The Design And Implementation Of Online Detection System For Quality Of LED Cup

Widely application of LED lighting drives the rapid development of LED cups and puts forward higher requirements for the yield and the consistency of product quality of the LED cups.However,at present,in the actual industrial production process,defect detection of LED cups still absolutely relies on the traditional manual detection.Low detection efficiency and visual fatigue are the disadvantages of manual detection which reduce yield and the consistency of product quality of LED cups.Fortunately,machine vision is becoming the mainstream detection technology because of its online detection,non-contact and high efficiency.It can not only enhance the consistency of product quality,but also improve the detection efficiency.Therefore,this paper puts forward the online quality detection system of LED cups based on machine vision.First of all,this paper designs the overall scheme of online quality detection system of LED cup from actual demand.Based on the technical indicators of the system,the selection and analysis of industrial camera,lens and light source are made in the paper.In order to trigger the camera to take a picture and separate the inferior cups,the control unit is designed and the working principle of each module in the main control board of the control unit is introduced in detail.On the basis of the above,the design scheme of the aperture detection system and the rivet detection system is proposed.Then,the aperture detection algorithm and the rivet detection algorithm are emphatically studied.In order to measure accurately the aperture of the hole,the method of extracting maximum inscribed circle based on the iteration of three points is proposed in the aperture detection algorithm.In the rivet detection method,a fitting circle algorithm based on the least square method is proposed which is able to calculate the hole spacing of the rivet and judge whether or not the hole of rivet is blocked.Also,sliding window algorithm proposed in the rivet detection method combines with the distribution of pixel value in binaryzation image can measure the actual depth of the groove of the LED cups.Next,on the basis of the hardware platform and algorithm,the whole framework of the system software is designed.The upper system is responsible for image processing and the lower system responsible for the control.Both achieve the function of separating the inferior cups through the serial communication.In addition,in order to be more friendly to the user,a human-computer interaction interface based on MFC is designed.Finally,the whole system and the internal modules were tested,and the test results were analyzed.In general,the basic functions of the system can be achieved.The resolution of the aperture detection system can reach 0.01 mm,and the resolution of the hole spacing in rivet detection system can reach 0.05 mm.However,after testing a large number of samples,it can be found that due to the accuracy and stability of the algorithm the misjudgement rate of bad product remains to be reduced.The misjudgement rate of the bad product of the aperture detection system is 1.74%,and the misjudgement rate of the good product is 0%.The misjudgement rate of the rivet detection system is 1.22%,and the misjudgement rate of the good product is 0%.