Research On Image Detection Method For Clay Content In Base Sand

Clay content is one of the important technical indicators to characterize the quality of base sand.Its size directly affects the performance of molding sand with basic sand as the main component,and then affects the quality of sand castings.At present,the detection of clay content in the base sand is still based on the detection method prescribed by the national standard,which is troublesome and takes a long time.Therefore,the development of a fast and accurate method for detecting the content of base sand and clay has important practical significance for improving the performance of molding sand.A method for quantitatively determining the content of base sand and clay using computer image analysis technology is proposed.Based on digital image analysis technology,the sand is separated by vortex washing,and the color of the sand washing liquid is different according to the clay content.Using digital image acquisition technology to image the basic sand turbid liquid after vortex washing.Combined with digital image processing and analysis technology,firstly extract the characteristic value from the image,then establish the mathematical model of the characteristic value and clay content,and finally realize the determination of the clay content of base sand through the established model.The feature of this method is that it can quickly test the clay content of foundry sand in a short time,and all the operation processes after the first vortex washing in the standard method are omitted.In this paper,the computer image acquisition and analysis technology was used to design the detection system of image method to detect the clay content of base sand.The system was based on MATLAB 2018 a as the programming platform.Firstly,the digital camera was used to obtain the turbidity image data with different clay content,and the image color space conversion algorithm was used to convert the RGB color space of the sand washing solution to the LCH color space,so as to construct the GUI interface for converting the sand washing solution image with different clay content from the RGB color space to the LCH color space;then the L,C and H components were obtained according to the GUI interface,the linear prediction model and BP neural network prediction model were established respectively.Finally,the image detection method and the national standard detection method were used to test and compare the base sand with different clay content.The results show that the prediction results of linear regression prediction model and BP neural network prediction model are more than 99% close to the test results of national standard detection method;and the image detection method shortens the test time,saves the follow-up steps of the first sand washing of national standard method,and realizes the efficient and accurate determination of base sand clay content,in which the stability and error of BP neural network are better than the line The regression model is good.