CN103411606B - A kind of star sensor star point cluster extraction method - Google Patents
A kind of star sensor star point cluster extraction method Download PDFInfo
- Publication number
- CN103411606B CN103411606B CN201310316798.6A CN201310316798A CN103411606B CN 103411606 B CN103411606 B CN 103411606B CN 201310316798 A CN201310316798 A CN 201310316798A CN 103411606 B CN103411606 B CN 103411606B
- Authority
- CN
- China
- Prior art keywords
- pixel
- asterism
- queue
- star
- criterion
- 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.)
- Active
Links
Landscapes
- Image Analysis (AREA)
Abstract
A kind of star sensor star point cluster extraction method, star sensor clap after a width star chart, read image element information in the picture signal that star sensor receives and in the picture signal received by ranks sequential storage star sensor gray scale higher than the pixel of image background; When asterism extracts, read the effective image element information of star chart successively, according to four clustering criterions and row-coordinate monotonically increasing feature, judge the attaching relation of pixel and asterism, cooperation dynamically removes technology with pixel index technology, noise and internal memory refills technology, is deposited by pixel by affiliated asterism different grouping; Finally pseudo-star judgement, centroid calculation carried out to the asterism pixel queue after cluster, close on asterism rejecting and energy ordering, finally complete extraction and the output of asterism queue.The invention enables star sensor star point to extract can not need to be equipped with jumbo full star chart storer, does not need to carry out full star chart pixel consuming time and travels through one by one, and ensures robustness in the situations such as star chart noise spot is more.
Description
Technical field
The present invention relates to a kind of star sensor star point cluster extraction method, be applicable to spacecraft Star Sensor field.
Background technology
Star sensor is that a kind of imaging type spatial attitude measures sensor, by extracting the asterism of clapped star chart, identify and calculate, export the three-axis attitude information of star sensor, because it has the feature such as high precision, high reliability, be widely used in all kinds of spacecraft such as satellite, airship.Asterism extracts the star sensor attitude initial acquisition time that directly affects, exports the important indicators such as attitude accuracy, is the important foundation work that star sensor obtains attitude information.
Existing star sensor star point extraction algorithm is as the celestial body segmented positioning technique study of star map image " in the star sensor " (author: Wei Xinguo, Zhang Guangjun etc.; Source: BJ University of Aeronautics & Astronautics's journal) etc. described in document, mostly be the full star chart pixel traversal formula asterism extraction algorithm based on connected domain, this algorithm has following open defect
(1) prior art needs star sensor to be equipped with jumbo full star chart storer, and as being 1024 × 1024 pixels by the face battle array of sensitive chip, gradation data length is 12 calculating, stores the capacity that the full star chart of a width needs 2M byte.
(2) during prior art extraction asterism, many need carry out full star chart pixel traversal, namely read and write pixel each in star chart, judge, be unfavorable for the simplification of star sensor hardware and the quick obtaining of asterism information.
(3) if do not carried out full star chart pixel traversal when prior art extracts asterism, then, in the situations such as star chart noise spot is more, algorithm committed memory space can significantly increase, and even causes memory overflow.
Therefore the asterism cluster extraction method that research is a kind of fast, effective, star sensor hardware requirement is lower is needed now.
Summary of the invention
Technology of the present invention is dealt with problems and is: overcome the deficiencies in the prior art, propose a kind of pixel index technology, noise of comprising and dynamically remove the star sensor star point cluster extraction method that technology and internal memory refill technology, star sensor star point is extracted can not be needed to be equipped with jumbo full star chart storer, do not need to carry out full star chart pixel consuming time to travel through one by one, and ensure robustness in the situation such as star chart noise spot is more.
Technical solution of the present invention is:
A kind of star sensor star point cluster extraction method, its step is as follows:
(1) read image element information in the picture signal that receives of star sensor and in the picture signal received by ranks sequential storage star sensor gray scale higher than the pixel of image background, if gray scale is more than or equal to 1 higher than the pixel number of image background, then create initial asterism and first gray scale is entered step (2) higher than the pixel of image background stored in the queue of asterism pixel, if pixel number is less than 1, then asterism extracts and terminates;
(2) read next gray scale and carry out affiliated asterism judgement higher than the pixel of image background, judge whether the row address of last pixel of built vertical asterism pixel queue and the difference of current pixel row address is greater than 1 and queue length is less than threshold value, if set up, this asterism pixel queue is judged as a noise spot and empties this asterism queue and enter step (4), if be false, enters step (3);
(3) pixel is carried out to the judgement of affiliated asterism, first carry out the judgement of criterion 1 and 2, if meet criterion 1 or 2, then the judgement carrying out criterion 3 enters step (4); If 1 and 2 do not meet, directly enter step (4), clustering criterion is as follows:
Criterion 1: pixel is identical with last pixel row address in the pixel queue of asterism, and column address is adjacent, then pixel belongs to this asterism; If any asterism that this pixel is traversed before not belonging to simultaneously, then pixel is placed in this asterism pixel queue end, in pixel queue, only preserves pixel index sequence number;
Criterion 2: pixel is identical or adjacent with last pixel row address in the pixel queue of asterism and there is adjacent with pixel row address, that column address is identical pixel in the pixel queue of asterism, then pixel belongs to this asterism; If any asterism that this pixel is traversed before not belonging to simultaneously, then pixel is placed in this asterism pixel queue end, in pixel queue, only preserves pixel index sequence number;
Criterion 3: if pixel is through judging to belong to some asterisms, judge that this pixel also belongs to another asterism and has been placed in another asterism pixel queue end simultaneously, then another asterism pixel queue is spliced and carry out asterism merging in the queue end of some asterisms, and empty another asterism pixel queue; Otherwise pixel is added and meets in the asterism pixel queue of criterion 1 or 2;
(4) judge whether that other asterisms judge in addition, if there is repetition step (2), if not, carry out step (5);
(5) judgement of criterion 4 is carried out: if pixel is through judging not belong to any asterism created, then create new asterism, pixel adds the queue of asterism pixel as this asterism first pixel;
(6) judge whether that in addition gray scale does not carry out affiliated asterism judgement, if the step of entering (2), if do not enter step (7) higher than the pixel of image background;
(7) pseudo-star judgement, centroid calculation carried out to the asterism pixel queue after cluster, close on asterism rejecting and energy ordering, finally complete extraction and the output of asterism queue.
Preferentially take in step (2) when newly setting up asterism queue in criterion 4 in described step (5) and be identified as noise spot and the asterism pixel queue of being cleared.
The present invention's beneficial effect is compared with prior art:
(1) The present invention reduces star chart storage capacity requirement, optimize star sensor hardware composition (decreasing reservoir), prior art is many need store full star chart, be 1024 × 1024 pixels by the face battle array of sensitive chip, gradation data length is 12 calculating, store the full star chart of a width and need the capacity of 2M byte (every 16 bit spaces store a pixel gradation data), then star sensor hardware is necessary for and stores star chart interpolation mass storage; The present invention only need store the effective image element information of star chart, and calculate by aforementioned sensitive chip parameter, can ensure that memory capacity is not more than 40K byte, the extensive logical device internal storage region of general star sensor just can meet storage demand.
(2) The present invention reduces asterism extraction time, during prior art extraction asterism, many need carry out full star chart pixel traversal, namely pixel each in star chart is read and write, judged, the present invention only need read and write the effective image element information stored, judge, in the identical situations such as sensitive chip, processor and object starry sky, the present invention extract that asterism required time is only full star chart ergodic algorithm about 1/50, and it is correct to extract result.
(3) this invention ensures that the robustness of extraction algorithm in the situations such as star chart noise spot is more, prior art travels through if do not carried out full star chart pixel when extracting asterism, then in the situation such as star chart noise spot is more, algorithm committed memory space can significantly increase, even cause memory overflow, the present invention can automatically remove noise in the situations such as star chart noise spot is more, correctly complete asterism and extract, and algorithm committed memory space size is controlled.
Accompanying drawing explanation
Fig. 1 is process flow diagram of the present invention;
Fig. 2 is clustering criterion 1 asterism schematic diagram;
Fig. 3 is clustering criterion 2 asterism schematic diagram;
Fig. 4 is clustering criterion 3 asterism schematic diagram.
Embodiment
Below in conjunction with accompanying drawing, the specific embodiment of the present invention is further described in detail.
As shown in Figure 1, a kind of star sensor star point cluster extraction method, its concrete steps are as follows:
(1) read image element information in the picture signal that receives of star sensor and in the picture signal received by ranks sequential storage star sensor gray scale higher than the pixel of image background, if gray scale is more than or equal to 1 higher than the pixel number (being called effective pixel) of image background, then create initial asterism and first gray scale is entered step (2) higher than the pixel of image background stored in the queue of asterism pixel, if pixel number is less than 1, then asterism extracts and terminates, and pixel storage format is as follows:
Table 1 is pixel storage format effectively
(2) next (reading according to the order stored in form is read, if only have a pixel gray scale higher than image background, then judging in step (1) after asterism belonging to first pixel that asterism extracts just to finish) gray scale carries out affiliated asterism judgement higher than the pixel of image background, judge whether that the row address of last pixel of built vertical asterism pixel queue and the difference of current pixel row address are greater than 1 and queue length is less than threshold value that (queue length threshold is depending on actual conditions, that extracts according to actual asterism needs setting threshold value), if set up, judge that this asterism pixel queue empties this asterism queue as a noise spot, follow-up newly-established asterism preferentially takies this pixel queuing memory space emptied, if set up, this asterism pixel queue is judged as a noise spot and empties this asterism queue and enter step (4), if be false, enter step (3).
(3) pixel is carried out to the judgement of affiliated asterism, first carry out the judgement of criterion 1 and 2, if meet criterion 1 or 2, then the judgement carrying out criterion 3 enters step (4); If 1 and 2 do not meet, directly enter step (4), clustering criterion is as follows:
Criterion 1: pixel is identical with last pixel row address in the pixel queue of asterism, and column address is adjacent, then pixel belongs to this asterism; If any asterism that this pixel is traversed before not belonging to simultaneously, then pixel is placed in this asterism pixel queue end; As shown in Figure 2, pixel P is identical with s (7) row address in asterism queue, and column address is adjacent, then pixel P belongs to this asterism, and pixel P is placed in after this asterism pixel queue s (7).
Criterion 2: pixel is identical or adjacent with last pixel row address in the pixel queue of asterism and there is adjacent with pixel row address, that column address is identical pixel in the pixel queue of asterism, then pixel belongs to this asterism; If any asterism that this pixel is traversed before not belonging to simultaneously, then pixel is placed in this asterism pixel queue end; As shown in Figure 3, the row address of the pixel s (7) in pixel P and asterism queue is adjacent, and identical with the column address of s (7), so pixel P belongs to this asterism, pixel P is placed in after this asterism pixel queue.
Criterion 3: if pixel is through judging to belong to some asterisms, judge that this pixel also belongs to another asterism and has been placed in another asterism pixel queue end simultaneously, then another asterism pixel queue is spliced and carry out asterism merging in the queue end of some asterisms, and empty another asterism pixel queue; Otherwise pixel is added and meets in the asterism pixel queue of criterion 1 or 2; As shown in Figure 4, if pixel p is through judging to belong to asterism s2, but before through judging belong to s1 and be placed in asterism s1 pixel queue end, is positioned at pixel queue s (4) place, then the queue end in asterism s2 is spliced in the queue of asterism s1 pixel, and empty the queue of s1 pixel.
(4) judge whether that other asterisms judge in addition, if there is repetition step (2), if not, carry out step (5).
(5) judgement of criterion 4 is carried out: if pixel is through judging not belong to any asterism created, then create new asterism, pixel adds the queue of asterism pixel as this asterism first pixel.
(6) judge whether that in addition gray scale does not carry out affiliated asterism judgement, if the step of entering (2), if do not enter step (7) higher than the pixel of image background.
(7) pseudo-star judgement is carried out (according to asterism pixel quantity to the asterism pixel queue after cluster, pixel region distribution rationality etc. judges whether the asterism of cluster is a fixed star picture point, not that namely fixed star picture point thinks pseudo-star), centroid calculation (adopting method of energy integral etc. to calculate asterism energy barycenter), closing on asterism rejects (by " two asterisms that energy barycenter is very close, reject energy lower asterism " etc. principle carry out) and energy ordering (asterism sequence is sorted from big to small by asterism energy, subsequent module is facilitated to use), finally complete extraction and the output of asterism queue.
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 (2)
1. a star sensor star point cluster extraction method, is characterized in that step is as follows:
(1) read image element information in the picture signal that receives of star sensor and in the picture signal received by ranks sequential storage star sensor gray scale higher than the pixel of image background, if gray scale is more than or equal to 1 higher than the pixel number of image background, then create initial asterism and first gray scale is entered step (2) higher than the pixel of image background stored in the queue of asterism pixel, if pixel number is less than 1, then asterism extracts and terminates;
(2) read next gray scale and carry out affiliated asterism judgement higher than the pixel of image background, judge whether the row address of last pixel of built vertical asterism pixel queue and the difference of current pixel row address is greater than 1 and queue length is less than threshold value, if set up, this asterism pixel queue is judged as a noise spot and empties this asterism queue and enter step (4), if be false, enters step (3);
(3) next gray scale is carried out to the judgement of affiliated asterism higher than the pixel of image background, first carry out the judgement of criterion 1 and 2, if meet criterion 1 or 2, then the judgement carrying out criterion 3 enters step (4); If 1 and 2 do not meet, directly enter step (4), clustering criterion is as follows:
Criterion 1: pixel is identical with last pixel row address in the pixel queue of asterism, and column address is adjacent, then pixel belongs to this asterism; If any asterism that this pixel is traversed before not belonging to simultaneously, then pixel is placed in this asterism pixel queue end, in pixel queue, only preserves pixel index sequence number;
Criterion 2: pixel is identical or adjacent with last pixel row address in the pixel queue of asterism and there is adjacent with pixel row address, that column address is identical pixel in the pixel queue of asterism, then pixel belongs to this asterism; If any asterism that this pixel is traversed before not belonging to simultaneously, then pixel is placed in this asterism pixel queue end, in pixel queue, only preserves pixel index sequence number;
Criterion 3: if pixel is through judging to belong to some asterisms, judge that this pixel also belongs to another asterism and has been placed in another asterism pixel queue end simultaneously, then another asterism pixel queue is spliced and carry out asterism merging in the queue end of some asterisms, and empty another asterism pixel queue; Otherwise pixel is added and meets in the asterism pixel queue of criterion 1 or 2;
(4) judge whether that other asterisms judge in addition, if there is repetition step (2), if not, carry out step (5);
(5) judgement of criterion 4 is carried out: if pixel is through judging not belong to any asterism created, then create new asterism, pixel adds the queue of asterism pixel as this asterism first pixel;
(6) judge whether that in addition gray scale does not carry out affiliated asterism judgement, if the step of entering (2), if do not enter step (7) higher than the pixel of image background;
(7) pseudo-star judgement, centroid calculation carried out to the asterism pixel queue after cluster, close on asterism rejecting and energy ordering, finally complete extraction and the output of asterism queue.
2. a kind of star sensor star point cluster extraction method according to claim 1, is characterized in that: preferentially take in step (2) when newly setting up asterism queue in criterion 4 in described step (5) and be identified as noise spot and the asterism pixel queue of being cleared.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310316798.6A CN103411606B (en) | 2013-07-25 | 2013-07-25 | A kind of star sensor star point cluster extraction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310316798.6A CN103411606B (en) | 2013-07-25 | 2013-07-25 | A kind of star sensor star point cluster extraction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103411606A CN103411606A (en) | 2013-11-27 |
| CN103411606B true CN103411606B (en) | 2015-10-21 |
Family
ID=49604634
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310316798.6A Active CN103411606B (en) | 2013-07-25 | 2013-07-25 | A kind of star sensor star point cluster extraction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103411606B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105004353B (en) * | 2015-06-17 | 2018-05-22 | 北京控制工程研究所 | A kind of star sensor dynamic star chart emulation mode |
| CN105424033B (en) * | 2015-10-30 | 2018-04-27 | 北京控制工程研究所 | A kind of method for improving Stellar orientation instrument dynamic instrumentation sensitivity |
| CN106441281A (en) * | 2016-09-09 | 2017-02-22 | 北京控制工程研究所 | Small high-precision star sensor with long service life |
| CN113674134B (en) * | 2021-08-03 | 2023-06-30 | 北京控制工程研究所 | Real-time reliable star point identification and storage method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101226113A (en) * | 2008-01-30 | 2008-07-23 | 北京航空航天大学 | Method and apparatus for testing star sensor function based on electric injection star map |
| CN101236651A (en) * | 2008-01-31 | 2008-08-06 | 北京控制工程研究所 | Self-adapting image background estimation method |
| CN102081738A (en) * | 2011-01-06 | 2011-06-01 | 西北工业大学 | Method for positioning mass center of spatial object star image |
| CN102096829A (en) * | 2011-01-06 | 2011-06-15 | 西北工业大学 | Iterative optimization distance categorization-based space weak and small target detection method |
| CN102903105A (en) * | 2012-09-17 | 2013-01-30 | 常州工学院 | Method for extracting star images from star chart |
| CN103020906A (en) * | 2012-12-03 | 2013-04-03 | 清华大学 | A pre-processing method of measured star image by star sensor during the day |
-
2013
- 2013-07-25 CN CN201310316798.6A patent/CN103411606B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101226113A (en) * | 2008-01-30 | 2008-07-23 | 北京航空航天大学 | Method and apparatus for testing star sensor function based on electric injection star map |
| CN101236651A (en) * | 2008-01-31 | 2008-08-06 | 北京控制工程研究所 | Self-adapting image background estimation method |
| CN102081738A (en) * | 2011-01-06 | 2011-06-01 | 西北工业大学 | Method for positioning mass center of spatial object star image |
| CN102096829A (en) * | 2011-01-06 | 2011-06-15 | 西北工业大学 | Iterative optimization distance categorization-based space weak and small target detection method |
| CN102903105A (en) * | 2012-09-17 | 2013-01-30 | 常州工学院 | Method for extracting star images from star chart |
| CN103020906A (en) * | 2012-12-03 | 2013-04-03 | 清华大学 | A pre-processing method of measured star image by star sensor during the day |
Non-Patent Citations (3)
| Title |
|---|
| APS图像传感器及其在星敏感器中的应用;钟红军等;《光学技术》;20090331;第35卷(第2期);204-208 * |
| 卫星转动条件下APS星敏感器星像目标中心提取精度分析;李晓等;《空间控制技术与应用》;20090831;第35卷(第4期);11-16 * |
| 李玉峰等.一种快速的星敏感器星点提取方法.《仪表,自动化及先进集成技术大会论文集(二)》.2007,第28卷 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103411606A (en) | 2013-11-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103411606B (en) | A kind of star sensor star point cluster extraction method | |
| CN108716916B (en) | Distributed parallel star point centroid extraction method based on superblock and FPGA implementation device | |
| Toft et al. | Long-term 3d localization and pose from semantic labellings | |
| WO2018233200A1 (en) | Rapid gate image processing system and method for star sensor | |
| CN110991311A (en) | Target detection method based on dense connection deep network | |
| CN103295034A (en) | Embedded type system and method for checking books being placed on shelf disorderly based on DSP | |
| CN113762455B (en) | Detection model training method, single word detection method, device, equipment and medium | |
| CN103678657B (en) | Method for storing and reading altitude data of terrain | |
| CN101572017A (en) | Method for extracting cruise section of flight path from air craft three dimensional position sequence | |
| CN110555432B (en) | Method, device, equipment and medium for processing interest points | |
| CN103021016B (en) | Magnanimity three-dimensional geographic information data processing method in a kind of render process | |
| CN110197113A (en) | A kind of method for detecting human face of high-precision anchor point matching strategy | |
| CN109708627A (en) | A kind of moving platform down space dynamic point target rapid detection method | |
| CN102314798B (en) | Method and device for automatically simplifying shape points of polygonal surface elements of electronic map | |
| CN104567865A (en) | Attitude capture method of star sensor under space particle interference condition | |
| Cao et al. | Small object detection algorithm for railway scene | |
| CN117083621A (en) | Detector training method, device and storage medium | |
| CN102903105A (en) | Method for extracting star images from star chart | |
| Peng et al. | Real-time traffic sign text detection based on deep learning | |
| CN110879063A (en) | Rapid star map identification method based on triangle matching | |
| CN117523120A (en) | Method for setting boundary points of natural resource right-determining registration unit based on three-dimensional scene | |
| CN110046632A (en) | Model training method and device | |
| CN113033578B (en) | Image calibration method, system, terminal and medium based on multi-scale feature matching | |
| CN111753625B (en) | Pedestrian detection method, device, equipment and medium | |
| CN106874837A (en) | A kind of vehicle checking method based on Computer Vision |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |