Research On Intelligent Maturity Identification Of Efficient Concrete Mixing Based On High Definition Photography

Concrete is an essential building material for modern engineering construction.Real-time detection of maturity is one of the very important aspects in the concrete mixing process.Due to the different ratios of concrete and the complex site working environment,the concrete maturity detection has not yet reached the real-time and intelligent detection requirements.At present,the maturity testing of concrete mixing mainly relies on manual measurement,which has the problems of inaccurate measurement,human resources consumption and labor intensity.The existing testing methods of concrete maturity are post-testing methods,which are complicated,difficult to implement and have great limitations.Therefore,further research on intelligent concrete maturity identification technology is needed to provide technical support for efficient production and intelligent construction of concrete mixing plants.With the development of high-definition imaging technology and artificial intelligence,real-time concrete inspection technology based on high-definition cameras will help meet the demand for real-time monitoring of concrete maturity.This paper investigates the concrete mixing cycle model and the mix state identification model,respectively,to obtain the key characteristics of the homogeneity and slump of concrete mixes through high-definition camera;Detection and localization of the concrete state using image feature fusion methods and model analysis based on the obtained real-time concrete state;The testing model of concrete uniformity and slump is constructed to provide a theoretical basis for the intelligent identification of concrete maturity.The main research contents and innovation points of this paper include:1.An intelligent maturity identification platform for concrete efficient mixing based on high-definition photography was built.The recognition effects of different image feature algorithms were compared comprehensively and compared with deep learning algorithms.Concrete maturity was divided into uniformity recognition stage and slump detection stage,key features of concrete mixing workability were extracted,and a concrete mixing model based on convolutional neural network was discussed.Judge the maturity of concrete mixing and realize visualization.2.The intelligent identification model of dynamic mixing concrete uniformity was designed.Through the analysis of the video stream data transmitted by high-definition cameras,the mixing state of concrete was regarded as the periodic Markov motion,the fluid wave characteristics of the mixture were extracted,and the Kolmogorov equation of the concrete mixing cycle in the diffusion process was established.The online detection of concrete maturity based on multi-feature fusion is realized,and the average accuracy of recognition is 97.14%.3.A concrete slump detection model based on multi-feature fusion of periodic time series is proposed.The concrete slump value is tested online and verified on the spot by using the data fusion of periodic state and multiple physical characteristics of concrete mixing.Moreover,multiple methods are applied to verify and compare each other,and the error of setting parameters of each model is compared.The optimal detection of concrete slump based on multi-feature fusion of periodic time series is realized.The error value of concrete slump is 4.57%,and the average absolute error is9.71 mm.Finally,this paper applies the above research results to Wuhan Minghua Concrete Company,verifies the feasibility and practicability of the research content,and provides technical support for the intelligent development of concrete mixing plant.