The Contour Size Detection System Research Of Automotive Airbag Based On Binocular Vision

Automotive airbags can effectively protect the safety of passengers in the process of traffic accident, their quality has been paid much attention all the time. it has important research value to fast and effective check whether the counter size of the automobile airbag is qualified or not. In the process of assembly, size deformation or oversize will lead to assembly tight pack, even can not put in. And it will leave gaps if it undersize. So it affects not only beauty and use of performance, but also will bring potential safety hazard, and lost car airbag proper function at a crucial moment.Dial gauge, contourgraph and DAB fixture are common detection methods of car airbag contour size. These detection methods have shortcomings such as low detection accuracy, poor repeatability, high misjudgment rate, and low efficiency. In this paper, we propose a car airbag contour size binocular vision detection method through in-depth study the principle of auto airbag contour size detection technology, combined with digital image processing technology.In this paper, we first introduce the research status at home and abroad of automobile airbags and the binocular stereo vision technology, then put forward the research content of this paper and design the technical route according to subject research requirements;Secondly, according to the actual demand to choose lighting system, cameras and other hardware, design the overall structure of the detection system and formed the whole testing experiment platform; By researching the binocular vision detection system algorithm, this paper proposes a car airbag contour size detection based on binocular vision method. Using MATLAB standard toolbox to proceed binocular stereo calibration; Calibrating right-and-left image parameters using OpenCV, Preprocessing the extractive automobile airbag images. Using Canny operator to extract the edge profile of airbags. Tracking and taging the edge profile of the automobile airbag, using Average of Square Difference and Normalized Cross Correlation method for edge contour coarse matching, using a dynamic programming algorithm for the airbag edge contour curve optimize matching. Finally,processing three-dimensional reconstruction for the optimal matching edge profile of airbags. Combined three-dimensional reconstruction coordinates of automobile airbag under test with the three-dimensional mathematical model of standard airbags, established the position tolerance Vx, Vy, Vz of automobile airbag contour testing point and form tolerance K of points on the contour curve segments as assembly accuracy parameter evaluation index, completed contour parameters calculation automatically, determine whether the airbag under test is qualified.Actual verification results show that, the established automotive airbags contour sizebinocular vision detection system and the evaluation index can evaluate the assembly quality of automobile airbags fastly and accurately. It saves the manpower, improves the efficiency, saves resources and costs compared with the traditional manual testing, and has good practical value, it is conducive to the development of auto industry.