Research On Visual Detection Method Of Automobile Safety Airbag

With the rapid development of Automobile industry modernization, thequality detection technique of airbag assembly, which combines computersoftware technique and intelligent technology, has become an important researchdirection in digital automatic automobile assembly. In this thesis, an airbag is thestudy object and an online quality monitoring system is designed. This system isbased on visual image processing and artificial neutral network technique andimproves the automatic inspection level of an automatic airbag assembly line.The airbag snap-fit detection system consists of data processing, display andmonitoring of visual image. An X-Y Cartesian-coordinate robot is used to movethe video camera with which visual image is collected and guaranteed theaccuracy of orientation. Graying and binarization of a visual image is performedusing MATLAB functions and then the feature in the joint is extracted throughHough transform. When the fault feature is input to the BP neural network, thefault type of edge detection is distinguished.A visual image is obtained using a facility which simulated the factoryequipment. A fault type is chose to conduct an experiment. During this period, theairbag snap-fit online quality inspection system runs well and is efficient. And thesystem has human-computer interface with full functions and is easy to operate.The result proves that the airbag snap-fit online quality inspection system reachthe expected goal and has good application prospect and practical value.