Font Size: a A A

Research On Underwater Vehicle Binocular Vision Distance Measurement And Manipulator Visual Servo Control

Posted on:2016-04-12Degree:DoctorType:Dissertation
Country:ChinaCandidate:D WangFull Text:PDF
GTID:1318330542974122Subject:Ships and marine structures, design of manufacturing
Abstract/Summary:PDF Full Text Request
In recent years,AUV-underwater manipulator system(AUV Manipulator System-AUVMS)carrying binocular stereo vision system plays an important role in promoting the exploration and development of the marine environment.Based on binocular stereo vision system,AUVMS can complete the following work:the investigations and sample analysis of sea creatures and environmental mineral resources in deep sea,the inspection of the oil and gas pipelines,the inspection of the submarine cable,the realization of the underwater engineering,etc.In order to realize the autonomous grasping function of AUVMS system to the underwater object,this paper studied the underwater robot binocular vision ranging technology and manipulator vision control technology.The main research contents are as follows:Firstly,it introduced the main research contents of robot visual control technology and the classification of robot visual system,pointed out the developing status and trends of robot visual control technology,and expounded the manipulator research status at home and abroad.For the study of the calibration technology of underwater camera,it conducted the stereo calibration experiment under the state of still water and flow turbulence respectively,calculated the experimental data,and compared and analyzed the calibration results.It conducted the comparative analysis of the distortion of the camera with the distorted image under the state of still water and flow turbulence.And the analysis results show that the flow turbulence has important effect on underwater camera calibration,and points out the necessity of the camera calibration under the underwater operating environment.Meanwhile,it studied the influence of the change of the number of calibration board images,the position and pose of calibration board on the results of camera calibration.From changing the sitting position and pose of the underwater calibration board,it tried to reduce the camera calibration error,so as to get the rectifying image which is more close to the actual scene.Before using the underwater binocular stereo vision system to measure to the distance of the object,the underwater images should be preprocessed firstly.Based on Histogram equalization,the enhanced technology of underwater image effectively improved the quality of the original underwater image,and laid a good foundation for visual measurement.Binocular stereo vision measurement in the specific implementation process,mainly through its parallax,allowing underwater camera capable of specific location where the object is measured,and then give the information to restore depth and three-dimensional shape of the target object.This paper adopts BM algorithm,SGBM algorithm and SIFT algorithm respectively to conduct the stereo matching experiment of binocular images of underwater object,and generated the corresponding disparity map,so as to further analyze the advantages and disadvantages of the disparity maps generated by these stereo matching algorithms.According to the characteristics of less feature points detected by the underwater images,it improved SIFT algorithm parameters and determined the matching principal curvature coefficients and Gaussian kernel parameters which are suitable for underwater binocular parallax.And it studied the algorithm of the conversion matrix between the camera coordinate system and the manipulator coordinate system,and put forward the ontology arm that uses the underwater manipulator.Based on geometry algorithm,combined with the separation algorithm between the position and pose of the manipulator,it calculated and calibrated the transformation matrix of the camera coordinate system relative to the manipulator coordinate system.In this paper,the underwater manipulator adopted the location-based visual control method and based on the underwater object image information to calculate 3D coordinates of underwater object in manipulator coordinates,through the solution of inverse kinematics to obtain the given positional value of manipulator revolute joint,and control the movement of underwater manipulator to the target location.In experimental pool,the measurement information based on the visual system completed the experiment of controlling and guiding the underwater manipulator to the target location of the part and grabbing the object,and verified the vision measurement method of underwater object and the vision control method of underwater manipulator.The experimental results show that the grasping task of underwater manipulator to underwater object can be realized with the distance measuring method and control method.The control scheme of the underwater object ranging and the underwater manipulator proposed in this paper is effective and feasible.
Keywords/Search Tags:underwater manipulator, visual control, vision measurement, stereo matching, underwater camera calibration
PDF Full Text Request
Related items