Research On Online Detection Method Of 94 Ice Scoop Defects

Ice scoop is a tool that people need to use when eating ice cream.Ice scoop is generally made of birch.Due to the influence of wood material and the process in the production of ice scoop,the surface defects of ice scoop will appear,which will affect the beauty of the product and produce potential safety hazards.Artificial vision detection is the earliest and most widely used method.Although advanced detection technologies such as artificial intelligence and machine vision have gradually matured,naked eye defect detection still accounts for a large proportion and widely exists in small and medium-sized enterprises.However,with the disappearance of the demographic dividend,as well as the boring work,low degree of freedom and low wages,fewer and fewer people are willing to engage in quality inspection,and the employment problem is becoming more and more prominent.From the current development trend,machine vision and other advanced detection systems will gradually replace the artificial.Aiming at the deficiency of manual detection of ice scoop surface defects,In this paper,the method of on-line defect detection for type 94 ice scoop is studied.Firstly,according to the selection of the current ice spoon detection hardware system,the relevant selection parameters of camera,lens,light source and other hardware are calculated to verify the feasibility of the current system for subsequent detection algorithm research.Then,the defects of ice scoop,such as burr,splitting,pollution,thin material and lack of material,are studied.It not only involves the traditional image processing algorithm,but also proposes the image segmentation algorithm based on direction space and the defect classification method based on SVM.Finally,in the aspect of software,a man-machine interface based on MFC is designed,and the test data of multiple cameras are synthesized,and the jet valve is controlled to complete the sorting according to the ice spoon grade.The image segmentation algorithm based on direction space has a good effect for small burr,which solves the problem of small burr missing detection before,and can detect the split on the surface,which makes up for the shortage of using only head image to detect the split;using shape based template matching to detect the lack of material has also achieved good results;at present,the detection effect of defect classification based on SVM is not ideal After testing,the software system can run well.