Research On The Recognition Method Of The Rem Oval Rate Of Abrasive Belt Grinding Based On Image

With high efficiency and low cost,abrasive belt grinding is widely used in various fields.However,manual grinding has the disadvantage of poor dimensional consistency.Because the theoretical prediction model does not take into account the impact of abrasion and vibration of grinding tools,it is impossible to accurately adjust the material removal rate.Therefore,this paper proposes a prediction method of material removal rate based on grinding spark images by studying the internal relationship between spark size,shape and material removal rate.In order to solve the above problems,the following research work is carried out in this paper:This paper designs and builds a belt grinding test bench.The test bench has realized the function of clamping workpiece and changing the depth of grinding.At the same time,industrial CMOS camera is used to capture the spark image.Lay the foundation for the follow-up research.The interference factors such as sand belt machine in the background can be removed by preprocessing the image.The contour,brightness,area and color of the spark image are extracted.After analysis,it is found that there is little difference between spark image characteristic under the same material removal rate,while there is great difference between spark image characteristic under the different material removal rate.In order to establish the prediction model,the correlation between spark characteristics and material removal rate was analyzed.Based on linear regression and polynomial regression,a single characteristic prediction model was established.Multiple linear regression and support vector machine regression were used to establish the prediction model of multi-feature.The experiment of 60 data sets shows that the SVM has the highest prediction accuracy.In order to verify the rationality and effectiveness of the method,150 sets of data of 60,120 and 240 model sand belts were collected and the prediction model was established based on SVM.The experimental results confirmed that this method was reasonable and effective.