Font Size: a A A

On Monocular Vision Measurement Technique Based On Single Photo

Posted on:2018-08-12Degree:MasterType:Thesis
Country:ChinaCandidate:Y Q WangFull Text:PDF
GTID:2348330563451251Subject:Geodesy and Survey Engineering
Abstract/Summary:PDF Full Text Request
Determining of the objects' position and orientation has a wide range of applications in the fields of large-scale equipment manufacturing.Mono-vision measuring system is an essential consisting part of vision measuring regin.Because of its simplicity,high precision,it has attracted many reseachers' interest.The research subject of this paper is vision measurement technology based on single photo.the research is mainly around the key technology of object's position and orientation parameter measurement based on the single photo,including the modeling of measurement model,the calibration of light pen and calculation of position and orientation parameter based on point feature.This paper designs and develops a complete position and orientation measurement system with single camera,which includes hardware and software.The system has the advantages of good stability,widely uses and high accuracy.The paper covers the following aspects:1.The measurement model of the system is constructed.Initial value of pose parameter is calculated by space resection,based on four non collinear control points,the single image space resection of least square iterative based on Lodrigues matrix is proposed,the system measurement model is built by coordinate transformation theory.2.In order to calibrate light pen for single-photo measurement system,a complete on-site light pen calibration method is proposed in this paper.According to the measuring principle the structure of light pen,material and the type,size and number of control points are analyzed,a new type of light pen with practical application are designed.The target point coordinate and the probe tip center coordinate are calibrated based on the camera of this system itself,the tripod,two calibration bars and one calibration sphere.Firstly,the true coordinates of target points on the light pen are implemented using two calibration bars;secondly,the location of the probe tip is not changed with the calibration sphere rotation,mathematical model of light pen probe calibration is built,according to the least square principle the coordinate of probe tip is acquired.3.The coordinate measuring technique based on single photo are improved.The transformation model between light pen and camera coordinate system is established based on single image space resection,the transformation parameters are solved accurately using the least square iteration principle;the use of retro reflection markers with flash photography to obtain high quality image point coordinates,It improves the precision of the mark point extraction and effectively increases the photography distance.In order to verify the measurement accuracy of the system,the indoor three-dimensional control network which is suitable for multisensor is built based on the visibility and the camera distance.4.A single photo pose measurement technique based on point feature is optimized.The rotation and translation parameters of the camera and the reference target and the target are solved linearly by decomposing the rotation matrix.take camera coordinate system as bridge,a linear formula for position and orientation is established.Aiming at the problem of target pose measurement based on texture information missing,The mark points instead of LED planar target are on the surface of the target5.Pose measurement system based on single photo is developed.With 2.4 m antenna positions as experimental object to simulate the FAST feed real-time position measurement,in the range of 6 meters,system outside the coordinate measuring precision is better than 0.1 mm,outside the Angle measurement precision is better than that of 20 ".
Keywords/Search Tags:single photo, 3D coordinate measurement, pose measurement, single image space resection, light pen calibration, coordinate transformation
PDF Full Text Request
Related items