The robotic arm needs to put clutches in automatic guided vehicle（ AGV） accurately after scraping clutches from the assembling line. As the storage location of AGV is uncertain and the robotic arm can ’t accurately realize the blanking

use the technology of machine vision to design a localization system of vision-based clutch unloading. Aiming at problem that the size of AGV is too large and localization of single camera is unstable and precision is low. Use binocular localization to improve the precision and stability. Practical application result show that the designed AGV feature recognition and location system which based on Vision Pro vision software

has wonderful locating results

the location precision is ± 0. 5 mm

and the run time is less than 2s

besides

it can guide robotic arm to realize the clutch blanking task precisely

and quickly

meeting the requirements of industrial application.