| Automatic docking system plays an important role in assembly of largecomponents.Test location technology is the key technology in automatic docking system,and it directly affects the docking results of assembly.In the detection of positioningtechnology, computer vision measurement is a wide prospect of application of testingtechnology.Through test simulation system, this paper studies the camera calibration, imagefeature extraction technology and position detection technology, etc., and is verified.Camera calibration can get the camera’s internal parameters and external parametersand the coefficient of optical distortion. It is the premise of the visual measurement.Comparative analysis of the current main calibration method, including the advantages anddisadvantages of the traditional calibration method and self-calibration method, this paperchooses Zhengyou Zhang calibration method which between the two method to completethe camera calibration. The calibration method is easy to implement, and the precision ofcalibration results is higher.Image feature extraction is the necessary process of image analysis, and the imagefeature extraction methods directly affect the accuracy of detection. Based on the analysisof the line feature extraction technique, especially of the characteristics of ellipse fittingtechnology, this paper compares the various ellipse fitting method, and puts forward amethod of combining RHT and the direct least square ellipse fitting.Position detection technology can get the position parameters of the object, and is thefoundation of automatic docking.This paper mainly studies the classic PNP problem,especially the P3P problem and the P4P problem.The solving method, which is based on theP4P problem of4o ’clock coplanar, can get the position parameters of the target object.In the experimental system, the camera calibration, feature extraction and targetposition of the object computing experiment are completed, and the feasibility of thealgorithm is verified. |
PDF Full Text Request