CN109978797B - Active pixel sensor star image denoising method based on dark channel noise template - Google Patents
Active pixel sensor star image denoising method based on dark channel noise template Download PDFInfo
- Publication number
- CN109978797B CN109978797B CN201910274241.8A CN201910274241A CN109978797B CN 109978797 B CN109978797 B CN 109978797B CN 201910274241 A CN201910274241 A CN 201910274241A CN 109978797 B CN109978797 B CN 109978797B
- Authority
- CN
- China
- Prior art keywords
- star
- noise
- star map
- template
- map
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20224—Image subtraction
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
- Navigation (AREA)
Abstract
The invention discloses an active pixel sensor star image denoising method based on a dark channel noise template, which comprises the following steps: calculating the frame number of the star map, and extracting a dark channel noise template of the star map of the active pixel sensor according to the minimum value of the pixels of the sequence star map images; carrying out difference on the sequence star map and a dark channel noise template to separate noise from an original star map, and realizing the image denoising of continuous multi-frame active pixel sensor star maps; and filtering out the residual single-point and multi-point noises of the star map image after the noise removal based on the dark channel template. The method provided by the invention can further improve the attitude determination precision of the star sensor by denoising the star map of the active pixel sensor, thereby being beneficial to improving the global uncontrolled mapping capability of the surveying and mapping satellite.
Description
Technical Field
The invention belongs to an active pixel sensor star map denoising processing technology, and particularly relates to an active pixel sensor star map denoising method based on a dark channel noise template.
Background
The star sensor is the attitude sensor with the highest precision at present, has the advantages of high precision, good reliability, strong capturing and tracking capability and the like, and is one of the key factors for determining whether the satellite can stably run on the orbit [ Luo L, Xu L, Zhang H.Improduced center extraction algorithm for autonomus star sensor [ J ]. IET Imageprocessing,2015,9(10):901 and 907 ]. In the attitude control system of the star sensor, the star sensor shoots a star map through a camera, and the orientation of the optical axis of the star sensor camera in a celestial coordinate system is determined through a series of work such as extraction of the centroid of a star, star map identification, attitude calculation and the like, so that the attitude [ Ho K.A surfey of algorithm for star identification with low-cost star tracker [ J ]. Acta Astronacartia, 2012,73: 156-; mo F, Xie J F, He Z N, et al.A pre-processing method of raw attribute for ZY-3satellite [ J ] Science of harvesting and Mapping,2016,41(1):127 and 132 ]. The star map denoising and the centroid positioning are key technologies for star sensitive attitude determination. The star map denoising is the premise of centroid extraction, and centroid positioning directly determines the attitude determination precision of the star sensor after the fact. Therefore, how to effectively remove noise and accurately extract the centroid is one of the research hot problems of the space remote sensing satellite in recent years [ Liuyu, Dadongkai, Marilibalance, and the like ] a star sensor calibration method based on an attitude correlation frame [ J ]. optical science and newspaper, 2017,37(11): 1128003; zhang W N, Quan W, Guo L.Blurredstar image processing for star Sensors under dynamic conditions [ J ]. Sensors,2012,12(5):6712-6726 ].
The existing star map denoising methods can be roughly divided into filtering denoising and threshold segmentation methods. Filtering and denoising [ ZHou M Y, Shi Y, Yang J G. denoising star map data video spatial representation and denoising [ J ]. Optik,2015,126(11/12):1133 and 1137); wangmi, Zhaojinyu, Chentao, etc. an extreme value median filtering star image denoising algorithm [ J ] based on an energy function, an electronic and informatics newspaper, 2017, (6) 1387 + 1393; nannuo, Cao Dong crystal, Zhang hong Wei, and the like, a space-time correlation-based star atlas noise reduction algorithm [ J ] space return and remote sensing, 2017,38(1):88-97 ], mainly comprises mean filtering, median filtering, Gaussian low-pass filtering, wiener filtering and the like, but star targets in the star atlas are in dotted distribution and have higher similarity with noise distribution, and the traditional image denoising method also loses original information of the stars while removing noise, thereby having obvious limitation. Thresholding [ Katake A B.modeling, imaging and estimation of high speed stage sensors [ D ]. Texas: TexaA & M University,2009: 103-) -104; the method comprises the steps of measuring a one-dimensional maximum entropy star point image segmentation algorithm [ J ] of a small-view field star map of a robot, surveying and mapping, 2018,47(4): 446-. Because the star targets are distributed in a dotted manner and have higher similarity with noise, the existing filtering denoising and threshold segmentation denoising method is not suitable.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide an active pixel sensor star map denoising method based on a dark channel noise template, which is based on the characteristics of noise, separates a star target and background noise by taking the capacity distribution of a dark channel as the noise template, and secondarily filters residual noise and dead spots, so that the star map can be effectively denoised to obtain star target information; one or more problems caused by active pixel sensor star map noise are substantially eliminated.
The purpose of the invention is realized by the following technical scheme:
an active pixel sensor star image denoising method based on a dark channel noise template comprises the following steps:
a, calculating the frame number of a star map, and extracting a dark channel noise template of the star map of the active pixel sensor according to the minimum value of the pixels of the sequence star map images;
b, carrying out difference on the sequence star map and a dark channel noise template, separating noise from the original star map, and realizing continuous multi-frame active pixel sensor star map image denoising;
and C, filtering out the residual single-point and multi-point noises of the star map image after the noise removal based on the dark channel template.
One or more embodiments of the present invention may have the following advantages over the prior art:
the method solves the problem that the strip noise and the background noise in the star map of the active pixel sensor influence the centroid extraction precision, can improve the fixed star centroid positioning precision and further improve the absolute attitude precision of the satellite, and adopts a dark channel noise template denoising method to realize the active pixel sensor star map denoising processing technology.
Drawings
FIG. 1 is a flow chart of an active pixel sensor star map denoising method based on a dark channel noise template;
FIG. 2 is a schematic diagram of dark channel noise template extraction provided by an embodiment of the present invention;
FIG. 3 is a schematic diagram of single-point and multi-point noise filtering according to an embodiment of the present invention;
fig. 4a, 4b, 4c and 4d are a diagram of an original star map, a de-noising result map, a three-dimensional display of the original star map and a three-dimensional display of the de-noising result according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the following embodiments and accompanying drawings.
As shown in fig. 1, a flow of an active pixel sensor star map denoising method based on a dark channel noise template includes the following steps:
and step 30, filtering the residual single-point and multi-point noises of the star map image after the noise removal based on the dark channel template.
The step 10 specifically includes the following steps:
calculating the number of star map frames exposed in the time required by the star facula to move out of one star map, wherein the calculation formula is as follows:
wherein R _ star is the star spot diameter, T is the star map exposure time, S is the displacement of the star, T is the accumulated time within the displacement S,
Is to round up upwards;
extracting a dark channel noise template, wherein the minimum value of pixel energy in the continuous c-frame star image is regarded as noise, and the mathematical expression for extracting the noise template is as follows:
wherein, InoiseFor extracted noise template information, x is the pixel value within the template, y is the pixel value of the star map image, JcFor a star map image sequence, Ω (x) is a window centered on pixel x.
The step 10 is to extract a star map noise template mainly through a dark channel: by utilizing the characteristic that the noise has small change in continuous multi-frame star maps, the capacity value of both strip noise and system background noise is far smaller than the capacity value of star points, and the problem that mass center extraction precision is influenced by a large amount of strip noise and background noise in star map images is solved. The method filters the residual monotonous and multipoint noises in the star map, further purifies the star map image, solves the problems of ring points and the like in the star map, and lays a foundation for improving the accuracy of centroid extraction and satellite attitude determination.
The step 20 specifically includes: denoising an active pixel sensor star map based on a dark channel noise template, carrying out difference (subtracting template noise from a sequence star map as shown in formula 3) through the sequence star map and the dark channel noise template, separating noise from stars and noise in an original star map, and realizing continuous multiframe active pixel sensor star map denoising, wherein the mathematical description is shown as the following formula:
wherein,
representing the de-noising result of the star map, JiRepresenting a sequence of active pixel sensor star maps.
The step 30 specifically includes:
the star map image denoised based on the dark channel template has residual single-point and multi-point noises, as shown in fig. 2, the star map image denoised by the dark channel template is converted into a binary image, and a target is marked through an eight-connected domain.
When the target area is larger than a threshold value delta, the target area is regarded as a star point; when the target area is smaller than the threshold value delta, the noise is regarded as noise, and the single and multi-point noises are filtered again through the threshold value, so that a star map image with the single and multi-point noises filtered is obtained, as shown in fig. 3.
The embodiment can randomly extract the original star map data acquired by the star sensor; preferably, the present embodiment takes the original star map data downloaded from the resource three-satellite as an example, but is not limited thereto. Original star map data downloaded by an orbit satellite are randomly extracted, the data acquisition time of the section is 2012,1, 14 days, the duration is about 62 seconds, the sampling frequency is 2Hz, and the track of a fixed star is theoretically considered to be a straight line in a short time.
According to a time calculation formula of a star spot moving out of a star map:
wherein,
is rounded up.
Taking the first frame star map as an example, the radius of a certain star is 5 pixels, the coordinates are (63.842,84.587), the coordinates of the star in the 20 th frame star map are (66.677,103.265), and the star displacement S
Shift of star frame number c of exposure in S
And taking out the minimum gray value of each pixel in the continuous 6 frames of star map images, and storing the minimum gray value into a two-dimensional matrix template which has the same size with the star map and is the dark channel noise template.
And carrying out difference on the 124-frame star images and the dark channel noise template in sequence to obtain a denoising result of the dark channel noise template.
And (3) carrying out binarization on the dark channel denoising result, calculating the area of each target, regarding the target as a star image point if the area is larger than a threshold value 6, and regarding the target as noise if the area is smaller than the threshold value 6, so as to realize re-denoising of the residual single-point and multi-point noise in the star image.
Fig. 4a, 4b, 4c and 4d are a graph showing an original star map, a denoising result map, a three-dimensional display of the original star map and a three-dimensional display of the denoising result in the present embodiment.
Although the embodiments of the present invention have been described above, the above descriptions are only for the convenience of understanding the present invention, and are not intended to limit the present invention. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (3)
1. An active pixel sensor star image denoising method based on a dark channel noise template is characterized by comprising the following steps:
a, calculating the frame number of a star map, and extracting a dark channel noise template of the star map of the active pixel sensor according to the minimum value of the pixels of the sequence star map images;
b, carrying out difference on the sequence star map and a dark channel noise template, separating noise from the original star map, and realizing continuous multi-frame active pixel sensor star map image denoising;
c, filtering out single-point and multi-point noises remained in the star map image after the noise removal based on the dark channel template;
the step A specifically comprises the following steps:
a1 calculating the exposed star map frame number in the time required for the star light spot to move out of a star map, the calculation formula is:
wherein R _ star is the star spot diameter, T is the star map exposure time, S is the displacement of the star, T is the accumulated time within the displacement S,
Is to round up upwards;
a2 considers the minimum value of pixel energy in the star map images of continuous c frames as noise, and the mathematical expression for extracting the noise template is as follows:
wherein, InoiseFor the extracted noise template information, x is the pixel value in the template, y is the pixel value of the star map image, JcRepresenting a sequence of star images, omega (x) representing a window centred on pixel x;
the step C specifically comprises the following steps:
c1, converting the star map image subjected to the dark channel template denoising into a binary image, and marking the target area in the binary image through an eight-connected domain;
c2 judges whether the marked target of the binary image is a star point or noise according to the size of the target area threshold value, and filters the single-point and multi-point noise in the star image as noise.
2. The method of claim 1, wherein in step B: the mathematical expression for realizing the denoising of the star map of the continuous multiframe active pixel sensor is as follows:
wherein,
de-noising result, J, for the star mapiIs a sequence of star maps of active pixel sensors.
3. The method of claim 1, wherein the target area is greater than a threshold δ and is considered as a star point; when the target area is smaller than the threshold δ, it is regarded as noise.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910274241.8A CN109978797B (en) | 2019-04-08 | 2019-04-08 | Active pixel sensor star image denoising method based on dark channel noise template |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910274241.8A CN109978797B (en) | 2019-04-08 | 2019-04-08 | Active pixel sensor star image denoising method based on dark channel noise template |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109978797A CN109978797A (en) | 2019-07-05 |
| CN109978797B true CN109978797B (en) | 2020-06-23 |
Family
ID=67083136
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910274241.8A Active CN109978797B (en) | 2019-04-08 | 2019-04-08 | Active pixel sensor star image denoising method based on dark channel noise template |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109978797B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111402176B (en) * | 2020-04-21 | 2023-02-14 | 中国科学院光电技术研究所 | Method for removing APS star sensor fixed mode noise in real time on orbit |
| CN112767392B (en) * | 2021-03-02 | 2024-04-09 | 百果园技术(新加坡)有限公司 | Image definition determining method, device, equipment and storage medium |
| CN113379629B (en) * | 2021-06-08 | 2024-08-16 | 深圳思谋信息科技有限公司 | Satellite image denoising method, device, computer equipment and storage medium |
| CN114820665B (en) * | 2022-06-30 | 2022-09-02 | 中国人民解放军国防科技大学 | A method, device, computer equipment and storage medium for suppressing star map background |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108010028A (en) * | 2017-12-27 | 2018-05-08 | 北京航空航天大学 | A kind of star observation method and device for satellite survey meter |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8218013B1 (en) * | 2008-05-27 | 2012-07-10 | Exelis, Inc | Star sensing for an earth imaging sensor |
| CN101763513A (en) * | 2010-02-26 | 2010-06-30 | 成都三泰电子实业股份有限公司 | Foreground extraction method for removing light effect |
| CN102103699B (en) * | 2011-01-27 | 2012-09-26 | 华中科技大学 | Method for detecting boll opening of cotton based on image detection |
| CN104599246B (en) * | 2014-12-29 | 2017-09-29 | 北京控制工程研究所 | A kind of adaptive filtering method of star sensor row pixel noise |
| CN105957058B (en) * | 2016-04-21 | 2019-01-04 | 华中科技大学 | A kind of preprocess method of star chart |
| CN107590777B (en) * | 2017-07-17 | 2020-01-17 | 中国人民解放军国防科学技术大学 | A kind of star sensor star point image enhancement method |
-
2019
- 2019-04-08 CN CN201910274241.8A patent/CN109978797B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108010028A (en) * | 2017-12-27 | 2018-05-08 | 北京航空航天大学 | A kind of star observation method and device for satellite survey meter |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109978797A (en) | 2019-07-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109978797B (en) | Active pixel sensor star image denoising method based on dark channel noise template | |
| CN115131420B (en) | Visual SLAM method and device based on keyframe optimization | |
| CN106530347B (en) | Stable high-performance circle feature detection method | |
| CN104077777B (en) | Sea surface vessel target detection method | |
| CN108510467A (en) | SAR image target recognition method based on variable depth shape convolutional neural networks | |
| CN107146239B (en) | Satellite video moving target detection method and system | |
| CN102663385B (en) | A method for point target detection on satellite | |
| CN101860664A (en) | A Moving Object Detection Method Robust to Illumination Changes | |
| CN114037968A (en) | Lane line detection method based on depth radar point cloud and image data fusion | |
| CN102750683B (en) | Filtering method for sea surface stripe noise and sea surface stripe cloud in moderate resolution imaging spectroradiometer (MODIS) remote sensing image | |
| CN102509276B (en) | Weighted constraint-based star atlas segmentation method | |
| CN111089586B (en) | All-day star sensor star point extraction method based on multi-frame accumulation algorithm | |
| CN111145198A (en) | A non-cooperative target motion estimation method based on fast corner detection | |
| CN118644535A (en) | A star map registration method based on semi-analytical Fourier transform of star point coordinates | |
| CN113554552B (en) | Micro-scanning super-resolution method for infrared point target detection | |
| CN112634295B (en) | A star point segmentation method of star sensor based on double gradient threshold | |
| CN109239384B (en) | A two-dimensional and three-dimensional fusion of non-cooperative target speed and shaft measurement method | |
| Xiang et al. | A modified joint trilateral filter based depth map refinement method | |
| CN118898793B (en) | A landmark rapid positioning and recognition method and system based on drone vision technology | |
| CN113850835A (en) | A near real-time moving target detection method and system in a large field of view in the wild | |
| Meng et al. | An automatic registration method for SAR and optical images based on line extraction and control points selection | |
| CN119832285B (en) | Vehicle matching method and system based on image recognition | |
| CN120876836B (en) | Methods and systems for close-range space target detection and intelligent identification of special components | |
| CN116664528B (en) | Method and device for detecting satellite-borne SAR time sequence image change | |
| CN120031907A (en) | A star image segmentation method based on local background estimation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |