Research On Concrete Surface Quality Evaluation And Decision-making Framework Based On Ontology And Computer Vision

The concrete surface defects occur repeatedly during the construction process,which have an impact on the structure,living environment and perception of the building.The current specifications require that the surface quality of concrete entities need to be inspected during construction acceptance,but the inspection is usually through manual judgment,requiring inspectors to evaluate the type,cause,and severity of defects,and propose corresponding repair suggestions.This inspection method is affected by the experience of the inspector and the on-site environment,is inefficient and the results are unreliable,and the formed experience and knowledge cannot be effectively reused.In addition,the judgment standards of different inspectors are not uniform,and the results of inspections are also different.Therefore,this article aims to propose a concrete surface quality evaluation and decision-making framework that combines ontology and computer vision technology,establishes an ontology model to reuse defect knowledge,then establishes evaluation criteria,and use computer vision to realize automatically identify and quantify defects.After that,use reasoning to achieve the automatic evaluation and decision-making of surface quality.This method improves the standardization and intelligence of surface defect inspection.Through literature research and interpretation of standards and norms,this paper analysis the gap.The current research lacks standardized concrete surface quality assessment and intelligent decision-making methods.Based on this,this paper clarifies the three research purposes,the structured expression of knowledge,automated multiple defect recognition and quantification,and standardized surface quality evaluation and decision-making.Therefore,the role of ontology and computer vision technology are clarified.And builds a concrete surface quality evaluation and decision-making framework combining the two technologies.The framework consists of three levels,three core functions,and five specific implementation steps.Secondly,clarify the relevant knowledge in the field of concrete surface quality defects,establish a domain ontology knowledge base based on Protégé,and clarify the evaluation and decision criteria according to relevant research and industry norms.Based on this,SWRL rules are established.This provide a knowledge basis for the reasoning of evaluation and decisionmaking.Then,to improve the efficiency of defect identification and quantification,based on the comparison with the more widely used computer vision models,a computer vision model suitable for the identification and quantification of concrete surface holes and cracks was established to automatically identify the types of defects and extract the parameter information.Finally,the application process of the joint application of two technical methods to support intelligent surface quality evaluation and decision is discussed,and verified the operability of the framework through examples in different scenarios.Based on ontology and computer vision technology,this paper proposes a method suitable for concrete surface quality evaluation and decision-making.This method can automatically identify and quantify defects,and infer the causes,responsibilities,severity,and repair methods of defects.It can be used as a tool for learning and defect storage,supporting knowledge sharing and reuse.This article discusses the surface of holes and cracks in concrete.With the enrichment of the defect data set and the update of knowledge,relevant personnel only need to expand and update the established ontology model,and train and optimize the computer vision model to suitable for actual construction projects.The results of this paper can improve the standardization and automation of concrete surface quality inspection and acceptance,and also have certain significance for promoting the application of AI in the construction industry.