CN-116929305-B - Monocular vision space non-cooperative target relative pose measurement method and system based on feature constraint set

CN116929305BCN 116929305 BCN116929305 BCN 116929305BCN-116929305-B

Abstract

A monocular vision space non-cooperative target relative pose measurement method and system based on a feature constraint set relate to the technical field of spacecraft relative navigation and pose estimation. The method aims to solve the problem of short-distance relative navigation in the task processes of on-orbit maintenance of a space non-cooperative target, active clearance of space fragments, capture of a non-cooperative target satellite and the like. The method comprises the steps of carrying out dynamic modeling on relative motion of a tracking spacecraft and a target spacecraft, establishing a measurement model, extracting target feature points, establishing a feature constraint equation based on a target feature constraint set, selecting parameters to be estimated, including non-cooperative target pose parameters and inertial parameters, expanding Kalman filter design, estimating pose parameters and inertial parameters, and designing a space non-cooperative target pose and inertial parameter measurement system. The invention adopts a monocular camera as a measuring sensor, extracts target feature points, establishes various constraint relations among target features to form a feature constraint set, inputs the constraint set as pseudo measurement into a filter to improve the observed information quantity, and finally realizes the rapid determination of the motion parameters and the inertia parameters of a space non-cooperative target.

Inventors

ZHANG ZEXU
CHEN XIANGGUI
GUO PENG
YUAN SHUAI
SU YU
WANG YISHI
YUAN MENGMENG

Assignees

哈尔滨工业大学

Dates

Publication Date: 20260512
Application Date: 20230713

Claims (8)

1. A monocular vision space non-cooperative target relative pose measurement method based on a feature constraint set comprises the following steps: Defining a reference coordinate system and a geocentric equatorial inertial coordinate system The origin is located at the centroid of the earth, The axis points to the direction of the spring point; The axis points to the direction of the north pole of the earth; The shaft is determined according to the right hand rule; LVLH coordinate system Tracking the center of mass of the spacecraft as the origin of a coordinate system, The axis points to the direction of the mass center of the tracking spacecraft from the earth center; the axis being in the plane of the track, perpendicular to The shaft and the included angle between the shaft and the speed direction are acute angles; The shaft is determined according to the right hand rule; tracking spacecraft body coordinate system Tracking the coordinate axis of the spacecraft body coordinate system and the inertia principal axis to coincide by taking the center of mass of the tracked spacecraft as the origin of the coordinate system; Target spacecraft body coordinate system The system takes the mass center of the target spacecraft as the origin, and assumes that the coordinate system coincides with the target principal axis of inertia, and takes the principal axis of minimum inertia as the principal axis of inertia The spindle with maximum inertia is The axis of the shaft is provided with a plurality of grooves, The shaft is determined according to the right hand rule; camera coordinate system The system uses the optical center of the camera as the origin of coordinates and the optical axis of the camera as the optical axis The axis of the shaft is provided with a plurality of grooves, The axis of the shaft is provided with a plurality of grooves, The axis is parallel to the image plane; Dynamic modeling is carried out on the relative motion of the tracking spacecraft and the target spacecraft; Constructing a measurement model, extracting target feature points, constructing target feature constraints so as to establish a feature constraint equation, and searching various constraint relations among target features based on the monocular vision measurement model to establish a feature constraint set equation; Based on a plurality of constraint relations among target features, a feature constraint set equation is established, the constraint equation is added into a measurement equation, measurement information measurement is improved, and therefore target estimation accuracy is improved; the method comprises the steps of selecting parameters to be estimated, including non-cooperative target pose parameters and inertial parameters, designing and expanding a Kalman filter based on different characteristic constraint measurement models, and estimating the target pose parameters and the inertial parameters; based on target characteristics, estimating pose parameters and inertial parameters of the spatial non-cooperative targets, and finally designing a spatial non-cooperative target pose and inertial parameter measurement model, so as to realize monocular vision spatial non-cooperative target relative pose measurement based on characteristic constraint sets; A measurement model based on monocular vision is established, target feature points are extracted, and a feature constraint set equation is established for various constraint relations among target features, and is specifically as follows: (1) Assuming that the projection center of the camera coincides with the center of mass of the tracking spacecraft, selecting a target The coordinates of the characteristic points are taken as observed measurement values, and the following steps are carried out on the tracking spacecraft body system: (7) wherein: Tracking a position vector of a target characteristic point under a spacecraft body coordinate system; tracking a vector from a spacecraft centroid to a target centroid under a spacecraft body tracking coordinate system; The position vector of the target feature point under the target body coordinate system; Target feature points; (2) Assume the projection coordinates of the target feature points on the image Is that The target feature point is in the coordinate of the target body coordinate system Is that The target feature point is co-ordinate in the camera co-ordinate system Is that The target feature point vector under each camera coordinate system is transferred to the tracking spacecraft body coordinate system and expressed as: , The method comprises the steps of obtaining a rotation matrix from a camera coordinate system to a tracking spacecraft body system, wherein the relation between target characteristic points and image projection coordinates in the camera coordinate system is as follows: according to the principle of projective geometry, the first Characteristic points Coordinates in the image are: (8) image coordinates Expressed in pixel units as: (9) Wherein the method comprises the steps of For the focal length of the camera, Is the pixel coordinates of the principal point, And Representing horizontal and vertical pixel density, respectively ; (3) The system observation equation is , For measuring sensor noise by equation expression Representative of A different pseudo-measurement equation, then the constraint set observation equation may be expressed as: (10) The feature constraint set equation is established by the constraint relations among the target features, namely, when part of features in the feature points are collinearly constrained, collinearly is established Constraint equation, assuming For the mass center of the object, characteristic points , .... , Collinear, feature point corresponding vector is , When (when) In the time-course of which the first and second contact surfaces, The collinearly constrained equation is: (11) The feature constraint set equation is established by the constraint relationships among the target features, and when part of the features in the feature points are the circle constraint equation, the co-circle is established Constraint equation, assuming For the mass center of the object, characteristic points , .... Coplanar and round, with a radius of ; The circle constraint equation is: (12) The feature constraint set equation is established by the constraint relations among the target features, namely, when part of features in the feature points are coplanar constraint equations, the coplanarity is established The constraint equation is set to be a function of, , , , For the coplanar four characteristic points on the target, any three points in the four characteristic points are taken to form two planes, , Normal vectors of two planes respectively; the coplanar constraint equation is: (13) the feature constraint set observation equation obtained by equation (10-13) is: (14) and establishing a characteristic constraint set equation for a plurality of different constraint relations among target characteristics, and inputting a constraint set as pseudo measurement into a filter for improving the observed information quantity so as to realize quick determination of motion parameters and inertia parameters.
2. The method for measuring the relative pose of a monocular vision spatial non-cooperative target based on a feature constraint set according to claim 1, wherein the target is a spatial non-cooperative target spacecraft.
3. The monocular vision space non-cooperative target relative pose measurement method based on the feature constraint set according to claim 1, wherein the dynamic modeling is performed on the relative motion of a tracking spacecraft and a target spacecraft under an inertial coordinate system, and the method is specifically as follows: According to Newton's second law, under the condition of no disturbance, the motion equations of the tracking spacecraft and the target spacecraft are respectively: (1) Wherein the method comprises the steps of Is the constant of the gravitational force, 、 The mass center position vectors under the inertia system respectively, and the sizes of the mass center position vectors of the target spacecraft and the tracking spacecraft can be expressed as follows: , is a half-long axis of the machine, In order to achieve the eccentricity ratio, Is true near angle, define Under the orbit coordinate system, the relative translational dynamics equation of the target spacecraft relative to the tracking spacecraft is as follows: (2) Wherein the method comprises the steps of Respectively is Is used for the first derivative of (c), For the angular velocity of the orbital coordinate system in inertial space, Is angular acceleration; relative attitude transformation matrix of target spacecraft relative to tracking spacecraft using euler quaternions Representing a gesture quaternion ; For a rotation matrix of the object specimen system to the tracking coordinate system, Expressed as: (3) Wherein the method comprises the steps of Is vector quantity Is used for the matrix of the anti-symmetry of (a), ; The attitude kinematics of the target spacecraft relative to the tracking spacecraft are represented by quaternions: (4) Wherein: , Relatively tracking an angular velocity vector of the spacecraft for the target spacecraft; in an orbital coordinate system, relative angular velocities of a target spacecraft and a tracking spacecraft Expressed as: (5) The rotation matrix from the target spacecraft coordinate system to the tracking spacecraft coordinate system, and the relative tracking spacecraft rotation dynamics equation of the target spacecraft: (6) wherein: 、 The moment of inertia matrixes of the target spacecraft and the tracking spacecraft are respectively; 、 External moment of the target spacecraft and the tracking spacecraft respectively; Angular velocity of a target spacecraft in a body coordinate system; Tracking the angular velocity of the spacecraft under an orbit coordinate system; angular velocity of the target spacecraft relative to the tracking spacecraft.
4. The monocular vision spatial non-cooperative target relative pose measurement method based on characteristic constraint set of claim 1, wherein the spatial non-cooperative unknown target is selected, and the parameters to be estimated comprise non-cooperative target inertial parameterization and estimated inertial matrix, and the method is specifically as follows: Assuming that the establishment of a target spacecraft body coordinate system is consistent with the principal axis of inertia, the rotational inertia of the target spacecraft is made to be , , The inertia matrix is not completely observable under the condition of no torque movement of the target, and only two out of three degrees of freedom are observable, so that the inertia comparison method is adopted, and the method has two degrees of freedom corresponding to two random variables, so that , And simplifying the diagonal moment of inertia matrix to obtain a scale factor expression: (15) Wherein: 。
5. The method for measuring the relative pose of the monocular vision spatial non-cooperative target based on the feature constraint set according to claim 4, wherein the design of the extended Kalman filter based on different feature constraint measurement models is as follows: Defining state vectors for extended Kalman filters Discretizing a relative motion dynamics state equation (2) of the target spacecraft, discretizing a posture kinematics equation and posture dynamics equations (4) and (6) of the target spacecraft, and designing an extended Kalman filter according to an extended Kalman filter theory, wherein the steps are as follows: The nonlinear state equation and the observation equation according to the above can be written as: (16) respectively linearizing the nonlinear equation to obtain: , Initializing: , State prediction: (17) covariance matrix prediction: (18) EKF gain matrix prediction: (19) and carrying out state updating according to the state prediction and the filtering gain, and rapidly and accurately estimating the relative position, speed, attitude, angular speed and inertia ratio information of the non-cooperative target through iteration.
6. The monocular vision spatial non-cooperative target relative pose measurement method based on characteristic constraint sets according to claim 1 or 5, wherein characteristic constraint set equations are established for different constraint relations among target characteristics, and the constraint sets are input into a filter as pseudo measurement to improve the observed information quantity and the estimation precision, so that the spatial non-cooperative target pose estimation and the rapid determination of inertial parameters are realized.
7. A monocular vision space non-cooperative target relative pose measurement system based on a feature constraint set is characterized in that the system is provided with a program module corresponding to the steps of any one of claims 1 to 6, and the steps in the monocular vision space non-cooperative target relative pose measurement method based on the feature constraint set are executed in running.
8. A computer readable storage medium, characterized in that it stores a computer program configured to implement the steps of a monocular vision spatial non-cooperative target relative pose measurement method based on feature constraints set as claimed in any of claims 1-6 when called by a processor.

Description

Monocular vision space non-cooperative target relative pose measurement method and system based on feature constraint set Technical Field The application relates to the technical field of relative navigation and pose estimation of spacecrafts, in particular to a method and a system for measuring relative pose of a monocular vision space non-cooperative target of a feature constraint set. Background The spacecraft relative pose and inertial parameter measurement technology has important significance in space task scenes, and the space non-cooperative target pose and inertial parameter estimation plays an important role in task implementation when tasks such as on-orbit maintenance, active space debris removal, enemy satellite capturing damage and the like are carried out on the space non-cooperative target fault spacecraft under the condition that a space non-cooperative target model is unknown. In the measurement of relative pose and inertial parameters, the visual navigation cost is relatively low, so that the method is widely applied to the relative navigation of the spacecraft. Compared with a laser radar, the monocular camera has the advantages of small quality, low power consumption and economic cost, relatively simple monocular vision system configuration, simple monocular vision structure, easy calibration and small platform area occupied by the monocular vision compared with stereoscopic vision, and based on the vision navigation characteristics, students at home and abroad have conducted extensive researches. At present, the research of the visual navigation technology at abroad has advanced, italian scholars Vincenzo Pesce and the like, and the pose and inertial parameters of a space non-cooperative target are estimated by adopting a stereoscopic vision measurement combined filtering method, but the space position of a platform occupied by a binocular camera for stereoscopic vision is larger. American scholars Sean Augenstein, etc., use monocular vision SLAM algorithm for pose tracking and shape reconstruction of unknown targets, but do not estimate the target inertial parameters. Domestic scholars Ge Dongming and the like adopt a stereoscopic vision measurement method, three characteristic points on a target are selected as measurement observance, and pose and inertial parameter information of a space non-cooperative target are estimated by combining a filtering algorithm, but the method is low in convergence speed, and can not realize pose estimation of the space non-cooperative target and rapid determination of inertial parameters. Disclosure of Invention The invention aims to solve the technical problems that: Aiming at the problems of short-distance relative navigation in the task processes of on-orbit maintenance of a space non-cooperative target, active clearance of space fragments, capturing and damaging enemy satellites and the like, the invention provides a monocular vision space non-cooperative target relative pose measurement method and system based on a feature constraint set when pose estimation and inertia parameters of the space non-cooperative target are required to be quickly determined in the intersecting process. The technical scheme adopted by the invention for solving the technical problems is as follows: A monocular vision space non-cooperative target relative pose measurement method based on a feature constraint set comprises the following steps: defining a reference coordinate system; modeling the relative kinematics and dynamics of the tracking spacecraft and the target spacecraft; based on the monocular vision measurement model, searching different constraint relations among target features to establish a feature constraint set equation; And establishing a characteristic constraint set equation for different constraint relations among the target characteristics, adding the constraint equation into a measurement equation, and improving measurement information measurement, thereby improving target estimation accuracy. And designing and expanding a Kalman filter based on different characteristic constraint measurement models to estimate the target motion parameters and the inertia parameters. And finally, designing a space non-cooperative target pose and inertial parameter measurement system (model) based on target feature filtering estimation on the space non-cooperative target pose parameters and inertial parameters, thereby realizing monocular vision space non-cooperative target relative pose measurement based on feature constraint sets. Further, the spatial targets are unknown non-cooperative. Further, the reference coordinate system is defined as follows: The earth center equator inertial coordinate system (I system) is characterized in that the origin is positioned at the earth centroid, the X I axis points to the direction of the spring point, the Z I axis points to the direction of the earth north pole, and the Y I axis is determined according to the right hand rule