Research On Hybrid Of Vision And Inclination Sensor Method For Estimating Relative Pose

The pose information of the object is the important parameters to reflect the space object,With the rising demand of pose measurement,pose measurement is becoming more and more important and has attracted people's attention,which has been widely used in industrial,military,large equipment manufacturing,aerospace and other fields.The traditional method of measuring the relative position and orientation from the principle can be divided into four categories: pose measurement method based on laser tracking systems,pose measurement method based on inertial navigation,pose measurement method based on mechanical positioning constraints principles,pose measurement method based on computer vision systems.To deals with the problem of estimating the relative pose between a camera and an object,this paper proposed a hybrid of vision and inclination sensor method for estimating relative pose.It is assumed that both the camera and the object are equipped with an inclination sensor,the resulting pose estimation problem can be seen as a Pn P(Perspective-n-Point)problem with inclination constraint.First,this paper introduces the principles of the proposed method and the structure parameter calibration.In order to improve the algorithm accuracy and noise immunity,further nonlinear optimization can be done on the basis of linear.Simulations results certify that the proposed method can improve the computing accuracy and depress the image noise comparing with other solutions to the Pn P problem.Finally,the experimental results show that the average re-projection errors of x and y directions for 15 feature points are less than 0.2 pixels by using the proposed method,which can meet the needs of pose measurement.The main research work and innovations are as follows:(1)Widely study relevant research literature at home and abroad,analysis the significance and research status of pose measurement,elaborate the research value and significance of this topic.(2)To deals with the problem of estimating the relative pose between a camera and an object,this paper study a hybrid of vision and inclination sensor method for estimating relative pose.The resulting pose estimation problem can be seen as a Pn P problem with camera and inclination information.Derive the principle of combined pose measurement and design the measurement algorithm in detail.In order to improve the algorithm accuracy and noise immunity,further nonlinear optimization can be done on the basis of linear.Set up simulation experiments and certify the correctness of the combined measuring principle.(3)The paper summarizes the basic theory and method of camera calibration,derive the principles of inside and outside parameters calibration according to the camera imaging model,complete the corresponding algorithms and simulation analysis.(4)Study the structure parameters calibration method according to the principle of the combined pose measurement,derive the principle of the structure parameters calibration.Set up simulation experiments and certify the correctness of the structure parameters calibration principle.(5)Design and write the corresponding algorithms of combined pose measurement,set up simulation experiments and analysis influences of the measurement factors.