Research On Sorting Workpiece On Machine Vision

Part sorting is an indispensable part on the industrial production line.Artificial sorting is usually used in traditional production,but with the popularization of automatic production to speed up production,manual sorting can no longer meet the demand of production.Aiming at this situation,machine vision is introduced into the sorting system of automotive symmetrical parts.This paper puts forward a scheme based on color recognition and shape recognition aimed at part's color characteristics and shape features of symmetrical parts.The tilt angle and location information of the parts can be get through process image's contour.For different color parts,we can quickly identify the parts' type by color recognition.After transforming the color space of the parts and calculating the HSV value of the unsaturated pixels,the type of the symmetrical parts can be identified.For the same color parts,through color to sort the parts is no longer applicable.At this time,image graying,image filtering,image edge extraction,and the smallest encircling rectangle are used to distinguish the left and right parts in the digital image processing.Output the four vertex coordinates of the smallest encircling rectangle of the part,and use the internal and external parameters of the camera,to convert the posture of the parts in the image coordinate system to the actual position and posture in the world coordinate system.In the experiment,a high precision CCD industrial camera is used to capture images under the condition of LED ring light source.OpenCV was configured to Visual Studio 2013 for processing the digital image.In this paper,the problem of sorting the plane mutual symmetry parts is solved by machine visual sorting method,and the position information of the smallest encircling rectangle is obtained to make the positioning precision of the parts reach the positioning precision requirement.