Research on the application of residual networks considering attention mechanism in concrete curing robot

Abstract

Manual curing after concrete pouring is mainly divided into two methods: plastic film curing and water spray curing, both of which have shortcomings, such as high working intensity and intense subjectivity. Current intelligent curing methods mainly focus on establishing intelligent water spray curing systems according to site needs and have poor versatility. There are currently no relevant intelligent curing measures for plastic film curing after concrete pouring. To solve the above problems, we designed and developed a concrete curing robot system based on the com-puter vision technique, including the concrete curing robot body and a corresponding control terminal, which realizes the intelligent operation of concrete plastic film curing and water spray curing. Through the classification model based on the improved ResNet-18 concrete curing surface image, the robot can make scientific judgments and decisions on the water spray curing independently. After the research and development of the robot were completed, we tested and verified the feasibility of the concrete curing robot system design at the project of Tianjin Lingang Hospital. This paper provides new ideas and technical schemes for intelligent and automatic control of concrete curing, especially for plastic film curing and water spray curing. It has essential scientific and practical significance for concrete curing, especially large-volume concrete curing.