Font Size: a A A

Research On Identification And Positioning Of Circular Mark Points In Machine Vision

Posted on:2019-03-13Degree:MasterType:Thesis
Country:ChinaCandidate:N N HeFull Text:PDF
GTID:2428330545477172Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
At present,artificial intelligence develops rapidly and is widely used in face recognition,robotics,gaming,intelligent control and machine vision.Humanity is moving toward an intelligent era.Industrial production intelligence depends on machine vision.Machine vision use machines instead of human eyes to complete functions such as identifying and positioning.Nowadays,machine vision is increasingly used in the manufacture of electronic products and high-precision inspection,especially for mount technology.It is a kind of circuit mounting technology that mounts assembled components on the surface of a printed circuit board and then assembles them by welding.With the miniaturization of electronic products,mounting components are getting smaller and smaller and more densely arranged,and the mounting accuracy has become higher and higher.Precise positioning of PCB boards is a prerequisite for accurate mounting.It is necessary to study the precise positioning method of the Mark point on the PCB board.This paper uses the improved RANSAC algorithm to detect circles.The main contents are as follows:(1)When RANSAC randomly selects points,the constraint condition is added,that is,the selected three points are greater than a certain distance,which promotes effective circulation.(2)Add a cyclic interrupt in the RANSAC iteration.Half of the total contour points are counted as the loop break condition.Once satisfied,the loop terminates.(3)Threshold optimization.Select the optimal shift value.(4)After RANSAC selects the model with the most votes,we do not see it as the best model,but instead use a point set that satisfies the model to do a least-squares fit.In this paper,the RANSAC algorithm is improved by the constrained random selection,cyclic interruption,threshold optimization,and least-squares optimization.The improved RANSAC achieves the positioning of the circular Mark point,making the accuracy of center detection achieve sub-pixel level.This paper takes the visual system on a solder paste printer as an example to verify the effectiveness of the algorithm.It has been verified that the improved RANSAC algorithm can achieve accurate positioning of circular Mark points.It is also time-consuming,robust,and universal for circular Mark detection.Compared with other traditional methods of detecting circles,the performance is improved a lot and has practical application value.
Keywords/Search Tags:machine vision, circular Mark point, positioning, RANSAC
PDF Full Text Request
Related items