutional neural network for object recognition, and the post-processing of the recognition results.