Study Of Wheels' Online Identification System Based On Machine Vision

With advantages of non-contact, high efficiency, high precision, online, real-time etc, machine vision identification technology plays an important role of improving flexibility and automation degree of production, and is also the foundation of computer integrated manufacturing technology. Based on analyzing the application of machine vision in modern production, in this paper the machine vision identification technology is studied on the application of on-line identification in molding line of car's light alloy wheel.Combining machine vision, image processing pattern recognition, and engineering control technology, using CCD sensor, motion control card and other hardware, the vision system is established. In Visual C++ 6.0 environment, according to extracting wheel's feature information, using algorithm of Hough transformation and edge extraction, the software of machine vision identification system is established to realize wheel's image acquisition, image processing, position recognition and characteristic window recognition. In image processing modules, the Intel open source library OpenCV is adopted for image processing. Wheel's position information is got by camera calibration. Through motion control card, manipulator is guided. Automatic transmitting, drilling, handling and stacking is realized.Through experiment, characteristic information of wheel is accurately extracted by this identification system. The wheel's molding production efficiency is improved. The disadvantage of low efficiency with manual clamping, drilling and sorting is overcome. For the application of machine vision in wheel molding line, it is the beneficial attempt.