Research On The Calculation Method Of Breakage Rate Of Refractory Materials In Thermal Shock Test Based On Machine Vision

With the development of high temperature industry,refractories will assume more and more functions.As an important property of refractories,thermal shock resistance is of great significance in the research of structural design and preparation technology of refractories.Thermal shock resistance to multiple cyclic heating-cooling test,in order to solve the traditional thermal shock resistance of refractory material detection,such as low efficiency,poor safety,no defects such as record process,in accordance with national standards,in this paper,many kinds of refractory samples,carried out the thermal shock resistance of high precision,automatic visual inspection research,mainly as follows:(1)Using the detection method based on machine vision,combined with automatic mechanical equipment,optical camera,computer and other fields of technology,developed a set of automatic detection device that can adapt to the thermal shock resistance of different types of refractory materials.(2)The acquisition of a variety of sample image feature analysis,in view of the sample is heated face will gradually appear different degree of crack and spalling,especially the outside edge of spalling cause of the image contour error detection problem,in this paper,the initial image edge detection,the mask of image and outside boundary M2 and M1.The mask image M1 and the subsequent images collected by the same group in the test were operated bitwise and to obtain the complete sample image with the surrounding background interference removed.Then binarization,morphology and other operations were carried out to complete the image preprocessing.(3)For sample images there are a lot of pitting and block the noise,cracks and large area of spalling and other issues,the samples taken after image preprocessing,the noise,cracks and peeling connected domain characteristics of K-means cluster analysis,and the clustering results,this paper proposes a connected domain through the calculation of setting initial error coefficient E feature segmentation threshold T algorithm,determine two groups of feature segmentation thresholds.After the preprocessing of the test image,the three-stage filtering of two connected domains was carried out based on the two groups of eigenvalues to accurately detect the cracks and spalling regions.(4)In the calculation of damage rate,skeleton refinement was carried out on the detected cracks and peeling areas,and chain code was introduced to measure the length and width of the damaged area.According to the particularity of the calculation standard of damage rate in the thermal shock resistance test of refractory samples,a calculation method of damage rate based on the interval division of skeleton line was proposed.The experimental results show that the proposed method is more accurate and faster than the existing grid method.The thermal shock resistance detection system established in this paper can greatly reduce the workload,automatically save all the process images and accurately calculate the damage rate results of each test.It is convenient for manufacturers and staff to analyze the location,length,peeling area,morphology and other information of cracks according to the data results of each test,and study the thermal shock resistance of the material sample,which provides an important basis for the optimization of product structure,efficient production and green development of the refractory industry.