Research On High-throughput Characterization Of Pore Parameters Based On Materials Genome Initiative

Materials Genome Initiative(MGI),developing new integrated computational,experimental,and data informatics tools,aims to accelerate materials discovery,develop-ment,and deployment.Due to the lack of high-throughput characterization measurement tools for pore parameters of material and coatings,a high-throughput characterization experimental determination platform were designed for pore parameters(including poros-ity,pore spacing,pore radius)based on machine learning,image processing algorithms,and parallel calculations.In addition,a high-throughput pore parameter determination platform were applied to characterize the AZ91D magnesium alloy micro-arc oxidation experimental coating characterization images and the 10Ni5CrMoV steel corrosion ex-periment substrate characterization images.The high-throughput determination platform were validated through online and offline experiments.The specific research results are as follows:(1)A high-throughput pore parameter determination platform were built based on machine learning,image processing technology,and multi-threaded parallel computing.Firstly,a high-precision binary image is obtained through machine learning.Then,the pore area in the obtained binary image is detected through Hough's principle,and the neighboring pore distance is calculated by image processing algorithm.The platform meets the online and offline high-throughput measurement conditions.(2)In the determination of the pore parameters(porosity,pore radius,and pore space)on the surface of anticorrosive functional coatings including 10Ni5CrMoV steel and AZ91D magnesium alloy,the high-throughput characterization platform of pore parameters achieves the same measurement accuracy as traditional measurement.The average error of the radius and pore spacing measurement is not more than 4%.(3)Based on the high-throughput characterization platform for pore parameters,high-throughput measurement of the pore parameters(porosity,pore radius,and pore space)on the surface of anticorrosive functional coatings including 10Ni5CrMoV steel and AZ91D magnesium alloy is performed.The platform meets the high-throughput real-time online measurement Demand.In addition,when faced with a large number of processing samples,the algorithm platform uses a parallel operation and data set segmentation scheme to accelerate the high-throughput characterization process.The high-throughput computing platform guarantees that each image calculation time does not exceed 2s.The platform has completed the measurement of pore parameters(porosity,pore spacing,and pore radius)in the form of high-throughput.The platform meets the online measurement conditions and offline high-throughput measurement conditions,which improves the reproducibility of the pore material characterization and improves the characterization experiments effectiveness.