Research On Automobile DAD Rapid-detection System Based On Machine Vision

With the development of society, automobile becomes one of the main vehiclemedium for people's traveling. But it brings to traffic accident hidden trouble whileit facilitates people's traveling. So it is seriously important for automobile securityperformance to be detected. And detection of Automobile DAD (difference of axledistance) is one of automobile detection items.Automobile DAD produces a great effect on auto using performanceespecially deteriorating automobile running stability. When it is braked or urgentlyaccelerated, automobile body inclines to side-slip, making it occupy more space ,affecting traffic security and increasing damage of tires, that brings un-negligibleeconomic damage and hidden danger. Consequently, it is of significance to realizeautomobile DAD detection and examine whether its detection result goes beyondthe limits of accepted result.At present, there are two traditional detection methods, photosensitivedetection and laser-detection. Photosensitive detection of automobile DAD adoptsthe way of locating wheel axis centre using reseau structure, and applies a theory ofrelatively measuring to get the test result of automobile DAD. Photosensitivereseau structure, apt to be affected by external factors, limited by mechanicalmachining technique, is made up of photosensitive tubes. And the testing accuracyof this detection method needs to be improved. Laser detection of automobile DADis difficult to realize the accurate location between transmitter and acceptor andensure the light intensity for radiation of light and interference among beams oflight. The premise of laser detection is that automobile runs in constant speed andfollows the direction parallels guiding line. But in fact, we can't make automobilerun straightly in constant speed and may not rectify the warp caused bycurve-running of automobile, and take error in detection result. Therefore, it needsa improved method, simple to operate, little external interference and moreaccurate detection result, to fulfill the need of detection of automobile DAD.Automobile DAD detection based on machine vision is designed for above reason.Machine vision is modern technology product of the development of imagetechnology, pattern recognition technology and computer technology, is oneadvanced technology realizing intelligence, automation and information. It canincrease producing efficiency by enhancing the producing flexibility andautomation with the testing advantage of accuracy, stability and less time. Somachine vision is introduced in detection of automobile DAD realizingcompensating the loss of error for automobile's curve-running, which makes thedetection of automobile DAD improved.This paper involves mainly the following aspects:1. The establishment of mathematics model of detection system: mathematicsmodel of automobile DAD detection base on machine vision is developed applyingthe theory of relative measurement of automobile DAD. It not only fits to thesituation that automobile go through the tester in a beeline but compensates errorfor automobile's curve-running. The purpose of compensation is to make testingresult more accurate and reliable by decreasing or avoiding warp caused by thechanging direction of automobile.2. The designing of hardware system: Hardware system is designed accordingto the selecting rules of detection system based on machine vision. The productsincluding WV-BP330 black-white CCD made by Japanese PANASONIC,DH-QP300 image acquisition card made by DaHeng and SE0484 camera lensmade by Japanese NSK are selected as main apparatus in image acquisition system.Illuminating project, front-illumination, is chosen selectively to get clear andaccurate image to improve image processing and data computing.3. Realizing image processing effectively: MATLAB is selected as a toolsoftware for its matrix computing ability in image processing. Tool box ofMATLAB provides image processing function abundantly to achieve the process ofimage with higher working efficiency, lack of work of matrix programming.Image processing technology about automobile DAD detection is analyzed indetail in this paper. Operators including Sobel,Prewitt.,Log and Canny arediscussed thoroughly for image edge detection in the field of machine vision. Theresult of image detection of the several operators is compared each other and theadvantage and disadvantage are given. Finally, Canny operator is adopted for itsability of anti-noise and higher accuracy of edge detection.4. The design of software system: In the development of detection system ofautomobile DAD, Visual C++ is adopted as developing platform for softwaresystem of DAD with the Matlab7.0. Application program is developed inmodularization pattern including equipment self-checking module, auto-detectionmodule, web transmission module, data manage module, error processing moduleand helping module. Auto-detection module, as the core of the detection system,comprises several subprogram modules like vehicle leading module, imageacquisition module, image processing module and evaluation displaying moduleand so on.5. Tested with experience: Detection system of automobile DAD based onmachine vision realized auto-detection of DAD. The detection system isexperienced through used vehicles. The result indicates that the detection system issatisfied to the requirement of automobile detection with accurate testing result.Automobile DAD is tested through machine vision technology with hightesting accuracy. The detection system is rather reliable and maintainable for itsmodularization design. With the further development of image edge detection ofimage processing technology and camera technology, automobile DAD detectionadopting machine vision technology is to have more and more advantages.