Research On Defect Recognition Of LCD Light Curing Molding Based On Machine Vision

With the continuous improvement of science and technology,3D printing technology has developed rapidly,especially the emerging 3D printing technology represented by LCD curing technology,which has developed rapidly in recent years.Compared with other consumer-grade 3D printers,LCD photo-curing technology 3D printers have the advantages of high molding accuracy,high printing efficiency,and low cost.In the 3D printing industry,research on LCD photo-curing technology 3D printers has continued to increase.The production of 3D printers with LCD light curing technology is increasing.However,there is currently no professional LCD light curing molding defect recognition and detection equipment in the industry,and manual quality inspection is still commonly used on production lines to inspect products.Although this detection method can meet the requirements of printer quality inspection to a certain extent,this detection method relies too much on manual detection,has large subjective influence factors,and has low detection efficiency,high labor costs,and cumbersome data statistics.In view of the above problems,this topic proposes a research on the defect recognition of LCD light curing molding based on machine vision.Aiming at this research,a set of product defect recognition system is designed in this subject.The system includes structure,electrical control,software and other parts.The structural part is designed using Solid Works 3D drawing software.The electrical control part is based on the cost-effective Cortex-M3 core chip STM32F103.The programming environment is keil for embedded programming.The system The host computer software is based on the Windows operating system,and uses the Opencv vision library and Qt as development tools for system software page design and algorithm implementation.The main work of this article is as follows:(1)Based on the requirements of the subject,by consulting the relevant literature of machine vision technology,investigate the development status and trends of the technology at home and abroad,and analyze the functional requirements of the defect recognition system initially;(2)According to the system function requirements,design a three-axis motion detection platform and build a physical object.The platform can carry various models,and the high-quality pictures of the model can be captured by adjusting the spatial orientation of the camera through the movement mechanism;(3)For the system structure platform,design an electrical control scheme that meets the functional requirements,select electronic components,and implement functions such as three-axis movement,serial communication,and axis movement limit through the underlying control algorithm;(4)A support vector machine(SVM)algorithm based on directional gradient histogram(HOG)feature extraction is used.This algorithm first needs to extract HOG features from a large number of light-cured molding samples,and then train to obtain an SVM with a recognition rate that meets the requirements.Classifier;(5)Designed with multifunctional upper computer software,to realize camera parameter setting,sample model training,model defect detection functions,and complete debugging and verification of each function.(6)Experiments are carried out on the defect detection system to verify the effectiveness of this algorithm.This system mainly completes the recognition of models with crack defects in appearance.The whole system debugging results show that the operation of each module of the system satisfies the predetermined functions,and it can better identify and classify the crack defects of the model,which is in line with the expected effect.