CN105243075A - Improved search method for star sensor full celestial sphere maximum group recognition - Google Patents
Improved search method for star sensor full celestial sphere maximum group recognition Download PDFInfo
- Publication number
- CN105243075A CN105243075A CN201510484843.8A CN201510484843A CN105243075A CN 105243075 A CN105243075 A CN 105243075A CN 201510484843 A CN201510484843 A CN 201510484843A CN 105243075 A CN105243075 A CN 105243075A
- Authority
- CN
- China
- Prior art keywords
- star
- nautical
- observation
- binary tree
- candidate
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/903—Querying
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Navigation (AREA)
Abstract
The invention provides an improved search method for star sensor full celestial sphere maximum group recognition. The method comprises: firstly, performing sub-interval division on all navigation stars obtained by search through observation stars under the condition of meeting matching requirements according to navigation star sequence numbers; and secondly, establishing a balanced binary tree in each divided sub-interval to realize quick sorting and search according to a sequence of the navigation star sequence numbers and make statistics on a matching frequency of the observation star and each searched navigation star. A search algorithm has the advantage of the balanced binary tree while meeting the condition of occupying few spatial resources, and is high in search speed and short in binary search time; and at the same time, a large amount of memory data are not required for exchange, so that the time performance is further improved and the method has great advantage in application to embedded systems. The method also can be popularized and applied in similar data search occasions in other fields.
Description
Technical field
The present invention relates to a kind of method for random number retrieval in big data quantity, can be applicable to star sensor importance in star map recognition technical field and other technologies field similar applications.
Background technology
The current algorithm for very high degree of precision star sensor importance in star map recognition mainly or by observation star mates with navigation star angular distance.But because the quick sensitivity of very high degree of precision star is high, visual field is little, so that take in the star chart that obtains and observe star intensive, in the star catalogue of used identification, nautical star substantial amounts, adds difficulty to importance in star map recognition at every turn.Traditional triangle shape matching algorithm due to asterism overstocked, between observation star pair, angular distance difference is little, and the quantity that there will be star triangle is many, and the probability producing Redundancy Match is in the matching process large, thus causes that algorithm recognition success rate is low can not be suitable for.The newer maximum match group algorithm of one has fast and high accuracy at present.First this algorithm is found out and the nautical star pair observing star maximum to matching times from star catalogue, generates coupling matrix; Then confirm coupling matrix in mate between two between matching relationship, composition coupling confirm matrix; Last foundation coupling confirms that matrix computations goes out maximum coupling group, obtains recognition result.This algorithm CCD very high precision star quick on be verified, there is very high discrimination and probability of misrecognition is low.
Owing to all using embedded system in actual AEROSPACE APPLICATION, this requires very high to the time performance that program realizes.First step in above-mentioned maximum group of recognizer, find out and the nautical star pair observing star maximum to matching times, because the nautical star searched out is to many in reality, cause search data amount during statistical match number of times very large, excessive with direct sequential search time loss, real time performance requirement cannot be met.It is exactly in mass data, repeatedly search out a random number that this situation is summed up, and records the searched number of times arrived of this random number, and improves search efficiency as far as possible when space resources is limited, reduces search time.
Searching method sets up the Hash table of navigational star number one-to-one relationship in nautical star candidate queue and star catalogue the most fast, directly carry out index by this nautical star sequence number when new coupling obtains nautical star, the general index that final each sequence number adds up to number of times be exactly matching times.Here according to actual conditions, if the nautical star in star catalogue is about 30000, generally every observation star retrieve at every turn the candidate's nautical star obtained be just about 500 (in space, because star chart image quality is higher, what K Vector search angular distance and magnitude criterion all can be specified is more reasonable, thus retrieve candidate's nautical star of obtaining at every turn will less), and hypothesis carries out mating (for ensureing accuracy of identification requirement with 10 observation stars, choose bright star as far as possible, generally be no more than 12) identify, then every candidate's nautical star just about 4500 observing star 9 angular distances obtain common retrieval, amount of redundancy is added when opening up between candidate's nautical star queue empty to prevent spilling, then every observation star distributes real space size also only needs the space of 5000 units (because matching result when will preserve observation star and a nautical star coupling, then every nautical star also needs at least size to be that the integer array of 20 is to store matching result information, namely here the space of a unit is equivalent to the space shared by 20 integer numbers), if open for star catalogue navigational star number relation one to one, need to open up the space that size is 30000 units, great waste to space resources, the embedded system limited for space resources just can not accept more.And following step retrieves the maximum navigational star number of every observation star matching times, just need travel through in this way and retrieve for 30000 times, be also a kind of to waste of time.
Dichotomy ranking and searching, this is a kind of comparatively popular lookup algorithm, also be the searching method that this maximum group of algorithm is used when Computer Simulation, its algorithm realization complexity is low, and space resources takies few, and principle is sorted from small to large by nautical star sequence number by candidate's nautical star of every observation star, retrieve from the centre position of queue after sequence afterwards, relatively size, at every turn all binary search, the time complexity of dichotomy is O (log
2n), wherein n is queue length.Although the dichotomizing search time is much smaller than sequential search, but whenever will insert a new node in queue (the nautical star sequence number newly matched is not in candidate's nautical star queue before), node larger than this new node numerical value in queue must be moved one all afterwards.Here moving with regard to probably carrying out secondary data up to a hundred if the nautical star sequence number newly increased is smaller, also just relating to up to a hundred internal storage datas and exchanging.And embedded system is due to own characteristic, carry out internal storage data swap operation and be quite time consuming, this have impact on the time performance of algorithm realization to a great extent.
Summary of the invention
The technical matters that the present invention solves is: in every candidate's nautical star queue observing star a large amount of, whether retrieval exists newly mates the nautical star obtained at every turn, if do not exist, insert candidate queue, if exist, recording this nautical star and add up matching times, for overcoming the deficiencies in the prior art, providing the improvement searching method of a kind of star sensor whole day ball maximum group of identification, while the satisfied resource that takes up room is few, have the advantage of balanced binary tree concurrently, seek rate is fast, is the binary chop time; Exchange without the need to a large amount of internal storage data simultaneously, improve time performance further.
The technical scheme that the present invention solves is: the improvement searching method of a kind of star sensor whole day ball maximum group of identification, and as shown in Figure 1, step is as follows:
(1) sort extracting the observation star obtained in star chart from big to small by energy, and order descending for energy is used for carrying out match cognization with the nautical star in star catalogue as observation star sequence number;
(2) calculating observation star angle sine value between two, generates observation star angular distance table;
(3) in nautical star angular distance his-and-hers watches, to select in the observation star angular distance table generated with step (2) often pair observes star to all nautical stars pair mated, and all nautical stars meet following two conditions to also needing simultaneously:
A) star angular distance that navigates is less than or equal to angular distance matching threshold with the absolute value of observation star angular distance difference;
B) the nautical star magnitude of nautical star centering observes the absolute value of the magnitude difference of star be less than or equal to magnitude matching threshold with observation star centering; Final composition candidate nautical star pair;
(4) by little for candidate's nautical star centering magnitude of obtaining little of observation star centering sequence number corresponding at every turn, magnitude large with large corresponding of sequence number, thus obtain candidate's nautical star of every observation star at every turn;
(5) carry out the division of candidate sub-range, every observation star in step (4) is mated the candidate's nautical star sequence number obtained at every turn and obtains business divided by interval range, divide between sub-candidate regions by quotient;
(6) searching loop every observation star, equilibrium establishment binary tree in each candidate sub-range of every observation star, the wherein corresponding candidate's nautical star sequence number of each node of balanced binary tree, final entry obtains the matching times of every nautical star in this observation star and its all candidate's nautical stars queue.
Maximum navigational star number/10 in described interval range=star catalogue.
In described step (5), candidate sub-range is 10.
The candidate's nautical star first obtained first when building at the beginning of described step (6) balanced binary tree, as root node, then inserts the new node at every turn searched one by one in relevant position according to binary tree establishment principle; Search for while setting up binary tree, if certain searched navigational star number is with to there is node identical, then add 1 to this node matching number of times, and no longer this nautical star is used as new node and inserts binary tree; While setting up binary tree, need record the degree of depth relative to root node left and right subtree, if the two depth difference is more than or equal to 3, then binary tree will no longer balance, and can reduce search efficiency like this, must re-establish balanced binary tree, and this just needs mobile root node; If existing left subtree depth ratio right subtree is more than or equal to 3, then only node maximum for navigational star number in left subtree need be used as new root node, primitive root node is put into the right subtree of new root node, then newly-established binary tree is balanced binary tree; Otherwise be more than or equal to the situation of 3 on the left of the depth ratio of right side then, node minimum for navigational star number in right subtree need be used as new root node, primitive root node be put into the left subtree of new root node.
Described step (3) to be selected in the observation star angular distance table generated with step (2) often pair and is observed star to all nautical stars pair mated in nautical star angular distance his-and-hers watches with K vector method.
The present invention's advantage compared with prior art:
(1) compared to the mode at whole candidate queue sequential search, the present invention first sets up sub-range and divides, and then searches for by the mode of balanced binary tree in the sub-range of each decile again, greatly improves search speed further.
(2) search compared to the mode of index of reference, the present invention retrieves in conjunction with balanced binary tree dividing on the basis between candidate regions, to divide between candidate regions when opening up candidate queue space not increase amount of redundancy for principle; To extract the observation star number that obtains in star chart for circulation, in circulation, every observation star all observes star form angular distance pair one by one with other at every turn, successively carry out K Vector search, by searching for candidate's nautical star of obtaining by sequence number size demarcation interval at every turn, be placed on respectively in such as 10 sub-ranges.Because nautical star angular distance is substantially even to star catalogue distribution, like this can by basic 10 deciles of data volume altogether needing to retrieve to reduce the data volume of each retrieval.Namely carry out identification by above-mentioned hypothesis with 10 observation stars to calculate, candidate queue need open up the space of 10*500 unit-sized, equilibrium establishment binary tree needs overhead 2 to be doubly used for depositing binary tree index point to the integer array of candidate queue size, but its gross space expense is also much smaller than the space of 30000 unit-sized.
(3) search compared in dichotomy sequence, equilibrium establishment binary tree in each sub-range of star is observed at every in the present invention, its seek rate is identical with dichotomy, but need the internal storage data related to exchange seldom, balanced binary tree only just need carry out 3 internal storage datas and exchange when mobile binary tree root node, this affects the time performance of algorithm hardly, and when being much better than dichotomy, issuable up to a hundred internal storage datas exchange, therefore its time performance is much more superior than dichotomy.General space and time complexity, optimizedly solve current problem.
Accompanying drawing explanation
Fig. 1 is realization flow figure of the present invention;
Fig. 2 is that the embodiment of the present invention describes in detail;
Fig. 3 is the binary tree before moving in the present invention;
Fig. 4 is the binary tree after moving in the present invention.
Embodiment
As shown in Figure 2, the present invention is implemented as follows:
(1) sort extracting the observation star obtained in star chart from big to small by energy, and order descending for energy is used for carrying out match cognization with the nautical star in star catalogue as observation star sequence number;
(2) calculating observation star angle sine value between two, generates observation star angular distance table;
(3) in nautical star angular distance his-and-hers watches, to select in the observation star angular distance table generated with step (2) often pair with K vector method observe star to all nautical stars pair mated, and all nautical stars meet following two conditions to also needing simultaneously:
A) star angular distance that navigates is less than or equal to angular distance matching threshold with the absolute value of observation star angular distance difference;
B) the nautical star magnitude of nautical star centering observes the absolute value of the magnitude difference of star be less than or equal to magnitude matching threshold with observation star centering.
(4) by little for candidate's nautical star centering magnitude of obtaining little of observation star centering sequence number corresponding at every turn, magnitude large with large corresponding of sequence number, thus obtain candidate's nautical star of every observation star at every turn;
(5) carry out sub-range division, every observation star in step (4) is mated the candidate's nautical star sequence number obtained at every turn and obtains business divided by interval range; Candidate's nautical star being put into this observation star candidate nautical star 10 sub-ranges (sub-range sequence number is 0-9) is that (such as star catalogue has 50000 nautical stars, then interval range=50000/10=5000 for of sequence number with quotient; If the 1st the nautical star sequence number observing star once search out coupling is 9000, the business of 9000/5000 is 1, then this nautical star is put into the candidate sub-range that the 1st observation star sequence number is 1).
(6) searching loop every observation star, equilibrium establishment binary tree in each candidate sub-range of every observation star, the wherein corresponding candidate's nautical star sequence number of each node of balanced binary tree, final entry obtains the matching times of every nautical star in this observation star and its all candidate's nautical stars queue.
Equilibrium establishment binary tree needs storage allocation, the nautical star sequence number corresponding to the child node of left and right of every nautical star in record candidate queue, Here it is additionally need apply for 2 times to the integer array of candidate's nautical star queue.The candidate's nautical star first obtained first when building at the beginning of balanced binary tree, as root node, then inserts the new node at every turn searched one by one in relevant position according to binary tree establishment principle.Can search for while setting up binary tree, if certain searched navigational star number is with to there is node identical, then add 1 to this node matching number of times, and no longer this nautical star is used as new node and insert binary tree.During each insertion new node, be all to the left and right child node assignment of each child node, do not relate to internal storage data and exchange.
While setting up binary tree, need record the degree of depth relative to root node left and right subtree, if the two depth difference is greater than 3, then binary tree will no longer balance, and can reduce search efficiency like this, must re-establish balanced binary tree, and this just needs mobile root node.The method of mobile root node has a variety of, and wherein most timesaving one as shown in Figure 3, Figure 4.
According to binary tree principle, if existing left subtree depth ratio right subtree large 3, then node that only need be maximum by the navigational star number in left subtree is used as new root node, and primitive root node is put into the right subtree of new root node, then newly-established binary tree is balanced binary tree.Otherwise be more than or equal to the situation of 3 on the left of the depth ratio of right side then, node minimum for navigational star number in right subtree need be used as new root node, primitive root node be put into the left subtree of new root node.
Moving method is, according to binary tree establishment principle, need by element maximum in left subtree, and namely right node is that empty child node is used as new root node, if L3 in Fig. 3 is exactly new root node.The left subtree of L3 is become the right subtree of L1, by the right node that primitive root node root is used as L3, former right subtree is used as the right subtree of root node, the left subtree of root node is set to sky.Using the left sibling of L1 as new node L3, other node is unchanged.The balanced binary tree re-established so still meets binary tree principle as shown in Figure 4.In moving process, only need travel through left subtree and find wherein right node to be empty child node, without the need to traveling through right subtree, this left subtree being 250 for the degree of depth also only needs 6 step traversals at the most; Step afterwards has also only done 3 internal storage data assignment when again putting the right node of the left and right node of L3 and L1, and access memory number of times is few, and moving root node as seen generally, to re-establish the process of balanced binary tree very low to time loss.In practical application, left and right subtree can be allowed to have certain degree of depth K vector method K vector method poor, this does not affect the search speed of binary tree substantially, such as depth difference can be relaxed is 10, can ensure that mobile rear left right subtree depth difference is 8 like this, so, the probability of mobile binary tree root node will step-down, can improve performance search time further.
Be actually passed through embedded system test, speed of the present invention is far away higher than sequential search and dichotomizing search.
The content be not described in detail in instructions of the present invention belongs to the known technology of those skilled in the art.
Claims (5)
1. an improvement searching method for star sensor whole day ball maximum group of identification, is characterized in that comprising step as follows:
(1) sort extracting the observation star obtained in star chart from big to small by energy, and order descending for energy is used for carrying out match cognization with the nautical star in star catalogue as observation star sequence number;
(2) calculating observation star angle sine value between two, generates observation star angular distance table;
(3) in nautical star angular distance his-and-hers watches, to select in the observation star angular distance table generated with step (2) often pair observes star to all nautical stars pair mated, and all nautical stars meet following two conditions to also needing simultaneously:
A) star angular distance that navigates is less than or equal to angular distance matching threshold with the absolute value of observation star angular distance difference;
B) the nautical star magnitude of nautical star centering observes the absolute value of the magnitude difference of star be less than or equal to magnitude matching threshold with observation star centering; Final composition candidate nautical star pair;
(4) by little for candidate's nautical star centering magnitude of obtaining little of observation star centering sequence number corresponding at every turn, magnitude large with large corresponding of sequence number, thus obtain candidate's nautical star of every observation star at every turn;
(5) carry out the division of candidate sub-range, every observation star in step (4) is mated the candidate's nautical star sequence number obtained at every turn and obtains business divided by interval range, divide between sub-candidate regions by quotient;
(6) searching loop every observation star, equilibrium establishment binary tree in each candidate sub-range of every observation star, the wherein corresponding candidate's nautical star sequence number of each node of balanced binary tree, final entry obtains the matching times of every nautical star in this observation star and its all candidate's nautical stars queue.
2. the improvement searching method of star sensor whole day ball according to claim 1 maximum group of identification, is characterized in that: maximum navigational star number/sub-range number in interval range=star catalogue in described step (5).
3. the improvement searching method of star sensor whole day ball according to claim 1 maximum group of identification, is characterized in that: in described step (5), candidate sub-range number is 10.
4. the improvement searching method of star sensor whole day ball according to claim 1 maximum group of identification, it is characterized in that: the candidate's nautical star first obtained first when building at the beginning of described step (6) balanced binary tree, as root node, then inserts the new node at every turn searched one by one in relevant position according to binary tree establishment principle; Search for while setting up binary tree, if certain searched navigational star number is with to there is node identical, then add 1 to this node matching number of times, and no longer this nautical star is used as new node and inserts binary tree; While setting up binary tree, need record the degree of depth relative to root node left and right subtree, if the two depth difference is more than or equal to 3, then binary tree will no longer balance, and can reduce search efficiency like this, must re-establish balanced binary tree, and this just needs mobile root node; If existing left subtree depth ratio right subtree is more than or equal to 3, then only node maximum for navigational star number in left subtree need be used as new root node, primitive root node is put into the right subtree of new root node, then newly-established binary tree is balanced binary tree; Otherwise be more than or equal to the situation of 3 on the left of the depth ratio of right side then, node minimum for navigational star number in right subtree need be used as new root node, primitive root node be put into the left subtree of new root node.
5. the improvement searching method of star sensor whole day ball according to claim 1 maximum group of identification, is characterized in that: described step (3) K vector method to be selected in the observation star angular distance table generated with step (2) often pair and observed star to all nautical stars pair mated in nautical star angular distance his-and-hers watches.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510484843.8A CN105243075B (en) | 2015-08-07 | 2015-08-07 | A kind of star sensor whole day ball greatly organizes the improvement searching method of identification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510484843.8A CN105243075B (en) | 2015-08-07 | 2015-08-07 | A kind of star sensor whole day ball greatly organizes the improvement searching method of identification |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105243075A true CN105243075A (en) | 2016-01-13 |
| CN105243075B CN105243075B (en) | 2018-08-31 |
Family
ID=55040725
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510484843.8A Active CN105243075B (en) | 2015-08-07 | 2015-08-07 | A kind of star sensor whole day ball greatly organizes the improvement searching method of identification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105243075B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106348119A (en) * | 2016-09-20 | 2017-01-25 | 广州特种机电设备检测研究院 | Isolated elevator running safety monitoring system and method based on Internet of Things |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO1997011882A2 (en) * | 1995-09-28 | 1997-04-03 | Lockheed Martin Corporation | Techniques for optimizing an autonomous star tracker |
| CN101034408A (en) * | 2007-04-16 | 2007-09-12 | 北京航空航天大学 | Star map matching recognizing method based on ant colony algorithm |
| CN101363733A (en) * | 2008-09-17 | 2009-02-11 | 北京航空航天大学 | Ultra-high accuracy star sensor |
| CN103954280A (en) * | 2014-04-08 | 2014-07-30 | 北京控制工程研究所 | Rapid, high-robustness and autonomous fixed star identification method |
-
2015
- 2015-08-07 CN CN201510484843.8A patent/CN105243075B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO1997011882A2 (en) * | 1995-09-28 | 1997-04-03 | Lockheed Martin Corporation | Techniques for optimizing an autonomous star tracker |
| CN101034408A (en) * | 2007-04-16 | 2007-09-12 | 北京航空航天大学 | Star map matching recognizing method based on ant colony algorithm |
| CN101363733A (en) * | 2008-09-17 | 2009-02-11 | 北京航空航天大学 | Ultra-high accuracy star sensor |
| CN103954280A (en) * | 2014-04-08 | 2014-07-30 | 北京控制工程研究所 | Rapid, high-robustness and autonomous fixed star identification method |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106348119A (en) * | 2016-09-20 | 2017-01-25 | 广州特种机电设备检测研究院 | Isolated elevator running safety monitoring system and method based on Internet of Things |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105243075B (en) | 2018-08-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101446962B (en) | Data conversion method, device thereof and data processing system | |
| CN102880854B (en) | Distributed processing and Hash mapping-based outdoor massive object identification method and system | |
| CN106033416A (en) | A string processing method and device | |
| CN103744934A (en) | Distributed index method based on LSH (Locality Sensitive Hashing) | |
| CN103049496A (en) | Method, apparatus and device for dividing multiple users into user groups | |
| CN103617217A (en) | Hierarchical index based image retrieval method and system | |
| CN106095920B (en) | Distributed index method towards extensive High dimensional space data | |
| CN106599091B (en) | RDF graph structure storage and index method based on key value storage | |
| CN106326475A (en) | High-efficiency static hash table implement method and system | |
| CN106649828A (en) | Data query method and system | |
| CN103049554A (en) | Parallel indexing technology for vector QR trees | |
| CN103345496A (en) | Multimedia information searching method and system | |
| CN106528790B (en) | The choosing method and device of supporting point in metric space | |
| CN104636349A (en) | Method and equipment for compression and searching of index data | |
| CN103678550A (en) | Mass data real-time query method based on dynamic index structure | |
| CN108549696B (en) | Time series data similarity query method based on memory calculation | |
| CN110069500A (en) | A kind of non-relational database dynamic hybrid index method | |
| CN111125396B (en) | Image retrieval method of single-model multi-branch structure | |
| CN111400301B (en) | Data query method, device and equipment | |
| CN110097581B (en) | Method for constructing K-D tree based on point cloud registration ICP algorithm | |
| CN105447064B (en) | Electronic map data making and using method and device | |
| Abbasifard et al. | Efficient indexing for past and current position of moving objects on road networks | |
| CN109711439A (en) | A kind of extensive tourist's representation data clustering method in density peak accelerating neighbor seaching using Group algorithm | |
| CN104424189A (en) | Positioning resolving method and positioning resolving system based on cloud platform | |
| CN105243075A (en) | Improved search method for star sensor full celestial sphere maximum group recognition |
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 |