A method for solving target roll angle based on straight line vector space reconstruction
By decoupling yaw and pitch angles through a linear vector space reconstruction method and directly solving the roll angle, the problem of three-attitude angle coupling in optical attitude measurement is solved, improving the measurement accuracy and stability of medium- and long-range dynamic targets.
Patent Information
- Application Number
- CN202610439648.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-03
- Publication Date
- 2026-07-24
AI Technical Summary
Existing optical attitude measurement methods suffer from three-attitude angle coupling problems in measuring the roll angle of medium- and long-range dynamic targets, resulting in complex calculations and low accuracy.
A method based on linear vector space reconstruction is adopted. By selecting coplanar linear vectors, the characteristic planes before and after roll are reconstructed in the dual measurement camera intersection system, decoupling yaw and pitch angles, and directly solving for the roll angle.
It achieves independent and accurate calculation of roll angle, improves dynamic measurement accuracy and stability, reduces dependence on external benchmark calibration, and adapts to local occlusion conditions.
Smart Images

Figure CN122448247A_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the field of optical attitude measurement technology, specifically relating to a method for solving the target roll angle based on linear vector space reconstruction. Background Technology
[0002] Optical attitude measurement technology has wide applications in target tracking, aircraft navigation and landing, and space rendezvous and docking. Roll angle, as one of the key parameters describing a target's spatial attitude, is crucial for accurate measurement and attitude calculation and control. Currently, roll angle measurement methods are mainly divided into contact and non-contact types. Measurement methods based on principles such as laser collimation, interferometry, polarization, and grating diffraction are mostly suitable for laboratory environments or close-range research, and are difficult to meet the measurement needs of medium- and long-range dynamic targets. For roll angle measurement of medium- and long-range targets, existing technologies mainly rely on inertial sensor combinations for remote in-situ measurement, extracting parameters through error compensation and dynamic calibration. However, this type of method requires the sensors to be embedded in the target body and precise initial alignment before measurement, limiting its application scenarios.
[0003] Optical measurement methods possess unique advantages in target attitude measurement due to their non-contact nature, high visibility, and flexible station deployment. With a well-planned measurement station layout, optical methods can achieve high-precision attitude measurement during critical flight segments. Existing instantaneous roll angle measurement methods based on optical images typically employ a 3D-2D feature matching strategy, optimizing the solution of three-axis attitude angles by constructing an overdetermined system of equations. However, this type of method faces the following problems in practical applications: the target's spin motion leads to a strong coupling relationship between pitch, yaw, and roll angles. During trigonometric function calculations, the decoupling of attitude information is complex, and errors in the calculation of any attitude angle can propagate to other angles, affecting the overall measurement accuracy and stability. Therefore, how to effectively decouple the three attitude angles under dynamic measurement conditions and achieve a direct and accurate solution for the roll angle is a pressing technical problem in this field. Summary of the Invention
[0004] This application addresses the shortcomings of existing optical attitude measurement methods, such as the coupling of three attitude angles and the complexity of roll angle calculation, by providing a roll angle solution based on linear vector space reconstruction. This method selects coplanar linear vectors, reconstructs their space to create characteristic planes before and after roll, and transforms the roll angle into a dihedral angle between the planes, achieving complete decoupling from yaw and pitch angles and improving dynamic measurement accuracy.
[0005] To achieve the above technical objectives, this application specifically adopts the following technical solution: In one aspect of this application, a method for solving a target roll angle based on linear vector space reconstruction is provided, comprising the following steps: S1. Select two or more coplanar spatial target line vectors; S2. In the dual-camera intersection measurement system, extract the projected straight line vector of the selected straight line vector from the images of each substation; S3. For each projected line vector, spatial reconstruction is performed based on the intersection principle of two measurement cameras to obtain each spatial line vector. The direction vector in the measurement coordinate system; S4. Based on the reconstructed spatial line vectors and their directions, solve for the normal vector of the plane they define. ; S5. Based on the target's attitude information before rolling, determine the normal vector of the plane before rolling. ; S6. Calculate the normal vector. Relative to the normal vector The angle, as the target roll angle .
[0006] In one implementation, the two or more coplanar spatial target line vectors selected in step S1 include the target's central axis vector and another non-central axis line vector that is coplanar with the central axis.
[0007] In one implementation, the two or more coplanar spatial target line vectors selected in step S1 include the target's central axis vector or two other coplanar non-central axis line vectors.
[0008] In one implementation, step S3, which involves spatial reconstruction of the projected line vector, includes: S31: Project the linear vector under the substation image plane coordinate system Transform to the camera coordinate system of the corresponding substation to obtain the straight line direction vector in the camera coordinate system; S32: Based on the optical axis direction vector in the coordinate system of each substation camera With the projected linear vector Given the direction vectors, calculate the normal direction vector of the plane determined by the two vectors. ; S33: Convert the normal direction vector of each substation Transform to a unified measurement coordinate system; S34: Perform a cross product operation on the normal direction vectors of the two stations in the measurement coordinate system to reconstruct the straight line vector of the spatial target. The direction vector.
[0009] In one implementation, in step S31, the projected straight line vector in the substation image plane coordinate system... The direction vector is represented as ,in, From the image plane coordinate system The angle from the axis to the normal to the projected line, with a range of values of [value missing]. .
[0010] In one implementation, in step S31, the straight line direction vector in the camera coordinate system is represented as: .
[0011] In one implementation, in step S32, the optical axis direction vector in the coordinate system of the substation camera is... Represented as The normal direction vector Depend on Calculated.
[0012] In one implementation, in step S33, the normal direction vector The rotation matrix that transforms the camera coordinate system to the measurement coordinate system is the azimuth angle pointed to by the optical axis of that substation. and pitch angle Sure.
[0013] In one implementation, the normal vector of the front plane is rolled in step S5. The method for determining the yaw and pitch angles is as follows: based on the reconstructed centerline vector, the yaw and pitch angles are analyzed to obtain the pre-roll rotation matrix, and then the direction of the non-centerline characteristic line vectors before the roll is determined. The cross product of coplanar line vectors yields the pre-roll plane normal vector. .
[0014] The beneficial effects of this application are as follows: Compared to existing roll angle measurement methods that rely on sensor combinations or feature matching optimization, this application constructs feature planes before and after roll through linear vector space reconstruction, directly calculating the roll angle as the dihedral angle between the two planes. This achieves complete decoupling of roll information from pitch and yaw information, avoiding error propagation caused by attitude angle coupling. This method inherits the spatial extension characteristics of linear vectors, enabling reconstruction even under partial target occlusion conditions, thus improving adaptability to dynamic measurement environments. Furthermore, the method of determining feature planes based on dual-camera intersection reconstructed vectors reduces the dependence on precise external benchmark calibration, providing a stable technical approach for optical non-contact roll angle measurement. Attached Figure Description
[0015] Figure 1 This is a flowchart illustrating the method for solving the target roll angle based on linear vector space reconstruction in this application. Detailed Implementation
[0016] The technical solution of this application will be clearly and completely described below with reference to specific embodiments. However, those skilled in the art will understand that the embodiments described below are only some embodiments of this application, not all embodiments, and are only used to illustrate this application, and should not be regarded as limiting the scope of this application. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0017] Existing optical roll angle measurement techniques generally face the challenge of attitude angle information coupling when dealing with dynamic targets. The pitch, yaw, and roll angles are intertwined during trigonometric function calculations, and errors in any one angle affect the accuracy of the others, making the decoupling process complex and unstable. This application, based on the geometric invariance of line vectors in space, uses the target's centerline vector and wingspan vector as characteristic lines carrying roll information, and achieves spatial reconstruction of the line vectors through the intersection of two measurement cameras. Based on this, the pre-roll rotation matrix is analyzed using the reconstructed centerline vector to determine the pre-roll wingspan vector direction; the dihedral angle between the normals obtained from the cross product of the two characteristic vectors before and after roll is directly defined as the roll angle. This method transforms roll measurement into a geometric problem of the angle between two planes, eliminating coupling with other attitude angles and achieving independent and accurate solutions for roll parameters.
[0018] In one implementation, this application provides a method for solving the target roll angle based on linear vector space reconstruction, referring to... Figure 1 As shown, it includes the following steps: S1. Select two or more coplanar spatial target line vectors, including two methods.
[0019] The first method involves selecting the target's central axis vector and another non-central axis straight line vector coplanar with the central axis. The central axis or a straight line vector parallel to it is mandatory and is used to determine the rotation matrix before roll. For the selection of the feature plane after roll, if it includes the central axis or a straight line vector parallel to it, only one other straight line vector coplanar with the central axis is needed; this straight line vector varies with the roll angle.
[0020] The second method involves selecting the target's central axis vector or two other coplanar non-central axis line vectors. If no central axis vector is available, two coplanar non-central axis line vectors must be selected to determine the characteristic plane after roll. The change in this plane before and after roll reflects the change in the roll angle.
[0021] Specifically, taking two coplanar straight line vectors as the target's central axis vector and another non-central axis vector as examples, the non-central axis vector is the wingspan vector. The central axis vector is along the longitudinal axis of the target fuselage, and the wingspan vector is along the direction of the target's wing extension. The central axis vector and the wingspan vector are perpendicular to each other and coplanar when the target is not rolling. This plane is the target's principal symmetry plane.
[0022] When the target rolls, the spatial orientation of the centerline vector does not change with the roll angle, while the wingspan vector rotates around the centerline vector, and the two always remain coplanar. This plane changes its spatial attitude with the roll angle. By selecting the centerline vector and the wingspan vector as coplanar straight line vectors, and utilizing the solvability of the attitude change of the centerline vector before and after the roll, as well as the fixed geometric constraint relationship between the wingspan vector and the centerline vector, vector input is provided to obtain the plane normal vector before and after the roll.
[0023] If the roll feature plane does not contain a central axis vector, then two coplanar non-central axis line vectors are selected to determine the plane after roll. The change in this plane before and after roll reflects the change in the roll angle.
[0024] S2. In the dual-camera intersection measurement system, the projected straight line vector of the selected straight line vector is extracted from the images of each substation.
[0025] In some embodiments, a unified measurement coordinate system is established in a dual-camera intersection measurement system. The origin of the measurement coordinate system is selected as the midpoint of the dual-camera baseline. The X-axis points from camera one to camera two along the baseline direction, the Y-axis is perpendicular to the baseline and points towards the zenith, and the Z-axis is determined according to the right-hand rule. The transformation relationship between the camera coordinate systems of the two cameras and the measurement coordinate system is determined by the optical axis pointing parameters calibrated when the cameras are deployed. The optical axis pointing parameters include the azimuth angle. and pitch angle ,in Camera number, azimuth angle The pitch angle is the angle between the projection of the optical axis onto the horizontal plane and the X-axis of the measurement coordinate system. This is the angle between the optical axis and the horizontal plane. (Based on azimuth angle) and pitch angle Determine the rotation matrix to transform any vector from the camera coordinate system to the measurement coordinate system.
[0026] In some embodiments, the projection vector of the straight line vector is extracted from the images of each substation in the following manner.
[0027] For each measuring camera, acquire the target image captured by that camera. Identify and extract the projected line corresponding to a pre-selected straight line vector on the target from the image. The extraction method uses edge detection and line fitting to obtain the linear equation of the projected line in the image plane coordinate system. The image plane coordinate system has its origin at the top left corner of the image, with the positive X-axis pointing horizontally to the right and the positive Y-axis pointing vertically downwards.
[0028] For the extracted projected line, calculate the angle between its normal direction and the positive Z-axis of the image plane coordinate system. , The range of values is Based on the included angle Determine the direction vector of the projected line vector in the image plane coordinate system. The direction vector is expressed as: This direction vector represents the direction in which the projected line points within the image plane.
[0029] Map the direction vector of the projected line in the image plane coordinate system to the camera coordinate system of the corresponding measuring camera. In the camera coordinate system, the image plane is located at the focal length along the positive Z-axis, and the direction vector of the projected line in the camera coordinate system is expressed as: The 0 component corresponds to the X-axis direction of the camera coordinate system. Corresponding to the Y-axis direction, Corresponding to the Z-axis direction.
[0030] Repeat the above process to extract the projected line vector corresponding to the same line vector from the images of each measuring camera, and obtain the projection direction vector of each station.
[0031] S3. For each projected line vector, spatial reconstruction is performed based on the intersection principle of two measurement cameras to obtain each spatial line vector. To measure the direction vector in the coordinate system, follow these steps.
[0032] S31: Transform the projected straight line vector in the image plane coordinate system of the substation to the camera coordinate system of the corresponding substation to obtain the straight line direction vector in the camera coordinate system.
[0033] The direction vector of the projected straight line in the substation image plane coordinate system is represented as follows: ,in The angle from the Z-axis to the normal to the projected line in the image plane coordinate system, with a range of values of... After transforming to the camera coordinate system, the straight line direction vector is represented as... .
[0034] S32: Based on the optical axis direction vector in the coordinate system of each substation camera With the projected linear vector Given the direction vectors, calculate the normal direction vector of the plane determined by the two vectors. .
[0035] The optical axis direction vector in the coordinate system of the substation camera is represented as: The cross product of the optical axis direction vector and the projection line vector direction vector is the normal direction vector, calculated as follows: ,in The direction vector of the optical axis. The direction vector is the projection line vector.
[0036] S33: Convert the normal direction vector of each substation Transform to a unified measurement coordinate system.
[0037] The rotation matrix that transforms the normal direction vector from the substation camera coordinate system to the measurement coordinate system, based on the azimuth angle of the substation's optical axis. and pitch angle Confirmed. Azimuth. The pitch angle is the angle between the projection of the optical axis onto the horizontal plane and the X-axis of the measurement coordinate system. The angle between the optical axis and the horizontal plane, based on the azimuth angle. and pitch angle Construct a rotation matrix to convert the normal direction vector Transform to the measurement coordinate system to obtain the normal direction vector in the measurement coordinate system.
[0038] S34: Perform a cross product operation on the normal direction vectors of the two stations in the measurement coordinate system to reconstruct the straight line vector of the spatial target. The direction vector.
[0039] Let the normal direction vector corresponding to the first measuring camera be... The normal direction vector corresponding to the second measuring camera is The cross product of the two vectors is the direction vector of the spatial target line vector, expressed as: The components of this direction vector in the measurement coordinate system are: This completes the reconstruction of spatial line vectors.
[0040] S4. Based on the reconstructed spatial line vectors and their directions, solve for the normal vector of the plane they define. .
[0041] After spatial reconstruction, the direction vectors of two coplanar spatial line vectors in the measurement coordinate system are obtained. If the two spatial line vectors are respectively the target's central axis vectors... and wingspan vector Central axis vector The direction vector is represented as Wingspan vector The direction vector is represented as .
[0042] Central axis vector With wingspan vector A spatial plane is jointly defined, which represents the instantaneous position of the target's principal symmetry plane after the roll. The normal vector of this plane is calculated. This is achieved by the cross product of the central axis vector direction vector and the wingspan vector direction vector. The formula for the cross product is: .
[0043] The cross product operation can be expanded into the following component forms: The X component is ; The Y component is ; The Z component is .
[0044] The normal vector is obtained through the above calculations. The direction vector in the measurement coordinate system represents the direction of the normal to the principal symmetry plane of the target in space after the roll.
[0045] S5. Based on the target's attitude information before rolling, determine the normal vector of the plane before rolling. .
[0046] The yaw and pitch angles of the target are analyzed using the reconstructed centerline vector direction vector. The centerline vector direction vector in the measurement coordinate system is expressed as... The yaw angle is calculated based on this direction vector. and pitch angle Yaw angle The pitch angle is the angle between the projection of the direction vector onto the horizontal plane and the X-axis of the measurement coordinate system. It is the angle between the direction vector and the horizontal plane.
[0047] Based on yaw angle and pitch angle Construct the rotation matrix before roll Rotation matrix Characterizes the attitude transformation relationship of the target from the target body coordinate system to the measurement coordinate system before roll. Rotation matrix before roll. The specific form is:
[0048] in, Yaw angle about the Y-axis The rotation matrix, The pitch angle is the rotation around the Z-axis. The rotation matrix.
[0049] Assuming the target is not rolling, the initial direction of the wingspan vector in the target's body coordinate system is known. For an aircraft target, the wingspan vector along the Z-axis in the body coordinate system is represented as... The wingspan vector in the body coordinate system is rotated using the pre-roll rotation matrix. Transform to the measurement coordinate system to obtain the direction vector of the forward roll wingspan vector. The transformation formula is:
[0050] Direction vector of the center axis vector before roll In the measurement coordinate system, this is represented by the reconstructed central axis vector and its direction vector. The principal plane of symmetry of the target before roll is determined by the central axis vector. and wingspan vector Together, determine and calculate the normal vector of the plane. This is achieved by the cross product of the central axis vector direction vector and the wingspan vector direction vector. The formula for the cross product is:
[0051] S6. Calculate the normal vector. Relative to the normal vector The angle, as the target roll angle .
[0052] After roll, the normal vector of the plane The direction vector in the measurement coordinate system is represented as: The normal vector of the plane before roll The direction vector in the measurement coordinate system is represented as: The angle between the two normal vectors is the angle by which the target rotates around its central axis from its state before the roll to its state after the roll; this angle is defined as the roll angle. .
[0053] Roll angle It is obtained by calculating the dot product of the two normal vectors and their magnitudes. The calculation formula is as follows:
[0054] in, The dot product of two vectors, expanded as , and These are the magnitudes of the two vectors, respectively.
[0055] Calculated The value is expressed in radians or degrees and represents the instantaneous roll angle of the target as it rotates around its central axis.
[0056] Although the embodiments of this application have been described above in conjunction with the accompanying drawings, this application is not limited to the specific embodiments and application fields described above. The specific embodiments described above are merely illustrative and instructive, not restrictive. Those skilled in the art can make many other forms based on the guidance of this specification and without departing from the scope of protection of the claims of this application, and these are all within the scope of protection of this application.
Claims
1. A method for solving the target roll angle based on linear vector space reconstruction, characterized in that, Includes the following steps: S1. Select two or more coplanar spatial target line vectors; S2. In the dual-camera intersection measurement system, extract the projected straight line vector of the selected straight line vector from the images of each substation; S3. For each projected line vector, spatial reconstruction is performed based on the intersection principle of two measurement cameras to obtain each spatial line vector. The direction vector in the measurement coordinate system; S4. Based on the reconstructed spatial line vectors and their directions, solve for the normal vector of the plane they define. ; S5. Based on the target's attitude information before rolling, determine the normal vector of the plane before rolling. ; S6. Calculate the normal vector. Relative to the normal vector The angle, as the target roll angle .
2. The method for solving the target roll angle based on linear vector space reconstruction according to claim 1, characterized in that, The two or more coplanar spatial target line vectors selected in step S1 include a) or b): a) The target's central axis vector and another non-central axis straight line vector that is coplanar with the central axis; b) The target's central axis vector or two other coplanar non-central axis line vectors.
3. The method for solving the target roll angle based on linear vector space reconstruction according to claim 2, characterized in that, Step S3, which involves spatial reconstruction of the projected line vectors, includes: S31: Project the linear vector under the substation image plane coordinate system Transform to the camera coordinate system of the corresponding substation to obtain the straight line direction vector in the camera coordinate system; S32: Based on the optical axis direction vector in the coordinate system of each substation camera With the projected linear vector Given the direction vectors, calculate the normal direction vector of the plane determined by the two vectors. ; S33: Convert the normal direction vector of each substation Transform to a unified measurement coordinate system; S34: Perform a cross product operation on the normal direction vectors of the two stations in the measurement coordinate system to reconstruct the straight line vector of the spatial target. The direction vector.
4. The method for solving the target roll angle based on linear vector space reconstruction according to claim 3, characterized in that, In step S31, the projected straight line vector in the substation image plane coordinate system The direction vector is represented as ,in, From the image plane coordinate system The angle from the axis to the normal to the projected line, with a range of values of [value missing]. .
5. The method for solving the target roll angle based on linear vector space reconstruction according to claim 4, characterized in that, In step S31, the direction vector of the line in the camera coordinate system is represented as: .
6. The method for solving the target roll angle based on linear vector space reconstruction according to claim 3, characterized in that, In step S32, the optical axis direction vector in the coordinate system of the substation camera Represented as The normal direction vector Depend on Calculated.
7. The method for solving the target roll angle based on linear vector space reconstruction according to claim 3, characterized in that, In step S33, the normal direction vector The rotation matrix that transforms the camera coordinate system to the measurement coordinate system is the azimuth angle pointed to by the optical axis of that substation. and pitch angle Sure.
8. The method for solving the target roll angle based on linear vector space reconstruction according to claim 3, characterized in that, In step S5, the normal vector of the plane before rolling The method for determining the yaw and pitch angles is as follows: based on the reconstructed centerline vector, the yaw and pitch angles are analyzed to obtain the pre-roll rotation matrix, and then the direction of the non-centerline characteristic line vectors before the roll is determined. The cross product of coplanar line vectors yields the pre-roll plane normal vector. .
