A servo star sensor structure parameter calibration method

Abstract

The application provides a kind of servo star sensor structure parameter calibration method, only theodolite is used, the conversion relationship of the prism reference system of star sensor and theodolite reference system is established, and the calibration of the structure parameter of star sensor is realized.Theodolite obtains the representation vector of three-axis of prism coordinate system under theodolite coordinate system using the autocollimation function of itself, and the representation vector of prism coordinate system is converted to the fixed theodolite coordinate system by mutual sighting with fixed theodolite;Theodolite obtains the miss-target vector by aiming function, converts it into the incident vector under the camera reference system of star sensor, and uses the same method as theodolite for measurement, the conversion relationship between the two reference systems is established by mutual sighting with fixed theodolite, the representation vector of crosshair target under the reference system of aiming theodolite is converted to the fixed theodolite reference system, and then converted to the incident vector of prism reference system, and finally the structure parameter is estimated according to the incident vector of two reference systems.

Description

TECHNICAL FIELD

[0001] The present application relates to the absolute attitude measurement technical field, and in particular to a kind of servo star sensor structure parameter calibration method. BACKGROUND

[0002] Servo star sensor is used to solve the problem of all-weather observation when ground astronomical position measurement is carried out.The introduction of servo makes star sensor appear multiple reference systems associated with structural features, such as camera reference system, benchmark prism reference system, and celestial observation vector must be converted through these reference systems step by step to carry out position calculation.Therefore, the accurate calibration of structural parameters characterizing the conversion relationship is the guarantee of the measurement accuracy of servo star sensor.

[0003] At present, the common method is to ensure that the installation precision meets the requirements during structure processing and assembly.Another calibration method uses the pointing observation of star sensor to external celestial bodies and the measurement results of theodolite to the actual attitude of star sensor after astronomical north-seeking to optimize the estimation of structural parameters.However, this method cannot be completed in indoor conditions, requires many conditions, and introduces many errors: first, if the installation precision of processing and assembly is not enough, the target celestial body is likely to be out of the field of view due to too large off-target amount;second, the number and distribution position of available celestial bodies during calibration process cannot be controlled, thereby the calibration accuracy of structural parameters cannot be guaranteed;in addition, due to the existence of atmospheric obscuration and turbulent disturbance, these random errors are directly introduced into the calibration results during calibration process.In addition, in order to find suitable celestial bodies, the experimental period needs to be accurately predicted in advance, and the observation process will also be affected by weather, so the calibration efficiency is low.

[0004] Therefore, the current structural parameter determination method of star sensor can improve the calibration accuracy, avoid the restriction and influence of environment, installation and observation conditions. SUMMARY

[0005] Therefore, the present application provides a servo star sensor structure parameter calibration method, which can calibrate the structural parameters of star sensor in laboratory, solve the problems of dependence on installation precision, existence of atmospheric disturbance and observation weather, improve the calibration efficiency and calibration accuracy of structural parameters, and reduce the complexity of calibration equipment.

[0006] To achieve the above application purposes, the technical solutions of the present application are as follows:

[0007] A calibration method of servo star sensor structure parameters, the specific steps include:

[0008] Step one, a fixed theodolite, an aiming theodolite and two measurement theodolites are arranged around the star sensor, and the two measurement theodolites are respectively aligned with two orthogonal side surfaces of the benchmark prism of the star sensor.

[0009] Step two, two measuring theodolites measure two axes of the reference system of the prism perpendicularly, and the two axes are characterized in the reference system of the measuring theodolite; the measuring theodolite and the fixed theodolite mutual sighting, and the conversion matrix of the fixed theodolite and the reference system of the prism is obtained.

[0010] Step three, the aiming theodolite is moved, so that the crosshair target is imaged on the image plane of the star sensor, and the crosshair target is changed into a characterization vector in the reference system of the aiming theodolite; the aiming theodolite and the fixed theodolite mutual sighting, and the conversion matrix of the fixed theodolite and the reference system of the aiming theodolite is obtained, and the conversion matrix of the aiming theodolite and the reference system of the prism theodolite is obtained by combining the conversion matrix in step two, and the characterization vector of the crosshair target is converted into an incident vector in the reference system of the prism.

[0011] Step four, the imaging center position of the crosshair target is extracted, and a miss-target vector is obtained.

[0012] Step five, the miss-target vector is converted into an incident vector in the camera reference system of the star sensor.

[0013] Step six, the structure parameters are calibrated according to the incident vectors in step three and step six.

[0014] Further, the specific process of measuring two axes of the reference system of the prism perpendicularly by two measuring theodolites is as follows:

[0015] The two orthogonal prism sides of the star sensor are measured perpendicularly by the first theodolite and the second theodolite, the azimuth angle α Xp and the pitch angle β Xp of the first theodolite are obtained as the characterization vector of the X p axis of the prism coordinate system O p -X p Y p Z p in the first theodolite coordinate system, and the azimuth angle α Yp and the pitch angle β Yp of the second theodolite are obtained as the characterization vector of the Y p axis of the prism coordinate system O p -X p Y p Z p in the second theodolite coordinate system.

[0016] Further, the specific process of mutual sighting of the measuring theodolite and the fixed theodolite to obtain the conversion matrix of the fixed theodolite and the reference system of the prism is as follows:

[0017] The azimuth angle α Xp is converted to the fixed theodolite reference system O s -X s Y s Zs O s -X s Y s Z s X p azimuth angle of the X Xp axis; based on the azimuth angle α Xp and the pitch angle β Xp , the representation vector of the X p axis in the fixed theodolite reference system is calculated

[0018] Convert the azimuth angle α Yp to the fixed theodolite reference system O s -X s Y s Z s O s -X s Y s Z s Y p azimuth angle of the Y Yp axis; based on the azimuth angle α Yp and the pitch angle β Yp , the representation vector of the Y p axis in the fixed theodolite reference system is calculated

[0019] Based on the representation vector and the representation vector , the vector representation of the Zp axis in the fixed theodolite coordinate system is obtained by using the cross product Further, the conversion matrix of the prism reference system to the fixed theodolite reference system is obtained

[0020]

[0021] Further, the representation vector in the conversion matrix is modified again by using the cross product That is

[0022]

[0023] Then, the value of the cross product modification is assigned to

[0024] Further, the imaging center position of the crosshair target is extracted, and the specific way of obtaining the miss distance vector is as follows:

[0025] The crosshair target image is subjected to gray uniformization and lifting, so as to highlight the crosshair target.

[0026] ​The Canny edge extraction algorithm is used to extract the crosshair target edge, and four straight line edges are obtained.

[0027] The Hough straight line transformation is used to obtain the parameters of the four straight line edges.

[0028] The four intersection points of the straight line equations are solved.

[0029] The mean of the four intersection points is the center position of the crosshair.

[0030] The position of the principal point in the optical imaging model of the star sensor is denoted as The miss distance vector is and is

[0031]

[0032] Further, the specific way of calibrating and estimating the structural parameters is as follows:

[0033] The incident vector obtained in step three and step six is substituted into formula (8) to calibrate and estimate formula (8), so as to obtain the installation error conversion matrix between the prism reference system and the servo zero position reference system and the installation error conversion matrix between the servo real-time generated position and the camera coordinate system wherein formula (8) is:

[0034]

[0035] wherein, is the incident vector of the camera reference system; is the incident vector of the prism reference system; is the servo rotation matrix, which is obtained from the servo code disc.

[0036] Further, the least square method is used to calibrate and estimate formula (8).

[0037] Further, the normal measurement is performed by using the positive and negative mirror method.

[0038] Beneficial effects:

[0039] 1. The application provides a kind of servo star sensor structure parameter calibration method, only theodolite is used, the conversion relationship of the prism reference system of star sensor and theodolite reference system is established, and the calibration of star sensor structure parameter is realized.Theodolite uses its autocollimation function, obtains the representation vector of three-axis in theodolite coordinate system in prism coordinate system, and is converted to fixed theodolite coordinate system by mutual sighting with fixed theodolite;Theodolite is obtained by aiming function, and is converted into incident vector in camera reference system of star sensor, and the same method as measurement theodolite is used, and the conversion relationship between the reference system is established by mutual sighting with fixed theodolite, the representation vector of crosshair target in the reference system of aiming theodolite is converted to fixed theodolite reference system, and then is converted into incident vector in prism reference system, and finally the structure parameter is estimated according to the incident vector of two reference systems.

[0040] 2. The application adopts positive and negative mirror method, repeatedly measures the normal of reference prism, and improves measurement precision.

[0041] 3. In the application, the edge of crosshair target is removed when extracting the center position of crosshair target, and the mean value of multiple positions is solved as the center position, so as to improve measurement precision.

[0042] 4. The method of the application realizes the calibration of structure parameter in laboratory by theodolite, avoids the strict control requirement of installation error during the assembly of star sensor, saves the prediction calculation of star map in field experiment, avoids the influence of atmospheric disturbance and air mass difference on the calibration precision of structure parameter. The calibration equipment only uses theodolite, and data acquisition is simple and easy to operate, so it is widely applicable to complex mechanism collimation measuring instruments and devices represented by servo star sensor.

[0043] 5. The application adopts least square method to optimize and estimate the installation error conversion matrix between prism reference system and servo zero position reference system and the installation error conversion matrix between servo real-time generated position and camera coordinate system, so as to improve estimation precision. DETAILED DESCRIPTION

[0044] Figure 1 The flow chart of the method of the application.

[0045] Figure 2 The prism coordinate system determination schematic diagram.

[0046] Figure 3 The crosshair target imaging determination schematic diagram.

[0047] Figure 4 The crosshair edge extraction schematic diagram.

[0048] Wherein, 1 - servo star sensor, 2 - star sensor hexahedron, 3 - first theodolite, 4 - second theodolite, 5 - fixed theodolite, 6 - collimating theodolite, 7 - computer. DETAILED DESCRIPTION

[0049] The application will be described in detail below with examples and drawings.

[0050] As Figure 1 shown, the application provides a calibration method for star sensor structure parameters, and the specific steps are as follows:

[0051] Step one, build a theodolite calibration system:

[0052] Set a fixed theodolite, a collimating theodolite and two measuring theodolites around the star sensor; the two measuring theodolites are respectively aligned with two orthogonal side surfaces of the reference prism of the star sensor.

[0053] Step two, the measuring theodolites perform normal measurement on the reference prism, express the three axes of the prism reference system O p -X p Y p Z p in the measuring theodolite reference system, and obtain the conversion relationship between the prism reference system and the measuring theodolite reference system:

[0054] Select two mutually orthogonal side surfaces on the reference prism of the star sensor to perform normal measurement. As Figure 2 shown, establish the prism coordinate system O p -X p Y p Z p , and the three axes are respectively coincided with the normals of the three side surfaces of the reference prism. Since the rear side surface normal is aimed, and the direction of the collimating light of the theodolite is opposite to the normal, the reading of the theodolite represents the direction of a certain axis of the prism coordinate system in the theodolite. Record the azimuth angle and the pitch angle of the aimed rear theodolite as the representation vector of the axis in the theodolite reference system.

[0055] The application adopts two measuring theodolites (first theodolite and second theodolite) to respectively measure two side surfaces of the reference prism. The first theodolite is used to perform normal measurement on a prism side surface of the star sensor, and the azimuth angle α Xp and the pitch angle β Xp of the theodolite in the collimating process are obtained as the representation vector of the X p axis of the prism coordinate system O p -X p Y p Z p in the first theodolite coordinate system; record the prism side surface as the first side surface, and the theodolite as the first theodolite.

[0056] Select the prism side orthogonal to the first side as the second side, and use another theodolite to measure the normal of the second side, and obtain the azimuth angle α Yp and the pitch angle β Yp of the second theodolite as the representation vector of the Y p axis of the prism coordinate system in the second theodolite coordinate system; and the theodolite is referred to as the second theodolite.

[0057] The application can also use the positive and negative mirror method, repeat step one, and take the average of multiple measurements to further eliminate measurement errors.

[0058] Step three, measure the mutual sighting of the theodolite and the fixed theodolite, and obtain the conversion relationship of the reference systems of the two; combine the conversion relationship of the theodolite and the prism reference system obtained in step two to obtain the conversion matrix of the prism reference system and the fixed theodolite reference system:

[0059] The first theodolite and the fixed theodolite mutual sighting, the azimuth angle α Xp is converted to the fixed theodolite reference system O s -X s Y s Z s , and the representation vector of the X p axis in the fixed theodolite reference system is obtained

[0060] O s -X s Y s Z s , the azimuth angle α' Xp of the X p axis is:

[0061] α' Xp = α Xp +( α s1 - α m1 +180°) (2)

[0062] Then the corresponding representation vector is expressed as:

[0063]

[0064] Wherein, the azimuth angle readings of the fixed theodolite and the first theodolite during mutual sighting are α s1 and α m1 .

[0065] Similarly, the representation vector of the Y p axis in the fixed theodolite reference system is obtained The second theodolite and the fixed theodolite mutual sighting, the azimuth angle α Yp is converted to the fixed theodolite reference system Os - X s Y s Z s down, get O s - X s Y s Z s down Y p azimuth angle of the axis α' Yp ; based on the azimuth angle α' Yp and the pitch angle β Yp , calculate Y p the vector representation of the axis in the fixed theodolite reference system

[0066] Using the properties of the cross product, Z p the vector representation of the axis in the fixed theodolite coordinate system

[0067]

[0068] In order to ensure the orthogonality of the Y p axis and the other two axes, the cross product can be used to correct again

[0069]

[0070] Finally, using the orthogonal matrix properties, the conversion matrix of the prism reference system to the fixed theodolite reference system can be obtained

[0071]

[0072] Step four, move the sighting theodolite:

[0073] As Figure 3 shown, the servo device is controlled by the computer to point the star sensor to a certain space direction. The moving sighting theodolite makes its crosshair target enter the field of view of the star sensor. At this time, the crosshair target of the sighting theodolite can be regarded as a calibration target at infinity in the space direction.

[0074] Step five, the crosshair target is imaged on the image plane of the star sensor to obtain the crosshair target image:

[0075] Fine-tune the azimuth and pitch angle of the sighting theodolite to make the center of the crosshair target image at five positions of the upper, lower, left, right and center of the image plane of the star sensor, respectively, and record the azimuth angle and pitch angle of the sighting theodolite at each time of imaging as the representation vector of the crosshair target in the sighting theodolite reference system. The fine-tuning of the theodolite combined with the angle change of the star sensor servo can well reduce the coupling of the structural parameters and improve the measurement accuracy.

[0076] Step six, extract the center of the crosshair target image, calculate the off-target vector:

[0077] Extract the center of the crosshair target image on the image plane, calculate the off-target amount. The crosshair target is two orthogonal thick lines on the star sensor image plane, four positions of the crosshair target image are extracted in step four, the center position of each image is solved, the average is obtained, the center position of the crosshair target is obtained, and the off-target vector is obtained. The specific solving steps are:

[0078] Step (6-1), gray equalization and lifting of the crosshair target image, highlighting the crosshair target;

[0079] Step (6-2), the edge of the crosshair target is extracted by using Canny edge extraction algorithm, and four straight line edges of two horizontal and two vertical are obtained, as shown by the dashed line of Figure 4 The edge of the crosshair target is removed in the application to avoid introducing errors due to inaccurate edge information at the intersection;

[0080] Step (6-3), the parameters of the four straight line edges are obtained by using Hough straight line transformation;

[0081] Step (6-4), the four intersection points of the straight line equations are solved by two-by-two;

[0082] Step (6-5), the average of the four intersection points is the center position of the crosshair

[0083] Step (6-6), the position of the principal point in the star sensor optical imaging model is The crosshair off-target vector is:

[0084]

[0085] Step seven, convert the off-target vector into the incident vector in the camera reference system:

[0086] Convert the off-target vector into the incident vector in the camera reference system of the star sensor In this embodiment, according to the focal length, distortion, and pixel size parameters of the camera, the off-target vector is converted into the incident vector in the camera reference system

[0087] Step eight, mutual sighting of the collimation theodolite and the fixed theodolite, obtain the reference system conversion matrix between the two, combine the conversion matrix between the fixed theodolite and the prism reference system in step three, and convert the representation vector of the crosshair target in the collimation theodolite reference system into the incident vector in the prism reference system:

[0088] In step five, the azimuth angle and the elevation angle of the collimating theodolite are collected at 5 positions of the crosshair image, and the representation vector of the crosshair target in the reference system of the collimating theodolite is obtained

[0089] By the same operation as in step three, the collimating theodolite and the fixed theodolite are mutual sighting, the azimuth angle of the collimating theodolite and the fixed theodolite are recorded, the conversion matrix between the reference system of the collimating theodolite and the reference system of the fixed theodolite is determined, and the representation vector is converted into the representation vector in the reference system of the fixed theodolite According to formula (5), the incident vector in the prism reference system is obtained

[0090] Step nine, according to the incident vector of the crosshair target and the incident vector of the miss distance vector, the structural parameters are optimally estimated:

[0091] In the imaging process, the installation error conversion matrix between the prism reference system and the servo zero position reference system exists The servo rotation matrix exists between the servo zero position and the servo real-time occurrence position The installation error conversion matrix exists between the servo real-time occurrence position and the camera coordinate system Wherein, The matrix and The matrix belongs to the structural parameters to be calibrated, and the servo rotation matrix The data [θ y θ z ] provided by the code disc of the servo device are calculated to obtain:

[0092]

[0093] The least square method is used to optimally estimate formula (8), and the matrix and The matrix and The matrix:

[0094]

[0095] In summary, the above is only a preferred embodiment of the present application, and is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for calibrating the structural parameters of a servo-type star sensor, characterized in that, The specific steps include: Step 1: Set up a fixed theodolite, an aiming theodolite, and two measuring theodolites around the star sensor; align the two measuring theodolites with the two orthogonal sides of the reference prism of the star sensor, respectively; Step 2: Normalize the two axes of the prism reference system of the two measuring theodolites and characterize the two axes under the measuring theodolite reference system; cross-align the measuring theodolite and the fixed theodolite to obtain the transformation matrix between the fixed theodolite and the prism reference system; Step 3: Move the aiming theodolite so that its crosshair target is imaged on the image plane of the star sensor, and transform the crosshair target into a representation vector in the aiming theodolite reference frame; the aiming theodolite and the fixed theodolite aim at each other to obtain the transformation matrix between the fixed theodolite and the aiming theodolite reference frames. Combined with the transformation matrix in Step 2, the transformation matrix between the aiming theodolite and the prism theodolite reference frames is obtained, and the representation vector of the crosshair target is converted into the incident vector in the prism reference frame. Step 4: Extract the imaging center position of the crosshair target and obtain the target miss vector; Step 5: Convert the off-target vector into the incident vector in the camera reference frame of the star sensor; Step 6: Based on the incident vectors from Steps 3 and 5, calibrate the structural parameters; The specific method for calibrating and estimating structural parameters is as follows: Substituting the incident vectors obtained in steps three and five, we perform calibration estimation on equation (8) to obtain the installation error transformation matrix between the prism reference system and the servo zero-position reference system. There is an installation error transformation matrix between the servo real-time position and the camera coordinate system. Equation (8) is: in, The incident vector is the one in the camera's reference frame. Let be the incident vector of the prism reference frame; This is the servo rotation matrix, obtained from the servo encoder.

2. The method as described in claim 1, characterized in that, The specific process of characterizing the two axes of the two theodolite normal measurement prism reference frames under the theodolite reference frames is as follows: Using the first and second theodolites, the normals of the two orthogonal prism sides of the star sensor are measured to obtain the azimuth angle α of the first theodolite during self-collimation. Xp and pitch angle β Xp As the prism coordinate system O p -X p Y p Z p X p The axis is represented by a vector in the first theodolite coordinate system, and the azimuth angle α of the second theodolite is obtained during collimation. Yp and pitch angle β Yp As the prism coordinate system O p -X p Y p Z p Y p The axis is represented by a vector in the second theodolite coordinate system.

3. The method as described in claim 1, characterized in that, The specific process of obtaining the transformation matrix between the measuring theodolite and the fixed theodolite and the prism reference system by mutual aiming is as follows: azimuth angle α Xp Switch to fixed theodolite reference frame O s -X s Y s Z s Below, we obtain O s -X s Y s Z s Next X p Azimuth angle α' of the axis Xp Based on the azimuth angle α' Xp and pitch angle β Xp Calculate X p The characterization vector of the axis in a fixed theodolite reference frame The azimuth angle α Yp Switch to fixed theodolite reference frame O s -X s Y s Z s Below, we obtain O s -X s Y s Z s Next Y p Azimuth angle α' of the axis Yp Based on the azimuth angle α' Yp and pitch angle β Yp Calculate Y p The characterization vector of the axis in a fixed theodolite reference frame Based on the characterization vector and characterization vector Z is obtained using the cross product. p Vector representation of axes in a fixed theodolite coordinate system This leads to the transformation matrix from the prism reference system to the fixed theodolite reference system.

4. The method as described in claim 3, characterized in that, The representation vector in the transformation matrix To be corrected again using cross product Right now Then the cross product is corrected. Assigned to 5. The method as described in claim 1, characterized in that, The specific method for extracting the imaging center position of the crosshair target and obtaining the target miss vector is as follows: The grayscale of the crosshair target image is uniformly increased to highlight the crosshair target; The edges of the crosshair target were extracted using the Canny edge extraction algorithm, and four straight edges (two horizontal and two vertical) were obtained and removed. The parameters of the edges of the four lines are obtained using the Hough line transform; Solve for the four points of intersection of the two pairs of lines by solving the equations of the lines simultaneously. The mean of the four intersection points is the center position of the crosshairs. The location of the principal point in the optical imaging model of the star sensor is: Off-target amount for:

6. The method as described in claim 1, characterized in that, The least squares method is used to calibrate and estimate equation (8).

7. The method as described in claim 1 or 2, characterized in that, The normal direction was measured using the upright and reverse mirror method.

Images