Research On The Technology Of Machine Evaluation And Classification Based On The Assembly Quality Of Standard Chassis

Chassis is the key protection structure and supporting components used in a large number of electronic equipments,data communication equipments,information technology equipment and almost all of the equipment.In view of the current evaluation of the assembly quality of the chassis,the artificial visual method is used to judge the defects of the incorrect assembly,leakage assembly and bad assembly.This paper,"Research on the technology of machine evaluation and classification based on the assembly quality of standard chassis",studies the framework of machine evaluation and classification system for chassis assembly quality,SVM machine evaluation technology of slow beat chassis assembly quaility,SoftMax evaluation technology for assembly quality of multi part fast beat chassis.It is of great academic value and practical significance to realize the automation of the detection of standard parts chassis and to promote the development of intelligent manufacturing technology.The research work was supported by Guangzhou Science and Technology Program(No.201701016064,201509010008).In this paper,the framework of machine evaluation and classification system for chassis assembly quality,SVM machine evaluation technology of slow beat chassis assembly quaility,SoftMax evaluation technology for assembly quality of multi part fast beat chassis are proposed.The validity and advancement of SVM dichotomy and SoftMax multi-classification methods are verified by experiments.Main research work of this paper includes followings:(1)Through analyzing the system requirements,the system framework is proposed for machine evaluation and classification of the chassis standard parts assembly quality.After determining the system working mechanism and process on this research,this paper researches on the key technologies of machine evaluation and classification for chassis assembly quality.(2)Applying support vector machine(SVM)to train multiple discrete SVM classifiers based on oneversus-rest(OVR)strategy,machine evaluation system with good generalization ability is established.This paper presents a machine evaluation system of inter class combined equilibrium tree,which can realize the chassis assembly quality evaluation of the slow production line and less kinds of standard parts.(3)Based on the characteristics of minor calculation and low resource occupancy rate of SoftMax,this paper puts forward the SoftMax full classification machine evaluation technology for the chassis assembly quality of the fast beat.The relationship between the training complexity and the types of chassis standard parts is analyzed.Further research on SoftMax multi-classifier structure optimization training complexity based on inter class combination is studied to construct SoftMax machine evaluation system for the chassis assembly quality of multi-site and fast beat,relizing the chassis assembly quality evaluation,which has the characteristics of fast production line,more kinds of standard parts and more sites of assembly detection parts.(4)Based on the construction of chassis standard parts assembly platform,Machine evaluation test of standard parts chassis assembly quality is done with Lenovo ThinkCentre G3 commercial computer case.Using SVM dichotomy and SoftMax multi-classification to classify the samples and to determine the error assembly,leakage assembly and poor assembly.The validity and advancement of SVM dichotomy and SoftMax multi-classification are verified.