Research On Appearance Defect Detection System Of Foamceramic Filters For Foundry

In this paper,the appearance defect of ceramic foam filter for foundry is taken as the research object,and the visual defect detection method is used to study the appearance defe-ct detection system of ceramic foam filter for foundry.The specific work is as follows:(1)Design and build an appearance defect imaging system for ceramic foam filter for foundry.Firstly,according to the requirements of detection standard and sample characteristics,the appropriate camera,lens and light source are selected through theoretical calculation.Secondly,the imaging system is designed and built to verify the optimal imaging system.Finally,the sample library is built based on the imaging system.(2)The method for detecting the appearance defects of ceramic foam filter for foundry was studied and realized.The foamed ceramic filter screen for casting was analyzed for its own sample defects and appearance defects such as corners,missing edges,pits and cracks.For image background interference,the background segmentation and area feature filtering are used to achieve background clipping and sample area separation.For the missing edge defect,the convex hull algorithm is used to fit the contour of the filter,and the depth between the front and back of the contour is calculated by calculating the distance between the front and back of the contour.For the corner defects,the contour of the filter is fitted by the minimum circumscribed rectangle algorithm of any orientation.And by calculating the distance between the contour vertices and the contour before fitting,the depth calculation of the corner defects is realized.For pits and crack defects,the extraction of defect areas and the calculation of defect length are performed by algorithms such as filling,difference set,closed operation,area feature extraction and smallest enclosing circle.(3)Software design.Design human-computer interaction interface,integrate function modules such as image acquisition and detection algorithms,and realize a set of detection processes integrating image acquisition,detection and output.At present,the average accuracy of the system for the detection of defects in rectangular samples is 98.7%,and the average accuracy for detecting defects in circular samples is 93.3%.