CN106123891B - Faint star autonomous classification method based on perimeter feature - Google Patents
Faint star autonomous classification method based on perimeter feature Download PDFInfo
- Publication number
- CN106123891B CN106123891B CN201610457479.0A CN201610457479A CN106123891B CN 106123891 B CN106123891 B CN 106123891B CN 201610457479 A CN201610457479 A CN 201610457479A CN 106123891 B CN106123891 B CN 106123891B
- Authority
- CN
- China
- Prior art keywords
- star
- triangle
- primary
- angular distance
- observation
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 26
- 238000010276 construction Methods 0.000 claims abstract description 13
- 238000004422 calculation algorithm Methods 0.000 claims description 17
- 230000000007 visual effect Effects 0.000 claims description 17
- 239000011159 matrix material Substances 0.000 claims description 9
- 230000003287 optical effect Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 5
- 230000001174 ascending effect Effects 0.000 claims description 4
- 238000005192 partition Methods 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 3
- 230000002194 synthesizing effect Effects 0.000 claims description 3
- 238000001514 detection method Methods 0.000 claims description 2
- 239000000284 extract Substances 0.000 abstract description 2
- 238000005259 measurement Methods 0.000 description 6
- 238000003909 pattern recognition Methods 0.000 description 3
- 239000000047 product Substances 0.000 description 3
- 230000007812 deficiency Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
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
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
The faint star autonomous classification method based on perimeter feature that the present invention relates to a kind of, includes the following steps：One, according to fundamental catalog, navigational star table is constructed using magnitude threshold method.Whole day ball is divided by several part overlappings day area using overlapping sphere moments method, two stars most bright in each overlapping day area are found out as primary pair, remaining star constructs triangle perimeter feature database as auxiliary star, presses the big little structure hash function of perimeter to the feature database of generation, carries out fragmented storage.Two, navigational star table and navigation characteristic library are read in, star chart extracts result and arranges by gray value size descending, and using two most bright stars as primary pair, remaining star is auxiliary star, construction, to the observation triangle for common edge, calculates triangle perimeter and using primary as two side length of triangle of common point with primary.Present invention substantially reduces the quantity of feature triangle, relatively traditional triangle recognizer based on star angular distance reduces star angular distance number of comparisons, improves recognition speed.
Description
Technical field
The faint star autonomous classification method based on perimeter feature that the present invention relates to a kind of, belongs to celestial navigation field.
Background technique
Star sensor is to be currently known the highest attitude of carrier sensor of precision.Star sensor by measurement fixed star be directed toward come
Resolve the posture of carrier.Importance in star map recognition is the key technology of star sensor, and rapidity and robustness are always the weight evaluated it
Want index.Typical star Pattern Recognition Algorithm mainly has triangle and its innovatory algorithm (primary algorithm, pyramid algorith etc.) and grid
Lattice algorithm.When star catalogue size is smaller, these algorithms have good robustness, and work as star sensor visual field (Filed of
View, FOV) it is smaller, when limited detective magnitude is higher, star catalogue size will sharply increase, and the limitation of related algorithm will show,
If triangle algorithm intrinsic dimensionality is low, the capacity that the increase of nautical star quantity will lead to feature database is sharply increased, to be easy
Existing error hiding or Redundancy Match；Pyramid algorith using inner product of vector as characteristic value, discrimination is owed within the scope of small angular distance
It is good；Raster is less in celestial body number and when pseudolite is more has robustness decline；Primary recognizer is in visual field
When bright star similar in the middle many magnitudes of appearance, recognition success rate will be reduced seriously.
Sometimes for higher attitude measurement accuracy is pursued in engineering, need star sensor that there is lesser visual field and darker
Magnitude detectivity, need to solve the problems, such as the importance in star map recognition in the case of faint star at this time.Solving faint star star Pattern Recognition Algorithm has two
Kind technological approaches：First is that local day area assists in identifying.Second is that the improvement to existing algorithm.Local day area, which assists in identifying, utilizes auxiliary
The correlation measurement information that system (such as inertial navigation system) provides determines the thick direction of star sensor, constructs Local Navigation feature three
It is angular, it is identified on this basis using existing star Pattern Recognition Algorithm.The main deficiency that local day area assists in identifying is identification
Process heavy dependence external information, loses the independence of star sensor.In terms of algorithm improvement, the importance in star map recognition side of small field of view faint star
Face document is few, and Li Hui improves pyramid algorith, using vector apposition as matching characteristic value, realizes the pole 8Mv
The importance in star map recognition (2.32 ° × 1.73 ° of visual field) of magnitude is limited, the deficiency of the algorithm is special using the vector apposition of two stars as matching
Sign, the identification of each triangle still needs to be searched for three times, in addition, setting a minimum to reduce the capacity of feature database
Critical star angular distance only retains a ludcida body within the scope of this angular distance, thus it is possible that observation star cannot identify
Problem.
Summary of the invention
The technical problem to be solved by the present invention is to provide a kind of faint star based on perimeter feature for the above-mentioned prior art
Autonomous classification method, the problems such as with matching speed when solving small field of view faint star importance in star map recognition is slow, Redundancy Match is more.
The present invention solve the above problems used by technical solution be：A kind of faint star autonomous classification side based on perimeter feature
Method the described method comprises the following steps：
Step 1: the building of nautical star feature database
A, the screening of nautical star
(1) the fixed star for being greater than star sensor limit detection star in fundamental catalog is rejected using magnitude threshold method value；
(2) double star is handled
When the star angular distance between two fixed stars be less than disc of confusion when, double star is handled as a star, and to fixed star right ascension,
Declination and magnitude are synthesized, then fixed star right ascension α, declination δ and the magnitude m after synthesizing are respectively：
Wherein, disc of confusion size is between 3 × 3~5 × 5 pixels, (α1, δ1) and (α2, δ2) it is respectively the red of two fixed stars
Through and declination, m1And m2The magnitude of respectively two fixed stars；
(3) the generation of navigational star table
Selected fixed star is renumberd, and the star catalogue number of each fixed star, right ascension, declination and magnitude are stored；
B, the building in navigation characteristic library
In conjunction with primary algorithm construction feature triangle：Using in visual field ludcida and time bright star as primary pair, remaining star
As auxiliary star, respectively with primary to constitutive characteristic triangle, each feature triangle includes：It is primary asterisk, secondary star asterisk, auxiliary
Help star angular distance between star number, primary and secondary star, star angular distance and triangle perimeter these data between primary and auxiliary star,
(1) the building of zone catalog
It is divided into several to be locally overlapped day area in entire celestial sphere using overlapping sphere moments method, ball square size should meet following item
Part：
In formula, Cα、CδRespectively right ascension range and declination range, α0、δ0For star sensor optical axis direction, R is visual field half
Diameter, need to be to C when calculating ball square sizeαZero passage processing, declination maximum value, minimum value are set to 90 ° and -90 °, establish zone star
When table, centered on the star sensor optical axis, ball square size is calculated, institute's any stars in ball square are arranged by magnitude ascending order, and deposits
It stores up using partition number as in the text file of filename, the calculation formula of subregion call number is as follows,
Index=(fix (α0/4)+1)×(fix(δ0/4)+23)
(2) the foundation of feature database
Zone catalog is read, with selected primary to other fixed star construction feature triangles, reading related star catalogue respectively
Number, right ascension, declination and stars and other information, then calculate related star angular distance d according to the following formulaij,
dij=arccos (Vi·Vj/(||Vi||×||Vj||))
In formula, " " is vector dot product operator；" | | x | | " is the norm of x, ViAnd VjThe reference of respectively fixed star i, j is sweared
Amount, when construction feature triangle, star angular distance dijFollowing condition should be met：
In formula, ξ is angle corresponding to disc of confusion, and FOV indicates field angle, if three sides of a triangle are all satisfied as broadcast TV programs by satellite
Angular distance restrictive condition then stores the triangle according to navigation characteristic data structure.
(3) the building of perimeter hash feature database
Judge whether all primarys calculate to finish, otherwise repeats (2), be then to all feature triangles according to perimeter size
Piecemeal storage, sub-block number construct hash function K (P) as the following formula：
K (P)=fix (P × N/ (6FOV)+1) (1)
In formula, P is triangle perimeter, and N is piecemeal sum；
Step 2: importance in star map recognition
(1) initialize：Navigational star table, navigational triangle feature database are read in, star sensor parameter is set and star angular distance matches door
Limit ε parameter；
(2) data are extracted to star chart to be numbered, and arrange by gray value descending；
(3) judge whether there is 3 or more and observe star, cannot then be identified less than 3；
If (4) observation star number is 3, primary pair and auxiliary star, building observation triangle, calculating observation are determined by gray value
The correspondence star angular distance and perimeter of triangle position sub-block number according to formula (1), search for two using primary as vertex in this sub-block
A star angular distance is that then matching result is stored in recognition matrix, wherein i-th with the presence or absence of the triangle for uniquely meeting following formula (2)
The star catalogue number of a element representation asterism i, then the element sets 0 when recognition failures；
In formula, A, B, C are respectively primary, secondary star and auxiliary star, | x | indicate the absolute value of x, lABIndicate primary A and secondary star B it
Between observation star angular distance, dABIndicate the reference star angular distance between primary A and secondary star B, lACIndicate the sight between primary A and auxiliary star C
Survey star angular distance, dACIndicate the reference star angular distance between primary A and auxiliary star C；
(5), if observation star is greater than 3, two observation triangles are constructed with asterism 1,2,3 and asterism 1,2,4 respectively, and press
(4) step is identified that, if two identical primarys of observation triangle existence anduniquess are to matching result, successful match will respectively
In the matching result deposit recognition matrix of preceding 4 stars, then basis has matched observation star and has projected to fixed stars other in star chart
Identification, and recognition result is verified using observation star angular distance；If two observation triangles there is no unique match as a result, if pick
Except asterism 3 or asterism 4, reads in next star and continue to identify；
(6) judge whether the frame star chart identifies to finish, otherwise turn (5), be then to be determined to see according to recognition matrix and navigational star table
The parameters for surveying star, are transferred to next frame importance in star map recognition, until entire identification mission finishes.
Compared with the prior art, the advantages of the present invention are as follows：
The invention firstly uses overlapping sphere moments methods to be divided into several part overlappings day area for whole day ball, finds out each overlapping day
Two most bright stars form primary pair in area, remaining star constructs triangle perimeter feature database as auxiliary star；Secondly, to generation
Feature database constructs hash function by triangle perimeter, carries out fragmented storage；In identification process, realized according to observation triangle perimeter
The identification of the quick positioning of sub-block, star angular distance only carries out in corresponding sub-block, to improve matching speed；Finally to algorithm into
It has gone verifying, has realized the fast star identification of 9.0Mv star.
Detailed description of the invention
Fig. 1 is coordinate system definition figure, wherein On-XnYnZnFor J2000.0 coordinate system, Os-XsYsZsFor star sensor coordinate
System, (α, δ, γ) are respectively star sensor yaw angle, pitch angle and roll angle, it reflect star sensor coordinate system with
Relationship between J2000.0 coordinate system.
Fig. 2 is that overlapping sphere moments method divides schematic diagram to celestial sphere, and FOV is the rectangular visual field of star sensor in figure.
Specific embodiment
The present invention will be described in further detail below with reference to the embodiments of the drawings.
To be described convenient for the content of present invention, at this to the definition of coordinate system involved in invention, related notion and symbol meaning,
The brief description such as single star sensor attitude determination method.
1, relative coordinate system defines
The present invention relates to following two coordinate systems：
Relationship between J2000.0 coordinate system and star sensor coordinate system is as shown in Figure 1.
2, related notion and symbol meaning
(1) measurement vector ViWith reference vector Wi
Assuming that the right ascension declination of fixed star i is respectively (αi, δi), the coordinate in corresponding star sensor coordinate system is respectively
(xi, yi), unit is pixel number, then the reference vector of fixed star i is respectively with measurement vector：
Vi=[cos αi cosδi sinαi cosδi sinδi]T (1)
In formula, D is imaging sensor pixel dimension, (x0,y0) be star sensor principal point coordinate), unit is pixel number.
(2) star angular distance
Star angular distance is the most basic characteristic value of importance in star map recognition, it reflects two fixed stars to the subtended angle of observer.Assuming that permanent
The reference vector of star i, j are respectively ViAnd Vj, measurement vector is respectively WiAnd Wj, then the reference star angular distance d of fixed star i, jijWith observation
Star angular distance lijCalculation formula be respectively：
dij=arccos (Vi·Vj/(||Vi||×Vj)) (3)
lij=arccos (Wi·Wj/(||Wi||×||Wj||)) (4)
In formula, " " is vector dot product operator；" | | x | | " is the norm of x.
Two, the building of nautical star feature database
Importance in star map recognition be the fixed star in star sensor current field is carried out with the fixed reference feature in navigation characteristic library it is corresponding
Matching, with parameters such as the star catalogues number, right ascension, declination that determine in visual field related fixed star, be subsequent attitude of carrier determine it is important before
It mentions.Therefore, the building of feature database is the basis of importance in star map recognition.
The building in navigation characteristic library is divided into following two step：
1. the screening of nautical star
(1) according to the limited detective magnitude of star sensor, using magnitude threshold method value (Visual Magnitude
Threshold, VMT) reject the fixed star for being greater than star sensor limited detective magnitude in fundamental catalog.
(2) double star is handled
Double star (Double Star) refers to two fixed stars that cannot be distinguished one another on star sensor imaging plane herein,
Namely lesser two fixed stars of star angular distance.Star sensor generally uses defocus technology, and disc of confusion size is in 3 × 3~5 × 5 pictures
Between member, therefore when the star angular distance between two fixed stars is less than disc of confusion, double star can be handled as a star at this time, at this time
Relevant parameter need to be synthesized.Assuming that the right ascension of two stars, declination are respectively (α1, δ1) and (α2, δ2), magnitude is respectively m1With
m2, then fixed star right ascension α, declination δ and magnitude m after synthesizing are respectively：
(3) the generation of navigational star table
Selected fixed star is renumberd, and is stored by following format：
2. the building in navigation characteristic library
It is the necessary condition of equivalent of triangle that perimeter is equal, therefore when triangle identifies, can be based on the think of of hash search
Think, first the equal triangle sets of search perimeter, searches for the matched triangle in both sides, in the set then so as to subtract
Few matching times.Primary algorithm construction feature triangle can be combined thus：Using in visual field ludcida and time bright star as leading
Star is to (Main Star-pair), remaining star is as auxiliary star, respectively with primary to constitutive characteristic triangle, specific data structure
For：
(1) the building of zone catalog
Theoretically n star can constructA feature triangle, even if plus constraint conditions such as double star and visual fields, feature three
Angular quantity is also very huge, and there are the feature triangles of many redundancies, constructs triangle perimeter feature database at this time, will
Face the calculating of magnanimity.For the calculation amount for reducing construction feature triangle, overlapping sphere moments method can be used and be divided into entire celestial sphere
Several are locally overlapped day area, to reduce influence of across the day area of visual field to recognition correct rate, as shown in Figure 2.
Ball square size should meet following condition：
In formula, Cα、CδRespectively right ascension range and declination range, α0、δ0It is directed toward (right ascension, declination) for the star sensor optical axis,
R is visual field radius.It, need to be to C when calculating ball square size in view of right ascension, the angular range of declinationαZero passage processing, declination are maximum
Value, minimum value are set to 90 ° and -90 °.Assuming that by stages is every taking 4 °, visual field radius R takes 4 °, then when establishing zone catalog, with
Centered on the star sensor optical axis (for right ascension from 0 °, declination is from -90 °), ball square size is calculated, institute's any stars in ball square are pressed
The arrangement of magnitude ascending order, and store to using partition number as in the text file of filename, the calculation formula of subregion call number such as formula (9)
It is shown, shown in text data format such as formula (6)：
Index=(fix (α0/4)+1)×(fix(δ0/4)+23) (9)
It follows that first star and second star in zone catalog are respectively primary and secondary star, other fixed star conducts
Assist star that can greatly reduce the quantity of feature triangle respectively with primary, secondary star construction feature triangle.In order to improve star chart
The adaptability of recognizer can regard fixed star brighter in zone catalog as primary pair, suitably to increase the quantity of feature database.
(2) the foundation of feature database
Read zone catalog, with selected primary to respectively with other fixed star construction feature triangles.Read related star catalogue
Number, the information such as right ascension and declination, then related star angular distance d is calculated according to formula (1)ij。
Optical system of star sensor uses Defocusing design generally to improve asterism segmented positioning precision, and disc of confusion size is general
It controls between 3 × 3~5 × 5 pixels, therefore when construction feature triangle, star angular distance dijFollowing condition should be met：
In formula, ξ is angle corresponding to disc of confusion.If three sides of a triangle are all satisfied such as angular distance restrictive condition of broadcasting TV programs by satellite,
Then the triangle is stored according to navigation characteristic data structure.
(3) the building of perimeter hash feature database
Judge whether all primarys calculate to finish, otherwise repeatedly (2) step, is then to all feature triangles according to perimeter
The storage of size piecemeal, sub-block number construct hash function K (P) as the following formula：
K (P)=fix (P × N/ (6FOV)+1) (11)
In formula, P is triangle perimeter, and N is piecemeal sum.
Three, importance in star map recognition
Triangle character library established above is to reduce volumes of searches to lay a good foundation, and it is fixed that a triangle need to only carry out once
Position and a partitioned searching can be completed, and specific algorithm process is as follows：
(1) initialize：Navigational star table, navigational triangle feature database are read in, setting star sensor parameter, star angular distance match door
Limit the parameters such as ε；
To star chart extract data be numbered (asterism i), and by gray value descending arrange；
(3) judge whether there is 3 or more and observe star, cannot then be identified less than 3；
If (4) observation star number is 3, primary pair and auxiliary star, building observation triangle, calculating observation are determined by gray value
The correspondence star angular distance and perimeter of triangle position sub-block number according to formula (11), are searched in this sub-block using primary as vertex
Two star angular distances whether there is the triangle for uniquely meeting following formula, be that then matching result is stored in recognition matrix Identify,
The wherein star catalogue number of i-th of element representation asterism i, then the element sets 0 when recognition failures；
In formula, A, B, C are respectively primary, secondary star and auxiliary star, | x | indicate the absolute value of x.
(5), if observation star is greater than 3, two observation triangles are constructed with asterism 1,2,3 and asterism 1,2,4 respectively, and press
(4) step is identified respectively.If two identical primarys of observation triangle existence anduniquess are to matching result, successful match will
In the matching result deposit recognition matrix Identify of preceding 4 stars, then basis has matched observation star to fixed stars other in star chart
Projection identification is carried out, and recognition result is verified using observation star angular distance；If two observation triangles do not have unique match
As a result, then rejecting asterism 3 or asterism 4, reads in next star and continue to identify；
(6) judge whether the frame star chart identifies to finish, otherwise turn (5), be then according to recognition matrix Identify and nautical star
Table determines the parameters such as star catalogue number, right ascension, declination and the apparent magnitude of observation star, is transferred to next frame importance in star map recognition, until entire identification
Task finishes.
Embodiment：
1, condition and parameter
Example platforms are computer, and operating system is Windows XP operating system, and programming tool software is
Matlab2011a, it is as shown in the table for other parameters：
2, nautical star feature database constructs example
(1) navigational star table screens
According to fundamental catalog, the navigational star table met the requirements is filtered out.Generally comprised in fundamental catalog star catalogue number, right ascension,
Information, the specific example such as declination, magnitude, spectral signature and right ascension, proper motion in declination parameter are as follows：
According to magnitude threshold value and star sensor system parameter, screening meets the 9.0Mv navigational star table of above-mentioned condition, to selected
Fixed star renumber, and stored by nautical star data format, specific example is as follows：
The nautical star for meeting condition shares 119528.
(2) the building in navigation characteristic library
A. the building of zone catalog
Centered on the star sensor optical axis, entire celestial sphere is divided into 4050 parts overlapping day according to overlapping sphere moments method
Area, wherein right ascension α0To 360 ° (being free of) from 0 °, declination δ0To 90 ° from -90 °, 4 ° are spaced, visual field radius R takes 4 °.It calculates
Institute's any stars in ball square are arranged by magnitude size ascending order, and stored to using partition number as the text of filename by ball square size
In file, wherein the specific fixed star such as following table institute in the 1st subregion (it is respectively -90 ° of 0 ° of right ascension and declination that the corresponding optical axis, which is directed toward)
Show：
B. the building in subregion navigation characteristic library
Zone catalog is read, using ludcida and time bright star as primary pair, other stars are respectively with primary to construction feature triangle
Shape.The information such as related star catalogue number, right ascension and declination are read, then related star angular distance d are calculated according to formula (1)ij。
By subregion navigational star table with navigation characteristic library it is found that the quantity of the feature triangle thus generated significantly reduces.
C. the building of perimeter hash feature database
Judge whether all primarys calculate to finish, otherwise repeatedly (2) step, is then to all feature triangles according to perimeter
The storage of size piecemeal.
It is 6054793, file size 267MB by the triangle perimeter feature database sum that above method constructs, by it
512 sub- block files are divided by perimeter size, block feature number is up to 45936, and minimum is 1, wherein the 60th piecemeal
Characteristic is as follows：
2, importance in star map recognition example
With 28 days 20 April in 2013:38:33:For star chart captured by 500：
(1) initialize：Navigational star table, navigational triangle feature database are read in, setting star sensor parameter, star angular distance match door
Limit the parameters such as ε；
(2) data are extracted to star chart and be numbered that (asterism i), and arranging by gray value descending, specific to extract result as follows
It is shown：
(3) star pattern matching
A. it is extracted according to star chart as a result, triangle perimeter characteristic value is calculated by formula (2), according to perimeter characteristic value by formula (11)
Navigation characteristic library text piece number corresponding to perimeter characteristic value is calculated, concrete outcome is as follows：
B. the 1st triangle is identified, perimeter feature database corresponding to the 1st triangle is read in, wherein the 218th
Perimeter feature database is as follows：
In this feature database, search observation star angular distance and reference star angular distance meet star angular distance match thresholding star angular distance AB,
Star angular distance AC data acquisition system, concrete outcome are as follows：
It follows that the matching result of asterism 1,2,3 is not unique, needs to introduce the 4th star at this time and assisted in identifying, i.e.,
Supplement identification is carried out by the 2nd triangle, further to exclude Redundancy Match as a result, the recognition result of the 2nd triangle is as follows
It is shown：
Two above triangle has common edge AB, therefore can pass through asterism 1,2 star catalogues in search two above star catalogue
Number identical set, can exclude most Redundancy Match results.Compare two above table, only one group of 1 He of asterism
The star catalogue number of asterism 2 simultaneously equal (overstriking data), therefore the star catalogue number of asterism 1,2,3,4 can be determined as 80284,64129,
64051 and 64049.Under normal circumstances, by two above triangle map, unique consequence can be obtained.If two above three
Angular no matching result illustrates in preceding four asterisms, matches thresholding beyond star angular distance with the presence of pseudolite or observed result, this
When can reject an asterism, introduce next star and matched again.
C. judge whether the frame star chart identifies to finish, be the identification for being transferred to next frame star chart, otherwise, introduce next star
Point, and asterism 1,2 constitutive characteristic triangles, and using the recognition result of asterism 1~3 as criterion, to accelerate the knowledge of subsequent asterism
Other speed.
The final recognition result of this example is as shown in the table：
(4) navigation data is exported
According to recognition result and navigational star table, the parameters such as star catalogue number, right ascension, declination and the apparent magnitude of observation star, tool are determined
Body result is as shown in the table, and the recognition time of this frame star chart is 1.3638 seconds.
In addition to the implementation, all to use equivalent transformation or equivalent replacement the invention also includes there is an other embodiments
The technical solution that mode is formed should all be fallen within the scope of the hereto appended claims.
Claims (1)
1. a kind of faint star autonomous classification method based on perimeter feature, it is characterised in that：It the described method comprises the following steps：
Step 1: the building of nautical star feature database
A, the screening of nautical star
(1) the fixed star that brightness in fundamental catalog is greater than the star brightness of star sensor limit detection is rejected using magnitude threshold method；
(2) double star is handled
When the star angular distance between two fixed stars is less than disc of confusion, double star is handled as a star, and to fixed star right ascension, declination
And magnitude is synthesized, then fixed star right ascension α, declination δ and the magnitude m after synthesizing are respectively：
Wherein, disc of confusion size is between 3 × 3~5 × 5 pixels, (α1, δ1) and (α2, δ2) be respectively two fixed stars right ascension and
Declination, m1And m2The magnitude of respectively two fixed stars；
(3) the generation of navigational star table
Selected fixed star is renumberd, and the star catalogue number of each fixed star, right ascension, declination and magnitude are stored；
B, the building in navigation characteristic library
In conjunction with primary algorithm construction feature triangle：Using in visual field ludcida and time bright star as primary pair, remaining star conduct
Star is assisted, respectively with primary to constitutive characteristic triangle, each feature triangle includes：Primary asterisk, secondary star asterisk, auxiliary star
Star angular distance and triangle perimeter these data between star angular distance between asterisk, primary and secondary star, primary and auxiliary star,
(1) the building of zone catalog
It is divided into several to be locally overlapped day area in entire celestial sphere using overlapping sphere moments method, ball square size should meet following condition：
In formula, Cα、CδRespectively right ascension range and declination range, α0It is directed toward for star sensor optical axis right ascension, δ0For star sensor view
Axis declination is directed toward, and R is visual field radius, need to be to C when calculating ball square sizeαZero passage processing, declination maximum value, minimum value are set to
90 ° and -90 °, when establishing zone catalog, centered on the star sensor optical axis, ball square size is calculated, by institute's any stars in ball square
It arranges, and is stored to using partition number as in the text file of filename, the calculation formula of subregion call number is as follows by magnitude ascending order
It is shown,
Index=(fix (α0/4)+1)×(fix(δ0/4)+23)
(2) the foundation of feature database
Zone catalog is read, with selected primary to other fixed star construction feature triangles, reading related star catalogue number, red respectively
Through, declination and stars and other information, then related star angular distance d is calculated according to the following formulaij,
dij=arccos (Vi·Vj/(Vi||×||Vj||))
In formula, " " is vector dot product operator；" | | x | | " is the norm of x, ViAnd VjThe respectively reference vector of fixed star i, j,
When construction feature triangle, star angular distance dijFollowing condition should be met：
In formula, ξ is angle corresponding to disc of confusion, and FOV is field angle, if three sides of a triangle are all satisfied such as angular distance limit of broadcasting TV programs by satellite
Condition processed then stores the triangle according to navigation characteristic data structure.
(3) the building of perimeter hash feature database
Judge whether all primarys calculate to finish, otherwise repeats (2), be then to all feature triangles according to perimeter size piecemeal
Storage, sub-block number construct hash function K (P) as the following formula：
K (P)=fix (P × N/ (6FOV)+1) (1)
In formula, P is triangle perimeter, and N is piecemeal sum；
Step 2: importance in star map recognition
(1) initialize：Navigational star table, navigational triangle feature database are read in, star sensor parameter is set and star angular distance matches thresholding ε
Parameter；
(2) data are extracted to star chart to be numbered, and arrange by gray value descending；
(3) judge whether there is 3 or more and observe star, cannot then be identified less than 3；
If (4) observation star number is 3, primary pair and auxiliary star, building observation triangle, calculating observation triangle are determined by gray value
The correspondence star angular distance and perimeter of shape position sub-block number according to formula (1), search for two stars using primary as vertex in this sub-block
Angular distance is that then matching result is stored in recognition matrix, wherein i-th yuan with the presence or absence of the triangle for uniquely meeting following formula (2)
Element indicates the star catalogue number of asterism i, and then the element sets 0 when recognition failures；
In formula, A, B, C are respectively primary, secondary star and auxiliary star, | x | indicate the absolute value of x, lABIt indicates between primary A and secondary star B
Observe star angular distance, dABIndicate the reference star angular distance between primary A and secondary star B, lACIndicate the observation star between primary A and auxiliary star C
Angular distance, dACIndicate the reference star angular distance between primary A and auxiliary star C；
If (5) observation star is greater than 3, two observation triangles are constructed with asterism 1,2,3 and asterism 1,2,4 respectively, and press step
(4) identified respectively, if two identical primarys of observation triangle existence anduniquess are to matching result, successful match, by preceding 4
In the matching result deposit recognition matrix of star, projection knowledge then is carried out to fixed stars other in star chart according to having matched observation star
Not, and using observation star angular distance recognition result is verified；If two observation triangles there is no unique match as a result, if reject
Asterism 3 or asterism 4 read in next star and continue to identify；
(6) judge whether the frame star chart identifies to finish, otherwise turn (5), be that observation star is then determined according to recognition matrix and navigational star table
Parameters, next frame importance in star map recognition is transferred to, until entire identification mission finishes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610457479.0A CN106123891B (en) | 2016-06-22 | 2016-06-22 | Faint star autonomous classification method based on perimeter feature |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610457479.0A CN106123891B (en) | 2016-06-22 | 2016-06-22 | Faint star autonomous classification method based on perimeter feature |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106123891A CN106123891A (en) | 2016-11-16 |
| CN106123891B true CN106123891B (en) | 2018-11-16 |
Family
ID=57269291
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610457479.0A Expired - Fee Related CN106123891B (en) | 2016-06-22 | 2016-06-22 | Faint star autonomous classification method based on perimeter feature |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106123891B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108469261A (en) * | 2018-02-07 | 2018-08-31 | 天津大学 | A kind of method for recognising star map suitable for boat-carrying ultra-large vision field celestial navigation system |
| CN109813302B (en) * | 2019-01-29 | 2020-12-04 | 中国人民解放军国防科技大学 | Method for quickly determining optimal available navigation satellite |
| CN110954085A (en) * | 2019-12-06 | 2020-04-03 | 中国科学院长春光学精密机械与物理研究所 | Star table partitioning and feature library building method and device applied to star sensor |
| CN110879063B (en) * | 2019-12-12 | 2021-02-26 | 中国科学院长春光学精密机械与物理研究所 | Rapid star map identification method based on triangle matching |
| CN111024063B (en) * | 2019-12-23 | 2023-07-18 | 上海交通大学 | Star map recognition algorithm based on star point re-extraction under large maneuvering condition |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN1896681A (en) * | 2005-07-15 | 2007-01-17 | 北京航空航天大学 | Fast tracting method of star sensor |
| CN101004349A (en) * | 2007-01-22 | 2007-07-25 | 北京航空航天大学 | Method for recognising not gauged star map |
| CN101441082A (en) * | 2008-12-22 | 2009-05-27 | 中国科学院长春光学精密机械与物理研究所 | Rapid triangle star map recognition method |
| CN102313549A (en) * | 2011-07-28 | 2012-01-11 | 北京航空航天大学 | Identification method for triangular star atlas based on characteristic of inertia ratio |
| CN102840861A (en) * | 2012-09-17 | 2012-12-26 | 常州工学院 | Navigational star screening method for star sensors |
| CN103148852A (en) * | 2013-03-01 | 2013-06-12 | 国家测绘地理信息局卫星测绘应用中心 | Star pattern recognition method based on directed loop |
| CN103335648A (en) * | 2013-06-27 | 2013-10-02 | 北京航天自动控制研究所 | Autonomous star map identification method |
| CN103868510A (en) * | 2014-03-27 | 2014-06-18 | 北京控制工程研究所 | Rapid autonomous all-sky map fixed star identification method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9423255B2 (en) * | 2014-06-19 | 2016-08-23 | The Boeing Company | System and method for mitigating an occurrence of a dry spot in a field of view of a star tracker |
-
2016
- 2016-06-22 CN CN201610457479.0A patent/CN106123891B/en not_active Expired - Fee Related
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN1896681A (en) * | 2005-07-15 | 2007-01-17 | 北京航空航天大学 | Fast tracting method of star sensor |
| CN101004349A (en) * | 2007-01-22 | 2007-07-25 | 北京航空航天大学 | Method for recognising not gauged star map |
| CN101441082A (en) * | 2008-12-22 | 2009-05-27 | 中国科学院长春光学精密机械与物理研究所 | Rapid triangle star map recognition method |
| CN102313549A (en) * | 2011-07-28 | 2012-01-11 | 北京航空航天大学 | Identification method for triangular star atlas based on characteristic of inertia ratio |
| CN102840861A (en) * | 2012-09-17 | 2012-12-26 | 常州工学院 | Navigational star screening method for star sensors |
| CN103148852A (en) * | 2013-03-01 | 2013-06-12 | 国家测绘地理信息局卫星测绘应用中心 | Star pattern recognition method based on directed loop |
| CN103335648A (en) * | 2013-06-27 | 2013-10-02 | 北京航天自动控制研究所 | Autonomous star map identification method |
| CN103868510A (en) * | 2014-03-27 | 2014-06-18 | 北京控制工程研究所 | Rapid autonomous all-sky map fixed star identification method |
Non-Patent Citations (3)
| Title |
|---|
| "New Robust Star Identification Algorithm";PABLO A. SERVIDIA，等;《IEEE TRANSACTIONS ON AEROSPACE AND ELECTRONIC SYSTEMS》;20060731;第42卷(第3期);1126-1131 * |
| "Star Recognition Algorithm for APS Star Tracker:Oriented Triangles";G. Lamy au Rousseau,等;《lEEE A&E SYSTEMS MAGAZINE》;20050228;27-31 * |
| "一种快速全天自主星图识别算法";李欣璐，等;《长春理工大学学报（自然科学版）》;20140630;第37卷(第3期);43-47 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106123891A (en) | 2016-11-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106123891B (en) | Faint star autonomous classification method based on perimeter feature | |
| CN102128623B (en) | Fast star map identification method of star sensor | |
| CN102986372A (en) | Picking object recognizing, classifying and space positioning device and picking object recognizing, classifying and space positioning method based on panoramic stereoscopic vision | |
| Momeni-k et al. | Height estimation from a single camera view | |
| CN111967337A (en) | Pipeline line change detection method based on deep learning and unmanned aerial vehicle images | |
| CN105928509B (en) | Triangle recognition methods based on maximum interior angle hash function | |
| Armagan et al. | Accurate Camera Registration in Urban Environments Using High-Level Feature Matching. | |
| Zou et al. | The comparison of two typical corner detection algorithms | |
| Cao et al. | Detection method based on image enhancement and an improved faster R-CNN for failed satellite components | |
| Chao-jian et al. | Image target identification of UAV based on SIFT | |
| Jian et al. | Target recognition and location based on binocular vision system of UUV | |
| CN116295363A (en) | Star point rapid extraction and high-precision positioning method | |
| CN113074719B (en) | Rapid and reliable star map identification method | |
| CN102004921A (en) | Target identification method based on image characteristic analysis | |
| WO2011054040A1 (en) | System and method for integration of spectral and 3-dimensional imaging data | |
| CN108109171A (en) | Unmanned plane aerophotograph revolves detection method, device, equipment and the storage medium of drift angle | |
| JP2007183908A (en) | Method for detecting object | |
| JPH1166321A (en) | Method for detecting work position | |
| CN114445468A (en) | Heterogeneous remote sensing image registration method and system | |
| Tomono | Building an object map for mobile robots using LRF scan matching and vision-based object recognition | |
| Sassi et al. | Skyline-based approach for natural scene identification | |
| Jung et al. | Matching aerial images to 3D building models based on context-based geometric hashing | |
| CN108898573A (en) | Infrared small target rapid extracting method based on multi-direction annular gradient method | |
| CN114972720B (en) | Autonomous image perception-based high-precision unmanned positioning method | |
| Almehio et al. | Level-line primitives for image registration with figures of merit |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20181116 |