Material Fracture Detection And Stress Information Estimation Based On Microscope Monitoring Video

With the increasing requirements for the manufacturing quality of various materials in the industrial field,higher requirements are put forward for the performance research of materials.In recent years,it has become a mainstream method to study the mechanical properties of materials by using surveillance video to identify surface defects,track fracture and detect stress through image analysis.Digital image correlation(DIC),also known as digital speckle image method,is an image method of stress measurement.DIC sprays irregular speckle image on the object studied,then obtains the deformation information of the region of interest through relevant calculation according to the two digital images before and after material deformation.As a very mature technology,DIC has been widely used in material science,architecture,aerospace and other fields,and its related technology has also made rapid development,including search algorithm and matching algorithm.However,in order to obtain more accurate deformation and stress characteristics of materials under external force,it is necessary to detect fracture information and estimate stress characteristics by using material tensile monitoring video with many times of magnification.The traditional DIC cannot detect the deformation and stress characteristics of microscope video with high magnification,because when the material video is magnified many times,the actual size of the corresponding material in the camera area will become very small.Using the traditional speckle injection method,it may result in the whole monitoring range of speckle injection and ignore other effective information.Therefore,in order to solve this problem,a method of material fracture detection and stress information estimation method based on microscope monitoring video is proposed by using feature point matching based on scale invariant feature transform(SIFT)and mixture Gaussian model.The details are as follows:(1)Firstly,SIFT feature points are extracted from the images with a certain number of frames in each interval of the microscope monitoring video,and the feature points of the adjacent images with a certain interval number of frames are matched.Then,the number of matched feature points is counted,and the time point of material fracture is determined according to the matching degree.Finally,the number of feature points of the image in different regions is counted to determine the region of material fracture.According to the change law of the number of matching feature points,the time and location of fracture can be accurately detected in the tensile microscope monitoring video with many times of magnification,and even the number law of matching feature points can be used to predict the location of fracture in advance.(2)The number of matching feature points in the detected fracture area is counted according to the block area,and then the image block with more matched feature points is modeled by mixture Gaussian model.According to the model parameters obtained from mixture Gaussian model,the stress information when the material is about to fracture is analyzed.By calculating the ellipse features corresponding to each model,the stress information of each region is estimated when the fracture is about to occur.In this paper,the proposed method is verified experimentally by using two pieces of microscope surveillance video of material tension with different magnification.For tensile video with different time,frame rate and resolution,the time point and region of fracture are detected very accurately in our experiments.Meanwhile,the stress and deformation trend of the material are analyzed by using the mixture Gaussian model.The experimental results show that the proposed method can accurately detect the fracture time and location of the material by using the microscope monitoring video of the material tension with many times of magnification,and give the estimation results of the stress information causing the fracture,which provides an important reference for the subsequent material tensile stress analysis.