CN113483751A - Radial triangle mapping matrix-based star map identification method - Google Patents
Radial triangle mapping matrix-based star map identification method Download PDFInfo
- Publication number
- CN113483751A CN113483751A CN202110723788.9A CN202110723788A CN113483751A CN 113483751 A CN113483751 A CN 113483751A CN 202110723788 A CN202110723788 A CN 202110723788A CN 113483751 A CN113483751 A CN 113483751A
- Authority
- CN
- China
- Prior art keywords
- star
- mode
- radial
- stars
- main
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/02—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by astronomical means
- G01C21/025—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by astronomical means with the use of startrackers
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Astronomy & Astrophysics (AREA)
- Automation & Control Theory (AREA)
- General Physics & Mathematics (AREA)
- Navigation (AREA)
Abstract
A star map identification method based on a radial triangular mapping matrix comprises the following steps: extracting a radial triangular mode of the navigation satellite and establishing a navigation satellite mode library; according to the geometric distribution of observation stars extracted from a star map, a radial triangle which is formed by taking a main star as a vertex and any two adjacent stars within a mode radius is constructed, the angular distance of the radial triangle is mapped to a mode matrix, voting is completed through matching of each grid of the mode matrix, a high-vote candidate star is selected as a main star label through voting, a rotation matrix and an attitude angle are determined for individual low-vote and low-vote candidate stars by utilizing double-vector attitude determination, the radial triangle is verified through a re-projection method, and an observation star verification identification result is output to obtain the star label. In the same way for other N close to the main point of the star mapmaxAnd identifying the star to be identified. The invention inherits the robustness of the triangle algorithm, constructs the mode matrix with rotation invariance, searches the whole mode library when voting is matched on the radial triangle mode every time, and has completeness.
Description
Technical Field
The invention relates to the field of star map identification in astronomical navigation of star sensors, in particular to a star map identification method based on a radial triangular mapping matrix.
Background
The star sensor is an important hardware component in a satellite attitude control system, and the acquisition of the three-axis attitude of the spacecraft is realized through star map recognition. The star sensor measurement does not need estimation prior knowledge, and the measurement precision can reach an angle second level, so the star sensor measurement method is widely applied to a star sky detection task.
The existing star map recognition algorithms are roughly classified into sub-map isomorphic algorithms and pattern algorithms according to different feature extraction modes. Sub-graph isomorphic algorithms such as triangle algorithm, pyramid algorithm, etc. look the star points as the vertices of the graph, with the edges corresponding to the angular distance between two vertices in the same field of view. The pattern class algorithm is used for identifying the main star by constructing a pattern on the relative position relation of star points around the main star. The method comprises the steps of selecting a main star, selecting a relative position relation between an adjacent star and the main star to construct a grid coordinate system, dividing grids within a mode radius, enabling star points to fall into the grids to form a specific star mode, and identifying the main points through mode matching. The pattern-class algorithm is, for example, a grid algorithm and a radial-circumferential algorithm, and most of the pattern-class algorithms mainly use angular distances in a construction manner. The pattern recognition algorithm has the main advantages that the capacity of the stored data pattern is small, and the matching pattern can be obtained through fast searching. The pattern-like algorithm needs to observe more star points stored in the star map to obtain a unique star point pattern, so that the performance of the pattern-like algorithm is obviously reduced when the number of the star points is less. Meanwhile, when the interference of large position noise exists, the performance of most mode type algorithms is obviously reduced. In addition, the pattern-like algorithm, such as the grid algorithm, needs to find the nearest neighbor to determine the pattern direction when constructing the pattern, and the radial annular star map recognition algorithm needs to find the minimum included angle in the annular direction as the starting x-axis direction of the coordinate axis when extracting the annular feature. In the practical use of the star sensor, the interference of noise such as atmospheric space and the like is often caused, the false star noise interfering with the nearest neighbor and the lost star noise are generated, and when the nearest neighbor encounters the interference of the false star noise, the mode is greatly influenced, and the identification fails.
According to the analysis, most pattern type algorithms need to find nearest neighbor stars, and for example, a grid algorithm construction pattern needs to rely on a star map to acquire sufficiently robust neighbor stars to obtain a good identification effect.
Disclosure of Invention
Aiming at the defect that the conventional radial algorithm cannot utilize annular information, the invention provides a radial triangular pattern feature star map recognition algorithm on the basis of the radial algorithm.
The technical scheme of the invention is as follows:
a star map identification method based on a radial triangular mapping matrix is characterized by comprising the following steps: the method comprises the following steps:
s100, constructing a radial triangular mode and simultaneously establishing a radial triangular mode star atlas identification database;
s200, constructing a radial triangular mode according to the positions of adjacent stars in the mode radius of each main star in the star map;
s300, sequentially extracting observation stars closest to a principal point, and identifying the observation stars by using a voting method according to a radial triangular mode matrix of each observation star in the star map;
and S400, selecting other eight observation stars close to the image main point as main stars, calculating the angular distances from the main stars to the satellite stars one by one, searching radial triangles matched within the mode radius in a radial triangle mode star map identification database by matching with the constructed index table, and identifying all the stars to be observed.
Further, the step S100 includes:
step S110, establishing a navigation satellite database;
step S120, constructing a radial triangular mode library;
step S130, a mode index library is established, and the compression mode library is rapidly searched.
Further, in step S110, the building a navigation satellite database includes:
determining the radius of a main satellite and a mode, calculating the distance from the main satellite to each adjacent satellite, sequencing the brightness and the angular distance of the adjacent satellites around the main satellite, screening the adjacent satellites, and keeping the brightest adjacent satellites not more than 9 adjacent satellites close to the main satellite in the main satellite mode.
Further, in step S110, the establishing a navigation satellite database further includes:
s111, removing stars which influence the construction of the robust mode in the star catalogue, and firstly removing dark stars such as stars and the like exceeding 6 Mv;
step S112, two navigation stars with the star point angular distance value smaller than 0.1 degree are removed as double stars;
step S113, removing the star catalogue navigation star containing less than three adjacent stars within the defined mode radius.
Further, in step S120, the step of constructing a radial triangle pattern library includes:
randomly selecting two stars around the main star and the main star to form a radial triangle, radially dividing the mode radius of the main adjacent star pair, and sequentially quantizing the angular distances of the two main adjacent stars of each radial triangle into 256 intervals according to the anticlockwise direction.
Further, in step S130, a mode index library is established, and the step of implementing fast search by the compression mode library includes:
forming an index from two sides of a radial triangle
Quantizing the value of the third edge
Mapping to 256-length mode matrix, and when multiple radial triangles are mapped to the grid represented by the same two-dimensional matrix, obtaining 1-norm of all elements in the grid matrix M
And stores the main star label and the number n of radial triangles mapped to the same grids. Storing the radial triangle patterns corresponding to all navigation stars with the pattern radius into a two-dimensional matrix gridForming a radial triangular pattern.
Establishing a radial triangular compression mode database, constructing a 256 x 256 mode matrix by the radial triangular modes extracted by each navigation satellite, extracting the non-zero values of the grid modes of all the navigation satellites only by the compression mode matrix database, wherein the non-zero value of each grid mode corresponds to a mode value
Major star number SnAnd the number of repetitions nsAnd stored uniformly in a pattern table. Numbering each section of the compression mode library, and constructing the radial triangle index and the compression mode library number into a corresponding lookup relation index f (e)i,ej) Each of each column of the compressed mode library can be quickly aligned by indexing
The value is looked up.
Further, in the step S200, constructing the radial triangle pattern according to the position of the neighboring star within the pattern radius of each main star in the star map includes:
calculating each observation star in the star map to the principal point (u)0,v0) Sequentially selecting the Euclidean distances not exceeding N according to ascending sequence of the distance valuesmaxThe star of (a) is an observation star, wherein N ismax≤8；
And sequentially calculating Euclidean distance values from adjacent stars to the main star in the mode radius by taking the main star as the center of a circle and taking the 6-degree angular distance as the mode radius, sequencing the Euclidean distance values in ascending order, screening the adjacent stars in the star map according to the method for selecting the adjacent stars in the step S120, and establishing a radial triangular mode for each navigation star after the angular distance is quantized.
Further, in the step S300, the step of sequentially extracting the observation stars closest to the principal point, and identifying the observation stars by using a voting method according to the radial triangular pattern matrix of each observation star in the star map includes:
and S310, sequentially extracting the observation stars closest to the principal point, and voting according to the radial triangular mode of the observation stars.
And S320, sequencing the corresponding counting table of each observation satellite in the radial triangular mode according to the counting value in a descending order, wherein the value in the counting table is the ticket number of each navigation satellite, and taking the first-ranked main satellite candidate as a main satellite identification result when the number of the first-ranked main satellite candidate ticket of the main satellite is two times or more higher than the number of the second-ranked main satellite candidate ticket of the main satellite.
And step S330, verifying and identifying other higher ticket identification results which cannot be confirmed in the identification view field by verifying and identifying links and utilizing the star result of identification and a double-loss reprojection method.
Further, in step S400, when the selected primary star is identified, the process returns to step S200, and the second observed star is selected as the primary star for identification until the nth star is identifiedmaxThe star is observed.
Compared with the prior art, the invention has the following beneficial technical effects:
(1) the invention improves the radial algorithm and the triangle algorithm, modifies the problem that the radial algorithm cannot utilize the annular information, makes up for the defect that the triangle algorithm needs to select a large number of triangles for storage, compresses the constructed radial triangle pattern database, and avoids the problems of overlarge space occupation and too low search speed of the pattern database.
(2) The invention defines the concept of the main star and carries out discretization treatment on three edges of a radial triangle formed by adjacent stars by utilizing the angular distance information between the adjacent stars in the mode radius. In the radial triangle mode, each triangle is formed by taking the main star as a vertex, so the frequency of the main star is the highest in the whole radial triangle group, and the identification of the main star is determined by adopting a voting method. The method has the advantages that the relative angular distance information in the construction mode is utilized to the maximum extent, and the completeness and the reliability of the star map identification algorithm are improved while the robustness of the triangle algorithm to various types of noise is kept through voting on the radial triangle mode.
(3) The method adopts a bright star extraction strategy when the adjacent star is screened, and has stronger robustness on star and other noises.
(4) The invention carries out discretization processing on three edges by using a radial triangle voting method, and makes an algorithm have stronger robustness on position noise and false star noise by fuzzy matching of a mode matrix.
(5) The invention utilizes the mode storage mode of the grid algorithm, and can improve the mode query efficiency.
(6) The radial triangular mode constructed by the invention has rotation invariance, and the problem that the radial annular star map identification algorithm is not stable enough is solved through the constructed mode matrix.
(7) The invention utilizes the mutual geometric information of all candidate stars in the voting process of each main star, has completeness and ensures that the identification result is more stable and reliable.
Drawings
FIG. 1 is a schematic diagram of the transformation relationship of the star vector from the geocentric inertial coordinate system to the image coordinate system according to the present invention;
FIG. 2 is a diagram of a radial triangle pattern used in the method of the present invention;
FIG. 3 is a diagram of a quantized radial triangle pattern library and a compressed storage method according to the present invention;
FIG. 4 is a schematic diagram of fuzzy search of pattern matrices according to the present invention;
FIG. 5 is a schematic diagram of a mode matrix storage method according to the present invention;
FIG. 6 is a diagram illustrating a voting process after matching the patterns according to the present invention;
FIG. 7 is a graph of the effect of the star point location noise on the identification method of the present invention;
FIG. 8 is a diagram illustrating the effect of noise such as star point and star on the identification method according to the present invention;
FIG. 9 is a graph of the impact of the star point pseudolite noise on the identification method of the present invention;
fig. 10 is a diagram of the simulation star atlas identification result of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the accompanying drawings in conjunction with the following detailed description. It should be understood that the description is intended to be exemplary only, and is not intended to limit the scope of the present invention. Moreover, in the following description, descriptions of well-known structures and techniques are omitted so as to not unnecessarily obscure the concepts of the present invention.
The star map identification method based on the radial triangular mapping matrix of the present invention is described in detail below by specific embodiments.
The star sensor imaging model is shown in FIG. 1, and the coordinate of the fixed star in the celestial coordinate system is assumed to be (alpha)i,δi) The direction vector under the star sensor coordinate system is (x)i,yi,zi) The imaging coordinate under the image coordinate system is (X)i,Yi). The imaging link includes two parts: converting the celestial coordinate system into the rotation transformation of the star sensor coordinate system; and (4) carrying out projection transformation from the star sensor coordinate system to an image coordinate system.
Converting the celestial coordinate system into a star sensor coordinate system: the attitude angle of the star sensor is assumed to be (alpha)0,δ0,φ0) In which α is0Is the right ascension, delta0Is declination, phi0For roll angle, the rotation matrix M from the star sensor coordinate system to the celestial coordinate system can be expressed as:
for the star point satisfying the above formula, the direction vector under the celestial coordinate system
Direction vector (x) to star sensor coordinatei,yi,zi) The transformation of (d) is represented as:
f is expressed as the focal length of the optical system, and the perspective projection transformation is expressed as:
from the image coordinate system (X)i,Yi) The conversion to the pixel coordinate system (U, V) can be expressed as:
wherein Dx、DyRespectively representing the sizes of the pixels in the horizontal and vertical directions, (U)0,V0) Representing the principal point coordinates.
And S100, constructing a radial triangular mode and simultaneously establishing a radial triangular mode star atlas identification database.
Specifically, a navigation satellite database, a radial triangle mode library and a mode index library are established.
The mode index library is an angular distance characteristic database index table.
Specifically, the step of establishing the radial triangular pattern star map identification database includes the following steps:
and step S110, establishing a navigation satellite database.
The invention adopts a Smithsonian space star chart (Sao star chart) as the basis for establishing a navigation basic star point characteristic database. The star table of the star table database comprises the following four types of information: star information, right ascension information, declination information, stars and the like. The quality of the star map identification effect depends on whether the selected star information is stable and reliable enough, so the star map navigation star needs to be screened before the identification step is started.
Specifically, a main star and a mode radius are determined, the distance from the main star to each adjacent star is calculated, the brightness and angular distance sorting is carried out on the adjacent stars around the main star, the adjacent stars are screened, and the brightest adjacent star which is not more than 9 adjacent stars close to the main star in the main star mode is reserved.
The establishing of the navigation satellite database comprises the following steps: and (4) screening a star catalogue.
In order to ensure that the algorithm can operate stably in real time, the number of navigation stars in the database is as small as possible, and therefore the navigation star library is preprocessed mainly for screening the following three types of stars: the star is a dark star with the star equal to or larger than 6Mv, the dual star with the angular distance between the stars smaller than 0.1 degree and the navigation star only contains 3 adjacent stars within the mode radius;
specifically, the star catalogue screening method comprises the following steps:
and S111, removing stars which influence the construction of the robust mode in the star table, and firstly removing dark stars such as stars and the like exceeding 6 Mv.
And step S112, two navigation stars with the star point angular distance value smaller than 0.1 degree are removed as double stars.
Because two satellites are interfered in the star map identification process, two satellites which are approximately overlapped cannot be identified, two navigation satellites with the star point angular distance value smaller than 0.1 degree need to be eliminated as the two satellites.
Step S113, removing the star catalogue navigation star containing less than three adjacent stars within the defined mode radius.
In particular, the defined mode radius RpatternradiusAnd 6 degrees, in order to ensure that at least 3 adjacent stars are contained in the radius of the tracking precision mode, the star table navigation stars containing less than three adjacent stars in the radius of the mode are removed.
And step S120, constructing a radial triangle mode library.
In order to obtain stable prior mode information in the identification process, radial triangle mode extraction needs to be carried out on the navigation satellite database, and a radial triangle mode library is constructed.
Specifically, two stars around the main star are randomly selected to form radial triangles with the main star, the mode radiuses of the main adjacent star pairs are divided radially, and the angular distances of the two main adjacent stars of each radial triangle are quantized to 256 intervals in sequence according to the anticlockwise direction, and are represented as
And taking each star in the navigation star database as a main star, and generating a simulation star map by using the imaging model. Using 6 degree angular distance as mode radius, aiming at all N in main star mode radiusneighThe adjacent star is screened out,calculating Euclidean distance values from each adjacent star to the main star, and sorting according to ascending order when N isneigh≤Nneigh_robustIs selected to be Nneigh_robust＝NneighThe adjacent star is used as the object of the mode construction when N isneigh＞Nneigh_robustWhen the number of the blind stars is the same, the blind stars far away from the main star are screened. The invention is provided with Nneigh_robustWhen the star is 9, the main star and any two adjacent stars can form a radial triangle, and the number of the radial triangle is
Wherein two edges with the main star as the vertex are arranged in the anticlockwise direction
Corresponding to the horizontal and vertical coordinates of the mode matrix and the edges corresponding to the vertexes
Corresponding to the values in the pattern matrix grid. When n issWhen the radial triangular mode values fall into the same grid, 1-norm of all elements in the grid is obtained
And respectively establishing a grid index table to mark grid index values
And the number n of modes in the grids。
Quantizing three edges of the radial triangle corresponding to all the navigation stars into
Mode storage to N N mode matrix MnIn, constructing a navigation star schema library, wherein
As shown in FIG. 1And (3) taking each star in the navigation star database as a main star, and generating a simulation star map by using the imaging model. With RpatternradiusThe mode radius is 6 degrees angular distance and is satisfied in the simulation star map by millimeter
f is the focal length of the lens, alphadIs the mode radius field angle. To minimize the effect of star-like noise on identification, all N within the radius of the primary star pattern are assignedneighScreening the adjacent stars, calculating the Euclidean distance value from each adjacent star to the main star, sorting according to ascending order when N isneigh≤Nneigh_robustIs selected to be Nneigh_robust＝NneighThe adjacent star is used as the object of the mode construction when N isneigh＞Nneigh_robustWhen the number of the blind stars is the same, the blind stars far away from the main star are screened, and N is setneigh_robustThe main star and any two stars can form a radial triangle pattern as 9, and the number of the radial triangles is
As shown in fig. 2, S is a radial pattern constructed by taking the quantization interval N as 10 as an example0Is a main star, S1～S5Is neighboring star, Δ S0S1S2、ΔS0S1S3、ΔS0S1S4、ΔS0S2S3、ΔS0S2S4、ΔS0S3S4、ΔS0S3S5、ΔS0S4S5、ΔS0S5S1、ΔS0S5S2Together forming a master star model. By Delta S0S1S2For example, each radial triangle in the figure is according to dx→dy→dxyQuantizes three sides of the triangle. The method respectively calculates three sides { x, y, z } of each radial triangle in a counterclockwise direction, and utilizes the following methodThe radial triangles are quantized by the formula:
wherein d ismax＝Rpr，
To round down, the method takes the quantization interval N as 256, when the focal length is 49.277mm, dmax430.42 mm. Two critical corner distances d of the quantized trianglex、dyAre respectively denoted by ei、ejOpposite angular distance dxyAfter quantization is denoted as eij. The quantized angular distance ei,ej,eijStoring in a temporary pattern matrix, recording index values, and characterizing by
When n issWhen the radial triangles are indexed into the same mode grid, 1-norm of all elements in the grid is obtained
And establishing a grid position index table containing the number of the non-zero position of each mode matrix
Number n of repeated radial triangles in the gridsAnd the matched main adjacent star number. When n issWhen the index number is larger than 1, the plurality of radial triangles fall into the same grid due to the same index number, the plurality of radial triangles share one vertex, and only the main star number is stored at the moment.
Step S130, a mode index library is established, and the compression mode library is rapidly searched.
In particular, the index is formed by two sides of a radial triangle
A third stripQuantized value of edge
Mapping to 256-length mode matrix, and when multiple radial triangles are mapped to the grid represented by the same two-dimensional matrix, obtaining 1-norm of all elements in the grid matrix M
And stores the main star label and the number n of radial triangles mapped to the same grids. And storing the radial triangular modes corresponding to all navigation stars with the mode radius into a two-dimensional matrix grid to form the radial triangular mode.
And compressing a radial triangular mode library to improve the searching speed. The number of matrix grids included in the pattern library constructed in step S120 is N × N, wherein the patterns of each navigation satellite are N × N square matrices, and the matrices include a large number of 0 values and occupy a large amount of useless space when stored, so that N × N tables constructed in step S120 are merged into one table after the matrices are compressed, wherein each column corresponds to an N × N index, and a nonzero value is arranged from small to large for each index corresponding to a pattern matrix
Main adjacent star number
And the number of repeated triangles in the grid
Each group of
Corresponds to a radial triangle. The relationship of each index number and index may be expressed as (e ═ e)i-1)×N+ej。
Index values obtained by observing a radial triangle pattern library constructed by stars
Can quickly find the star value corresponding to the index and then pass the star value
Corresponding to
Corresponding to observation stars
Comparing to determine whether the star is the candidate star.
And establishing a radial triangular compression mode database. The radial triangle mode extracted by each navigation satellite constructs a 256 x 256 mode matrix, the compression mode matrix database only extracts the non-zero values of the grid modes of all the navigation satellites, and the non-zero value of each grid mode corresponds to a mode value
Major star number SnAnd the number of repetitions nsAnd uniformly storing the index values into a mode table, numbering each interval of the compression mode library, and constructing the radial triangle index and the compression mode library number into a corresponding lookup relation index f (e)i,ej) Quickly aligning each of each column of the compressed mode library by indexing
The value is looked up.
The number of matrix grids included in the radial triangular pattern library constructed in step S120 is N × N, wherein the pattern of each navigation satellite is an N × N square matrix, and the maximum number of non-zero patterns included in one pattern matrix is N × N
Then a large amount of unused space is occupied when the matrix contains a large amount of 0 value storage. As shown in fig. 3, in order to improve the storage efficiency, the matrix is compressed, and an index relationship is established for the compressed lookup table. Will be described in detailThe N pattern matrices constructed in S120 are regarded as N × N tables, and the N × N tables are merged into one table, where each column corresponds to one grid of the N × N pattern matrices, and the grid pattern values are arranged from small to large for the corresponding pattern matrices under each index
Main adjacent star number group
And the number of repeated triangles in the grid
Each group of
Corresponds to a radial triangular pattern. The compressed pattern matrix is shown in table 1. The relationship of each index number and index may be expressed as (e ═ e)i-1)×N+ej. Let each grid obtain a maximum non-zero value of nmaxCompressed pattern library size of nmaxxNxN, in which N is 256, NmaxIs much smaller than n, and the index value is obtained by a radial triangle pattern library constructed by observation stars
The star value corresponding to the index can be quickly found.
TABLE 1 compression mode matrix
And S200, constructing a radial triangular mode according to the positions of the adjacent stars in the mode radius of each main star in the star map.
In the real star map, a mode radius is defined for each main star, the size of the mode radius is the same as that in step S120, and a radial triangular mode is constructed for the main star.
The process is as follows:
step S210, calculating each observation star to principal point (u) in the star map0,v0) Sequentially selecting the Euclidean distances not exceeding N according to ascending sequence of the distance valuesmaxThe star of (a) is an observation star, wherein N ismax≤8；
Calculating each star to the main point (u) of the star map in the star map0,v0) The identification order selects the star closest to the principal point in turn to be close to the center of the image (u)0,v0) The star of (1) is the main star.
And S220, sequentially calculating Euclidean distance values from adjacent stars to the main star within the mode radius by taking the main star as the center of a circle and taking the 6-degree angular distance as the mode radius, sequencing the Euclidean distance values in ascending order, screening the adjacent stars in the star map according to the method for selecting the adjacent stars in the step S120, and establishing a radial triangular mode (mode matrix) for each navigation star after the angular distance is quantized.
A12-degree circular field of view is taken by taking the main star as the center of a circle, and other stars in the mode radius are taken as adjacent stars. And screening by the step of screening the adjacent stars by the main star in the S120. The main star and any two adjacent stars form a radial triangle. And storing the serial numbers of the three navigation stars into index units corresponding to the two directional indexes according to the sequence of the main star, the first adjacent star in the anticlockwise direction and the second adjacent star, and repeating the above steps to store all the radial triangles into a mode matrix to form a main star mode.
And S300, sequentially extracting the observation stars closest to the principal point, and identifying the observation stars by using a voting method according to the radial triangular mode matrix.
And identifying the observation star of the main star by using a voting method for a radial triangular mode constructed in the mode radius, and voting to obtain the star number with the most votes as the star number of the main star.
The method comprises the following steps:
and S310, sequentially extracting the observation stars closest to the principal point, and voting according to the radial triangular mode of the observation stars. Specifically, according to the radial triangular mode matrix of the observation star, voting is performed on the candidate star of the observation star once, and the process is as follows:
establishing a 1 XN star by taking all navigation stars in the navigation star library as candidate stars of the generated radial triangle mode of each observation starcandiAnd initializing the counter to 0, NcandiThe number of navigation stars in the star catalogue.
Because the radial triangle mode has rotation invariance, a star closest to the main star is selected as an initial star to construct a radial triangle in the counterclockwise direction, two edges taking the main star as a vertex after the radial triangle is quantized are selected as the index value of the mode matrix in the step S220, when the index value is not at the edge of the matrix, the radial triangle mode values stored in eight adjacent units around the index are taken as the candidate mode of the observation star, and if the index unit is at the edge of the lookup table, only the index unit and the adjacent grids around the index unit are considered.
Each radial triangle mode of the observation star is searched by using the mode matrix index, after two adjacent edges of the radial triangle about the principal point are quantized, the obtained value is used as the index value of the mode matrix in the step S130, and the mode value is subjected to fuzzy search. When the index value is at the center of the pattern matrix, the index unit and the radial triangle pattern values stored in the surrounding 8 units are used as candidate patterns of the observation star, and when the index value is at the edge of the pattern matrix, only the pattern values of the index grid and the surrounding adjacent grids in the pattern matrix are considered.
After the radial triangular mode of the observation satellite is matched with the mode library, adding 1 to a counter of a main satellite serial number corresponding to each observation satellite mode, and simultaneously adding 1 to a counter of an adjacent satellite A, B serial number; when the pattern includes nsWhen there is a radial triangle, the number counter of the main star is increased by nsAnd storing the serial numbers of the main adjacent stars of the candidate radial triangles into the index unit of the serial numbers of the main adjacent stars in the temporary lookup table.
And S320, sequencing the corresponding counting table of each observation satellite in the radial triangular mode according to the counting value in a descending order, wherein the value in the counting table is the ticket number of each navigation satellite, and taking the first-ranked main satellite candidate as a main satellite identification result when the number of the first-ranked main satellite candidate ticket of the main satellite is two times or more higher than the number of the second-ranked main satellite candidate ticket of the main satellite.
Specifically, when the pattern matrix corresponding to each observation star is identified, the grid where each radial triangle is located corresponds to a candidate star number and two adjacent star numbers, when the angular distance library indexes corresponding matching, 1 is added to the counter of the main star counting table corresponding to the observation star, and simultaneously, 1 is added to the counter of the adjacent star counting table corresponding to the adjacent star counting table. And when the main star label is in the repeated index table, adding n to the corresponding counter, and then storing the candidate star into the corresponding main star sequence number index unit in the main star temporary table.
And (3) arranging the votes of the counting tables corresponding to the main stars in a descending order, and taking the main star candidate star with the first rank as the main star identification result when the votes of the main stars reach two times or more of the second main star candidate votes for each main star.
And step S330, verifying and identifying other higher ticket identification results which cannot be confirmed in the identification view field by verifying and identifying links and utilizing the star result of identification and a double-loss reprojection method.
Specifically, when the first-ranked main star candidate result is the same as the second-ranked result, the identified main star is selected from the star map, and the top k candidate stars with similar votes are selected for verification and identification. The specific method comprises the following steps: selecting one identified main star in the center of the star map as a reference star, establishing a verification star pair, determining a star sensitive rotation matrix and an attitude angle corresponding to each verification star pair by using double-loss attitude determination, projecting navigation stars near the identified star onto a focal plane, generating star points on a reference star map, identifying the observation star correctly when coordinates of the star points in the reference star map and the star points in the observation star map are within a smaller error range, and selecting other observation stars as the main stars and then performing an identification step if the verification fails.
And S400, selecting other eight observation stars close to the image main point as main stars, calculating the angular distances from the main stars to the satellite stars one by one, and searching the radial triangles matched within the mode radius in a radial triangle database by matching with the constructed index table.
And selecting secondary adjacent stars and taking S220-S330 as identification steps until all the stars to be observed are identified.
The specific embodiment is as follows:
initialization: and setting a state identifier for each navigation satellite in the navigation satellite list, and initializing the state identifier in the matching process. Establishing a 1 XN observation star for each observation starcandiThe counter of (3) searches the value of the index position in the radial triangle pattern library through the grid index obtained in the step (S120), and simultaneously establishes a temporary lookup table to store the adjacent star label corresponding to the main star in the candidate radial triangle pattern.
In the radial triangular pattern matrix corresponding to each observation satellite generated in step S200, the index of the pattern matrix corresponding to the observation satellite at the non-zero position is extracted, and the index unit is used to search for the index value in the pattern matrix in step S100. By looking up pattern library asterisk values
Corresponding to
Corresponding to observation stars
Comparing to determine whether the star is the candidate star.
The conditions for successful matching are as follows:
1) pattern matching:
where function abs () represents the absolute value, pats(a, b) and patcAnd (i, j) respectively represent a pattern library radial triangular pattern and an observation star radial triangular pattern, the min _ match is a threshold value for successful matching, the recognition success rate and the false recognition rate are determined by the size of the threshold value, 4 is selected from the min _ match according to an experiment, and the probability of false matching is almost zero.
In order to meet the requirement that the model is robust to position noise, a matching error tolerance k is set, the selection of k is related to star point positioning noise, and the larger the star point positioning noise is, the larger the value of k is. It should be noted that larger k values increase the star map recognition time. According to the result of the paper "Improved Grid Algorithm Based on Star Pattern and Two-dimensional array distance for Full-Sky Star Identification", as shown in FIG. 4, performing fuzzy search on the Pattern matrix, the present invention takes k as 1, i.e. the radial triangle Pattern in the index unit of the Pattern matrix and the eight adjacent units around the index Grid as the candidate radial triangle Pattern of the observation Star, and only the index unit and the adjacent grids around the index unit are considered as the candidate radial triangle Pattern if the index unit is located at the edge of the matrix.
2) The number of radial triangles in the two grids is the same, namely nss＝nscWherein n isssThe number of radial triangles in a certain grid in an observation mode is nscThe number of radial triangles in a certain grid of the pattern library.
Matching a pattern matrix grid corresponding to the observation star with a radial triangular pattern library, and specifically comprising the following steps of:
when acquiring the grid pattern of the observation star, judging n of the grid patternsThe value is obtained. And when the number of the triangle modes in the grid is 1, adding 1 to a counter of the serial number of the main star corresponding to the radial triangle corresponding to the observation star, simultaneously storing the candidate serial numbers of the adjacent stars into a temporary lookup table corresponding to the index of the main star, and adding 1 to a counter of the serial number of the adjacent star corresponding to the counting table of the observation star of the adjacent stars. When the number of the triangular modes in the grid is more than 1, adding n to the counter of the mark number of the main star corresponding to the observation stars。
Fig. 6 shows a process of voting for a radial triangle formed by observation stars 0, 1, and 2 in fig. 2. Firstly, from two sides S of the radial triangle0S1、S0S2And determining index units (147, 78), and taking the index grid and the surrounding adjacent 8 units as candidate radial triangle patterns. When a candidate radial triangle pattern finds a match in the pattern library, voting is performed under a navigation star number counter corresponding to a main adjacent star corresponding to the candidate radial triangle pattern, for exampleSuch as candidate radial triangle 3408,3467,3534]And if 3408 is the main star number and is the candidate navigation star corresponding to the observation star No. 0, adding 1 to the count of the navigation star No. 3408 under the observation star No. 0.
When falling within the same grid, where the candidate triangle has multiple triangles sharing a dominant star, only the dominant star candidate is voted, e.g., multiple radial triangle combinations [234,342,456]，[234,468,456]N in the grid index table when the index units are at the same positionsEqual to 2, when the grid match is successful, the probability of the primary star match being correct for the grid is higher, so the star counter number 234 under the primary star observation is incremented by 2. Similarly, voting is carried out on all radial triangular modes in the grid once, and when the grid index table is matched with a plurality of radial triangular modes, the star counter under the observation star of the main star is added with ns。
And for the main star observation stars, when the number of the votes of the main star candidate navigation stars with the first ranking is two times or more than that of the votes of the second main star candidate navigation stars, selecting the candidate navigation stars with the first ranking as a main star identification result. Table 2 shows the voting results of the observation stars generated by taking the navigation star 650 as an example.
TABLE 2 matching set with navigation satellite 650 as the dominant satellite
Under the condition that the interference of the false stars exists around the main star or the number of star points contained in the mode radius is small, the front k stars in the candidate navigation stars obtain the same ticket number, the main star cannot be distinguished and identified, and the front k candidate navigation stars are selected for verification.
The verification method comprises the following steps: according to a pinhole model, a star closest to the center is taken as a main star and any one recognized adjacent star to form a verification star pair, a double-loss attitude determination method is adopted to perform attitude calculation on two stars to obtain an attitude vector, then a navigation star near the recognized star is projected onto a focal plane, star points are generated on a reference star map, the coordinates of the star points in the reference star map and the corresponding coordinates of the star points in an observation star map are within a small error range, the recognition is considered to be successful, otherwise, other candidate stars are selected as the main star to restart the recognition step, and when all k candidate stars are failed to be verified, other to-be-recognized stars are selected as the main star to restart the recognition step. Table 3 shows that the navigation star 23256 is the main star, and the voting result of two random pseudo-star noises is added to the pattern radius, and as can be seen from table 3, the navigation star 23555 can obtain more than twice the number of votes of the second star, so that the recognition result can be obtained. Table 4 shows the verification and identification result of the navigation satellite 23256, and it can be seen from the table that the navigation satellite 23256 cannot obtain the same number of votes as the navigation satellite 2770 after voting, and at this time, the navigation satellite 23555 satellite pair 23256 and 2770 within the mode radius are used as the main satellite numbers to perform verification, and the verification satellite pairs 23555 and 2770 are respectively obtained as the verification satellite pairs to obtain the attitude vectors, and the main satellite number is 23256 after the verification step.
TABLE 3 matching set with navigation satellite 23555 as the dominant satellite
Table 4 verification identification of a matching group of navigation stars 23256 as the dominant stars using navigation star 23555
When the selected main star identification is finished, returning to the step S200, selecting the second observation star as the main star to perform the steps S210-S340 until the Nth observation star is identifiedmaxThe star is observed. In the real star map recognition task, NmaxThe attitude can be determined by taking 3-8 in general, and in the method, N is takenmax＝8。
Performance analysis
In order to evaluate the performance of the method, a simulation star map is adopted to perform a simulation experiment, and the parameters of a simulation model are shown in table 5.
TABLE 5 Star sensor parameters
In order to verify the influence of noise on the method, random position noise, star and other noise and pseudo star noise are added to the simulated star map respectively. Gaussian position noise with the mean value of 0, standard deviation sigma of 0 to 4 pixels and star-like noise with the mean value of 0 and the standard deviation sigma' of 0.4Mv are respectively added to the real positions of the simulated star map. And (4) counting the 5000 generated simulated star maps, wherein the successfully identified standard values are 4 observation stars successfully identified. After experimental verification, the average recognition time of each frame was calculated to be 8.65 ms. The recognition rate is counted, and the recognition results are shown in fig. 7 to 8. Fig. 7 shows the recognition rate of the present invention under the position noise, and it can be seen from the figure that the recognition rate can reach more than 98% when the position noise is less than 2 pixels. Fig. 8 shows the recognition result of the present invention under the star noise, since the present invention screens the dark stars greatly affected by the brightness at the beginning of the star map recognition, it can be seen from the figure that when the star noise reaches 1Mv, the present invention still has a recognition rate of more than 99.0%, and shows a certain robustness to the brightness noise. The star sensor is easily influenced by sunlight reflected by space debris and other close-range space objects in space, and a false star with light spots similar to a real star can be generated in a view field. Fig. 9 shows the identification result of the present invention under the pseudostellar noise, and it is seen from the figure that after 5 pseudostellars are added to the simulated stellar map, the present invention still obtains the identification rate of more than 98.5%, and can meet the identification requirement. Fig. 10 shows the identification result of the simulation star map used by the method, and the star with the mark number is the observation star identified by the method, and the serial number of the navigation star corresponding to the observation star. Generally, the invention can embody better recognition effect on the simulated star map and can meet the experiment requirement.
In summary, the present invention provides a star map identification method based on a radial triangle mapping matrix, including: extracting a radial triangular mode of the navigation satellite and establishing a navigation satellite mode library; according to the geometric distribution of observation stars extracted from a star map, a radial triangle which is formed by taking a main star as a vertex and any two adjacent stars within a mode radius is constructed, the angular distance of the radial triangle is mapped to a mode matrix, voting is completed by matching each grid of the mode matrix, a high-vote candidate star is selected as a main star label through voting, a rotation matrix and an attitude angle are determined for individual low-vote and low-vote candidate stars by utilizing double-vector attitude determination, the radial triangle is verified by a re-projection method, and an observation star verification and identification result is output to obtain the star label. In the same way for other N's close to the main point of the star mapmaxAnd identifying the star to be identified. The invention combines the relative angular distance information between adjacent stars to form a radial triangular mode on the basis of a radial algorithm, inherits the robustness of the triangular algorithm, constructs a mode matrix with rotation invariance, searches the whole mode library when the radial triangular mode is matched and voted every time, and has completeness. The experimental verification is carried out on the simulated star map, and the invention is verified to have better robustness on star point position noise, star and other noises and pseudostar noise and have application value.
It is to be understood that the above-described embodiments of the present invention are merely illustrative of or explaining the principles of the invention and are not to be construed as limiting the invention. Therefore, any modification, equivalent replacement, improvement and the like made without departing from the spirit and scope of the present invention should be included in the protection scope of the present invention. Further, it is intended that the appended claims cover all such variations and modifications as fall within the scope and boundaries of the appended claims or the equivalents of such scope and boundaries.
Claims (9)
1. A star map identification method based on a radial triangular mapping matrix is characterized by comprising the following steps: the method comprises the following steps:
s100, constructing a radial triangular mode and simultaneously establishing a radial triangular mode star atlas identification database;
s200, constructing a radial triangular mode according to the positions of adjacent stars in the mode radius of each main star in the star map;
s300, sequentially extracting observation stars closest to a principal point, and identifying the observation stars by using a voting method according to a radial triangular mode matrix of each observation star in the star map;
and S400, selecting other eight observation stars close to the image main point as main stars, calculating the angular distances from the main stars to the satellite stars one by one, searching radial triangles matched within the mode radius in a radial triangle mode star map identification database by matching with the constructed index table, and identifying all the stars to be observed.
2. The radial triangle mapping matrix-based star map identification method according to claim 1, wherein: the step S100 includes:
step S110, establishing a navigation satellite database;
step S120, constructing a radial triangular mode library;
step S130, a mode index library is established, and the compression mode library is rapidly searched.
3. The star map identification method based on radial triangle mapping matrix as claimed in claim 2, wherein the step S110, establishing the navigation star database comprises:
determining the radius of a main satellite and a mode, calculating the distance from the main satellite to each adjacent satellite, sequencing the brightness and the angular distance of the adjacent satellites around the main satellite, screening the adjacent satellites, and keeping the brightest adjacent satellites not more than 9 adjacent satellites close to the main satellite in the main satellite mode.
4. The star map identification method based on radial triangle mapping matrix as claimed in claim 3, wherein the step S110, establishing the navigation star database further comprises:
s111, removing stars which influence the construction of the robust mode in the star catalogue, and firstly removing dark stars such as stars and the like exceeding 6 Mv;
step S112, two navigation stars with the star point angular distance value smaller than 0.1 degree are removed as double stars;
step S113, removing the star catalogue navigation star containing less than three adjacent stars within the defined mode radius.
5. The radial triangle mapping matrix-based star map identification method according to claim 4, wherein the step S120 of constructing a radial triangle pattern library comprises:
randomly selecting two stars around the main star and the main star to form a radial triangle, radially dividing the mode radius of the main adjacent star pair, and sequentially quantizing the angular distances of the two main adjacent stars of each radial triangle into 256 intervals according to the anticlockwise direction.
6. The radial triangle mapping matrix-based star map identification method according to claim 5, wherein said step S130 is to establish a mode index database, and the step of implementing fast lookup by the compression mode database comprises:
forming an index from two sides of a radial triangle
Quantizing the value of the third edge
Mapping to 256-length mode matrix, and when multiple radial triangles are mapped to the grid represented by the same two-dimensional matrix, obtaining 1-norm of all elements in the grid matrix M
And stores the main star label and the number n of radial triangles mapped to the same grids. And storing the radial triangular modes corresponding to all navigation stars with the mode radius into a two-dimensional matrix grid to form the radial triangular mode.
Establishing a radial triangular compression mode database, constructing a 256 x 256 mode matrix by the radial triangular modes extracted by each navigation satellite, extracting the non-zero values of the grid modes of all the navigation satellites only by the compression mode matrix database, wherein the non-zero value of each grid mode corresponds to a mode value
Major star number SnAnd the number of repetitions nsAnd stored uniformly in a pattern table. Numbering each section of the compression mode library, and constructing the radial triangle index and the compression mode library number into a corresponding lookup relation index f (e)i,ej) Each of each column of the compressed mode library can be quickly aligned by indexing
The value is looked up.
7. The star map identification method based on the radial triangle mapping matrix according to claim 6, wherein the step S200 of constructing the radial triangle pattern according to the positions of the neighboring stars within the pattern radius of each main star in the star map comprises:
calculating each observation star in the star map to the principal point (u)0,v0) Sequentially selecting the Euclidean distances not exceeding N according to ascending sequence of the distance valuesmaxThe star of (a) is an observation star, wherein N ismax≤8；
And sequentially calculating Euclidean distance values from adjacent stars to the main star in the mode radius by taking the main star as the center of a circle and taking the 6-degree angular distance as the mode radius, sequencing the Euclidean distance values in ascending order, screening the adjacent stars in the star map according to the method for selecting the adjacent stars in the step S120, and establishing a radial triangular mode for each navigation star after the angular distance is quantized.
8. The star map identification method based on the radial triangular mapping matrix according to claim 7, wherein the step S300 sequentially extracts observation stars closest to the principal point, and the step of identifying the observation stars by using a voting method according to the radial triangular pattern matrix of each observation star in the star map comprises:
and S310, sequentially extracting the observation stars closest to the principal point, and voting according to the radial triangular mode of the observation stars.
And S320, sequencing the corresponding counting table of each observation satellite in the radial triangular mode according to the counting value in a descending order, wherein the value in the counting table is the ticket number of each navigation satellite, and taking the first-ranked main satellite candidate as a main satellite identification result when the number of the first-ranked main satellite candidate ticket of the main satellite is two times or more higher than the number of the second-ranked main satellite candidate ticket of the main satellite.
And step S330, verifying and identifying other higher ticket identification results which cannot be confirmed in the identification view field by verifying and identifying links and utilizing the star result of identification and a double-loss reprojection method.
9. The star map recognition method based on radial triangle mapping matrix as claimed in claim 8, wherein in step S400, when the selected primary star recognition is finished, the method returns to step S200, and selects the second observation star as the primary star for recognition until the nth star is recognizedmaxThe star is observed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110723788.9A CN113483751B (en) | 2021-06-29 | 2021-06-29 | Radial triangle mapping matrix-based star map identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110723788.9A CN113483751B (en) | 2021-06-29 | 2021-06-29 | Radial triangle mapping matrix-based star map identification method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113483751A true CN113483751A (en) | 2021-10-08 |
| CN113483751B CN113483751B (en) | 2022-07-05 |
Family
ID=77936184
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110723788.9A Active CN113483751B (en) | 2021-06-29 | 2021-06-29 | Radial triangle mapping matrix-based star map identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113483751B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114719844A (en) * | 2022-04-06 | 2022-07-08 | 北京信息科技大学 | Space projection-based all-celestial star map identification method, device and medium |
| CN115546323A (en) * | 2022-08-31 | 2022-12-30 | 广东工业大学 | Image compression reconstruction method based on spherical coordinate principal component analysis |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4104722A (en) * | 1976-09-10 | 1978-08-01 | Evans Dow J | Method and apparatus for celestial navigation |
| CN1796938A (en) * | 2004-12-28 | 2006-07-05 | 北京航空航天大学 | Method for recognising star map based on triangle character |
| CN102072730A (en) * | 2010-11-04 | 2011-05-25 | 北京航空航天大学 | Method for identifying triangular star map |
| CN103363987A (en) * | 2013-06-26 | 2013-10-23 | 哈尔滨工业大学 | Star map identification method of multi-view-field star sensor |
| CN103453905A (en) * | 2013-08-07 | 2013-12-18 | 北京航空航天大学 | Star pattern recognition method based on principal component analysis of plane triangles |
| CN106931964A (en) * | 2017-01-19 | 2017-07-07 | 中国人民解放军国防科学技术大学 | Attitude determination method and star sensor based on compressed sensing imaging |
| CN107816986A (en) * | 2017-09-05 | 2018-03-20 | 湖北航天技术研究院总体设计所 | A kind of star sensor whole day ball fast star identification method |
| CN107883947A (en) * | 2017-12-28 | 2018-04-06 | 常州工学院 | Star sensor method for recognising star map based on convolutional neural networks |
| CN108614237A (en) * | 2018-03-19 | 2018-10-02 | 西北大学 | Fingerprint base based on compressed sensing is established and RFID localization methods |
| CN112179335A (en) * | 2020-09-17 | 2021-01-05 | 北京航空航天大学 | Star map identification method based on simplest general subgraph |
| CN112729277A (en) * | 2020-12-29 | 2021-04-30 | 中国科学院微小卫星创新研究院 | Star sensor star map identification method based on dynamic included angle matching |
-
2021
- 2021-06-29 CN CN202110723788.9A patent/CN113483751B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4104722A (en) * | 1976-09-10 | 1978-08-01 | Evans Dow J | Method and apparatus for celestial navigation |
| CN1796938A (en) * | 2004-12-28 | 2006-07-05 | 北京航空航天大学 | Method for recognising star map based on triangle character |
| CN102072730A (en) * | 2010-11-04 | 2011-05-25 | 北京航空航天大学 | Method for identifying triangular star map |
| CN103363987A (en) * | 2013-06-26 | 2013-10-23 | 哈尔滨工业大学 | Star map identification method of multi-view-field star sensor |
| CN103453905A (en) * | 2013-08-07 | 2013-12-18 | 北京航空航天大学 | Star pattern recognition method based on principal component analysis of plane triangles |
| CN106931964A (en) * | 2017-01-19 | 2017-07-07 | 中国人民解放军国防科学技术大学 | Attitude determination method and star sensor based on compressed sensing imaging |
| CN107816986A (en) * | 2017-09-05 | 2018-03-20 | 湖北航天技术研究院总体设计所 | A kind of star sensor whole day ball fast star identification method |
| CN107883947A (en) * | 2017-12-28 | 2018-04-06 | 常州工学院 | Star sensor method for recognising star map based on convolutional neural networks |
| CN108614237A (en) * | 2018-03-19 | 2018-10-02 | 西北大学 | Fingerprint base based on compressed sensing is established and RFID localization methods |
| CN112179335A (en) * | 2020-09-17 | 2021-01-05 | 北京航空航天大学 | Star map identification method based on simplest general subgraph |
| CN112729277A (en) * | 2020-12-29 | 2021-04-30 | 中国科学院微小卫星创新研究院 | Star sensor star map identification method based on dynamic included angle matching |
Non-Patent Citations (3)
| Title |
|---|
| JIAN LI 等: ""Improved Grid Algorithm Based on Star Pair Pattern and Two-dimensional Angular Distances for Full-Sky Star Identification"", 《IEEE ACCESS》 * |
| WANG, GY 等: ""Star Identification Based on Hash Map"", 《IEEE SENSORS JOURNAL》 * |
| 魏新国等: "利用径向和环向分布特征的星图识别方法", 《光电工程》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114719844A (en) * | 2022-04-06 | 2022-07-08 | 北京信息科技大学 | Space projection-based all-celestial star map identification method, device and medium |
| CN115546323A (en) * | 2022-08-31 | 2022-12-30 | 广东工业大学 | Image compression reconstruction method based on spherical coordinate principal component analysis |
| CN115546323B (en) * | 2022-08-31 | 2023-06-09 | 广东工业大学 | Image compression reconstruction method based on spherical coordinate principal component analysis |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113483751B (en) | 2022-07-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| AU2016315938B2 (en) | Systems and methods for analyzing remote sensing imagery | |
| US9875404B2 (en) | Automated metric information network | |
| CN113483751B (en) | Radial triangle mapping matrix-based star map identification method | |
| David et al. | Softposit: Simultaneous pose and correspondence determination | |
| CN103337052B (en) | Automatic geometric correcting method towards wide cut remote sensing image | |
| CN109813303B (en) | Star map identification method independent of calibration parameters based on angular pattern cluster voting | |
| CN107816986B (en) | Rapid star map identification method for all celestial sphere of star sensor | |
| CN107816987B (en) | Star pattern recognition method based on spider web pattern and convolutional neural network | |
| CN111768498A (en) | Visual positioning method and system based on dense semantic three-dimensional map and mixed features | |
| CN112150358A (en) | Image feature matching method for resisting large geometric distortion | |
| CN114166211B (en) | Double-view-field star sensor star map identification method | |
| Du et al. | Star identification based on radial triangle mapping Matrix | |
| CN112179335B (en) | Star map identification method based on simplest general subgraph | |
| Zhu et al. | A robust and fast star identification algorithm based on an ordered set of points pattern | |
| CN103925922A (en) | Method for identifying fixed stars in ICCD star atlas under high-dynamic condition | |
| CN111174776B (en) | Radial and dynamic circumferential mode-based star map identification method | |
| Wiguna et al. | Evaluation of deep learning models for building damage mapping in emergency response settings | |
| CN116659485A (en) | Quick robust global star map identification method | |
| US20220276046A1 (en) | System and method for providing improved geocoded reference data to a 3d map representation | |
| CN107341151B (en) | Image retrieval database generation method, and method and device for enhancing reality | |
| CN115290100A (en) | Partitioned fast star map identification method based on attitude information | |
| CN115507844A (en) | Star map identification method suitable for ultra-large field of view | |
| KR101749231B1 (en) | Celestial Navigation Method of a Satellite using Star Tracker and Horizon Sensor | |
| Yan et al. | Horizontal velocity estimation via downward looking descent images for lunar landing | |
| CN111024063B (en) | Star map recognition algorithm based on star point re-extraction under large maneuvering condition |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |