Research And Improvement Of Camera Calibration And Characteristic Points Extraction In Binocular Vision

In recent years,vision measurement technology has developed rapidly,which is widely used in many fields,and is getting closer to our lives.Binocular vision is an important branch of visual measurement technology.Theoretical research on binocular vision provides technical support for industrial manufacturing,commercial activities and new technology generation.Camera calibration technology is the most important part of binocular vision system.Camera calibration accuracy The quality of the entire vision system directly affects the measurement accuracy of the entire vision system.It is important to study the camera calibration technology to improve the overall measurement accuracy of the vision measurement.In the traditional camera calibration process,the calibration block is often used as the structural information of the known scene,and most of the feature points on the calibration block or the calibration target are used as reference points,and the center position is accurately extracted through the change of the gray scale information.The center position extracted at this time can be regarded as a feature point,and the camera calibration is performed using the information of the extracted characteristic points.In order to simplify the camera calibration process,this paper presents a flexible calibration method which is simple in principle and easy to implement.This method is based on Hough algorithm to extract calibration points,and achieve high accuracy.During the traditional calibration of cameras,calibration blocks are often used as the structural information of a known object.In most cases,the characteristic point on calibration block or calibration target is taken as the reference point,to accurately extract the center through the change of grey information.The accuracy of characteristic point extraction greatly affects the whole calibration accuracy.Usually,an error of a pixel leads to a totally different calibration result.In the traditional calibration,a checkerboard is often taken as the calibration pattern.The checkerboard calibration is the corner detection based on gray images,binary images,and contour curves.For checkerboard pattern,during the characteristic point extraction,it is easy to extract pseudo corners,to affect the whole calibration accuracy.To extract characteristic points with higher accuracy,the characteristic point extraction method can be further optimized.In this thesis,the calibration pattern with crossed sine fringe pattern is used and the improved phase extraction method is adopted,to improve the positioning accuracy of pattern characteristic points,thus improving the calibration accuracy.