| Automobile stabilizer bar is a common parts of automobile, the spatial shapeerror is an important control index. The traditional detection method is artificialdetection, it detects low speed, low efficiency and greatly influenced byartificial factors, it is hard to meet production needs. So, this paper proposes acar stabilizer bar non-contact measurement based on machine vision method,design the whole system, and further study of the key technologies.First of all, for the auto stabilizer bar size is larger, irregular characteristics,this paper designed a three axis servo moving system. Uses the camera along thespecified path to collect images, and design the structure of the three-axis servomobile system and the control unit. Secondly, in view of this article researchobject auto stabilizer bar with the imaging features of simple structure, so thispaper proposes a grayscale differences determine method based on thresholdsegmentation algorithm, and has carried on the experiment to the algorithm, bycomparing with the traditional edge detection algorithms, it is concluded thatthe algorithm possesses the advantages of accurate and fast. Then, based on thesymmetric point method to extract the image of the center line, respectively bythe least squares method to design automobile stabilizer bar’s straight line andarc part error calculation method, and the method is verified through theexperiment, this method has been successfully applied to the automobilestabilizer bar error evaluation. Finally, the system was calibrated, calculate theper pixel, and the distortion error of the lens is analyzed.The visual inspection system can improve the flexibility and automation ofthe production, is suitable for large size and irregular shape parts detection, itcan solve the low efficiency of manual detection, the precision is not high,unable to real-time online measurement shortcomings. It has very broadapplication prospect. |
PDF Full Text Request