CN102236792B - The characteristic point of lunar surface image is selected and extracting method - Google Patents
The characteristic point of lunar surface image is selected and extracting method Download PDFInfo
- Publication number
- CN102236792B CN102236792B CN201010160706.6A CN201010160706A CN102236792B CN 102236792 B CN102236792 B CN 102236792B CN 201010160706 A CN201010160706 A CN 201010160706A CN 102236792 B CN102236792 B CN 102236792B
- Authority
- CN
- China
- Prior art keywords
- center
- radius
- brightness
- shade
- large area
- 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
Landscapes
- Image Analysis (AREA)
Abstract
The characteristic point that the present invention relates to a kind of lunar surface image is selected and extracting method, for realizing the automatic three-dimensional reconstruction of the face whole month. The method, according to the feature of lunar crater, has been selected 5 category feature points, respectively: the line center at center, large area shadow region, large area highlight regions center, shade center, brightness center, shade center and brightness center; And from lunar surface image, extract this five category features point. The steps include: the shade of each image pixel intensity and setting, highlighted luminance threshold to compare, extract center, large area shadow region and large area highlight regions center. To lunar surface image scan, in extraction column and row direction, brightness is all the pixel of minimum or maximum, is shade center or brightness center. According to sunshine incident direction, shade center and brightness center are matched, extract their line center. Advantage is: the method has been utilized the natural characteristic of lunar surface dexterously, calculates simple and quickly, and the characteristic point density of extraction is high.
Description
Technical field
The invention belongs to the technical field of computer vision and GIS-Geographic Information System, relate to the feature point extraction of image, particularly relate toAnd in reconstructing three-dimensional model process, the method that boat (defending) sheet is carried out to characteristic point selection and extraction.
Background technology
The successful launch of " Chang'e I " satellite, has caused showing great attention to of International Politics, military affairs and academia, its task itOne is exactly the 3-dimensional image figure that obtains lunar surface.
Feature point extraction is prerequisite and the basis of corresponding image points coupling, and corresponding image points coupling is to realize automatically building the moon by imageThe key of face 3-dimensional image figure, quantity, stability and the accuracy of feature point extraction directly affect corresponding image points matching rate, go forward side by sideOne step determines resolution ratio and the precision of the lunar surface 3-dimensional image figure automatically building. This is because generating three-dimensional striograph need to obtain notWith the corresponding image points on image, and calculate the three-dimensional coordinate of each pixel according to principle of stereoscopic vision, conventional method needs moreManual intervention, and to utilize the detection data of laser altimeter supplement and merge. If can be automatically from bidimensional imageExtract the characteristic point of sufficient density, and to CCD stereoscopic camera captured face, forward sight and backsight three width different angles imagesCharacteristic point mate, obtain corresponding image points, and then calculate the three-dimensional coordinate of this point, so just can avoid a large amount of peopleWork is intervened, and improves one-tenth figure speed and the quality of lunar surface 3-dimensional image figure. As can be seen here, extracting lunar surface image feature point should possess to the greatest extentDensity that may be high, and meet higher accuracy simultaneously. The at present extraction of characteristic point normally for different characteristics of image andDifferent application backgrounds, adopts various extraction operators. Moravec operator, Forstner operator are calculated as traditional feature point extractionSon, principle is directly perceived, algorithm is simple, convenience of calculation, but for complicated image, extraction of low quality. Harris operatorAlgorithm is simple, accurate positioning, is not subject to the impact of image rotation, grey scale change, but nonshrinkly put, anti-affine deformation function a little less than.SIFT operator is not subject to the impact of graphical rule and rotation, and light variation, noise, affine deformation are all had to robustness. SUNSANOperator is compared the effective of rim detection to the detection of angle point, is applicable to the image registration based on corners Matching. But above-mentioned these are normalWith method can not adapt to the image feature of lunar surface completely, and cannot meet lunar surface 3-dimensional image figure and automatically build the number to characteristic pointThe demand of amount, stability and accuracy.
Summary of the invention
The object of the invention is to solve in the automatic building process of moon 3-dimensional image figure, from lunar surface bidimensional image, which type of extractsFeature and the problem how to extract, the characteristic point that proposes a kind of lunar surface image is selected and extracting method.
Characteristic point selection and the extracting method implementation of lunar surface image of the present invention are as follows:
1. according to the feature of menology lunar crater, 5 category feature points have been designed: be respectively center, large area shadow region, large area is highlightedThe line center at regional center, shade center, brightness center, shade center and brightness center.
2. the selection of above-mentioned 5 category feature points and extracting method comprise the following steps:
Step 1: determine respectively shade and highlighted luminance threshold, according to the magnitude relationship of the each pixel intensity of lunar surface image and threshold value,Calculate center, large area shadow region and large area highlight regions center, and the parameter of computational representation feature sizes---halfFootpath. Wherein, center (x, y) and their characteristic radius (radius) are determined by following formula:
In formula, n is the pixel number in large area shade or large area highlight regions, (ci，ri) be i shadow spots or highlightedThe pixel coordinate of point, lum[i] be the brightness of i shadow spots or high bright spot, weight[lum] be that gray scale is the pixel of lumWeights.
Step 2: lunar surface image is carried out respectively to line scanning and column scan, extract brightness on column direction and line direction and be all minimumPixel, be shade center; On extraction column direction and line direction, brightness is all the pixel of maximum, is brightness center;Then the parameter of computational representation feature sizes---parameter of radius and characteristic feature intensity of variation---gradient.
The calculating that characterizes the characteristic parameter radius (radius) of shade center or brightness center P is divided into two steps: first with shade centerOr brightness center P in the row direction with column direction on characteristic length smaller as initial radium radius, then taking P as roundThe heart, radius is radius, sets up circular window, according to from distance of center circle from trying to achieve successively AverageLum[i], i represents from the center of circleDistance, AverageLum[i] represent be the mean flow rate of all pixels of i apart from the center of circle, i=radius ..., 1,0, ifAverageLum[i] < AverageLum[i-1], radius subtracts 1, the like, obtain final radius.
Characteristic parameter gradient diff computing formula is as follows:
diff＝(AverageLum[0]-AverageLum[radius])/radius
Step 3: according to sunshine incident direction, shade center and brightness center are matched, be extracted into right shade center andThe parameter of the line center at brightness center, and the parameter of computational representation feature sizes---radius and characteristic feature intensity of variation---gradient.
The establishment at line center is according to sun almanac data, calculates the incident direction of sunray, is specified to right shade centerWith brightness center, and paired shade center and brightness center have uniqueness. Wherein line central feature parameter radius(radius) and gradient (diff) determined by the parameter at its corresponding shade center and brightness center:
radius＝radiuss+radiusb
diff＝(diffs+diffb)/2
Wherein, radiussThe radius at shade center, radiusbIt is the radius at highlighted center; DiffsIt is shade centerGradient, diffbIt is the gradient at highlighted center.
The characteristic point selection of lunar surface image of the present invention and the advantage of extracting method:
1. because utilize dexterously the natural characteristic of lunar surface, selected 5 category feature points, thereby broken away from structure moon 3-dimensional imageWhen figure, conventional method needs the drawback of a large amount of manual interventions, and automaticity is high;
2. the extraction algorithm of 5 class lunar surface image feature points of design is simple and quick, and extraction efficiency is high;
3. according to five class lunar surface image feature point extraction algorithms of design, the characteristic point density of extracting is high, is moon 3-dimensional imageThe basis that figure autocreating technology is realized.
Brief description of the drawings
The center, large area shadow region (square frame) that Fig. 1 obtains according to characteristic point selection and the extracting method of lunar surface image and large faceLong-pending highlight regions center (triangle) schematic diagram;
The shade center that Fig. 2 obtains according to characteristic point selection and the extracting method of lunar surface image and the (side, line center at brightness centerFrame) schematic diagram.
Shade center (square frame) and brightness center (triangle that Fig. 3 obtains according to characteristic point selection and the extracting method of lunar surface imageShape) schematic diagram;
Detailed description of the invention
Now in conjunction with the accompanying drawings and embodiments characteristic point selection and the extracting method of lunar surface image of the present invention are described in further detail. RealExecute the feature point extraction of the 168th rail image of taking taking " Chang'e I " satellite in example as example. Lunar surface image feature of the present invention clicksSelecting and extracting method, is according to the various lunar craters of the menology feature that image has under solar light irradiation, designs and select five class spiesLevy a little, respectively: center, large area shadow region, large area highlight regions center, shade center, brightness center, relevant cloudyThe line center at shadow center and brightness center; And designed the extraction algorithm of 5 category feature points, can be rapidly from lunar surface two dimension shadowIn picture, extract a large amount of characteristic points, Fig. 1, Fig. 2, Fig. 3 are respectively the variety classes characteristic points that go out from the 168th rail Extraction of ImageSchematic diagram, because image size is very large, has only intercepted partial image. The characteristic point of lunar surface image of the present invention is selected and extracting methodBe described in detail as follows:
1. a large amount of impact crater of lunar surface has formed some large-area shadow region and highlight bars under solar light irradiation, and their features are brightAobvious, positioning precision is high, and uniqueness and high conformity, so select large area shadow region and the center conduct of large area highlight regionsCharacteristic point.
2. shade center and brightness center are image local area brightness extreme points: in the time that this point is brightness maximum point, be called in brightnessThe heart, is called shade center during for brightness minimum point, this two category feature is because the landform that lunar surface height rises and falls is subject to solar light irradiationUneven and form, their exist in a large number in image, and positioning precision is higher, and uniformity is better, can make up large area shadeRegion and the sparse shortcoming of large area highlight regions central distribution, therefore select shade center and brightness center as characteristic point.
3. a lot of shade centers and brightness center are to occur in pairs, and the implicit information in their line center is than shade center and brightnessCenter is more, in specific region, can filter out the line center with uniqueness according to certain condition, therefore selectsThe line center at shade center and brightness center is as a category feature point.
4. the extracting method at center, large area shadow region is:
1) establishing large area shadow region luminance threshold is MASS_SILL, and brightness weight is f.
2) get 1 P on image, its coordinate is (c, r), and brightness is I (c, r). If I (c, r) > is MASS_SILL, PPoint does not belong to shadow region, and algorithm is got back to step 2) continuation detection consecutive points; Otherwise, by stacked this point, enter step 3).
3) judge whether stack is empty. In the time of stack non-NULL, stack top element is popped, and as current detection point, enters step 4); When stack is emptyEnter step 5).
4) if current detection point has been inserted the queue of large area shadow region, return to step 3); Otherwise current detection point is inserted largeThe queue of area shadow region, then checks whether current detection point around has not point and this point after testing meetI≤MASS_SILL, has stacked this point, and returns to step 3).
5) from the queue of large area shadow region, (queue length is to take out successively each pixel n), calculates large area shade according to following formulaThe parameter radius r adius of the center (x, y) in region and sign provincial characteristics size.
In formula, n is the pixel number in large area shadow region, (ci，ri) be the pixel coordinate of i shadow spots,Lum[i] be the brightness of i shadow spots, weight[lum] be that gray scale is the weights of the pixel of lum, by formulaweight[lum]＝f(MASS_SILL-lum)Determine; In the present embodiment, large area shadow region luminance threshold is
MASS_SILL=32, brightness weight is f=1.5.
1. the extracting method at the extracting method at large area highlight regions center and center, large area shadow region is similar, and difference isTest point brightness is greater than large area highlight regions luminance threshold HIGH_SILL and is considered to this point and belongs to a large area highlight regions;In the present embodiment, large area luminance area luminance threshold is HIGH_SILL=110, and brightness weight is f=1.5.
2. the extracting method at shade center is:
1) to image line scanning, upper structure brightness slope (brightness monotone increasing or decline pixel in row or column successive range in the row directionThe interval that point forms), find out the slope that slope aspect rises, getting its left end point is minimum point, gets in the length of grade on these left and right two slopes of pointOne of smaller as the parameter l en1 that characterizes this minimum characteristic length size on line direction.
2) to image column scanning, using the method identical with previous step to find out brightness is minimizing pixel and characteristic length len2 thereof.
3) find out on column direction and line direction on brightness be all minimizing pixel, be shade center.
4) the characteristic attribute parameter at computational shadowgraph center, mainly comprises that characteristic feature puts big or small parameter radius and characteristic feature point slopeThe parameter d iff of degree. Get the value that len1, len2 smaller are parameter radius. But at this moment only investigate this shade centerFour direction, the value of parameter radius is inaccurate, and must further determine: shade center is designated as a P, and the first step is with a PCentered by, set up the circular window that radius is (2 × radius+1), obtain successively and put P distance for i's (0≤i≤radius)Point average brightness AverageLum[i]; Second step, according to i order from big to small, compares AverageLum[i successively] andAverageLum[i-1], some P is shade center, so above-mentioned average brightness meetsAverageLum[i] > AverageLum[i-1], if do not met, radius subtracts 1. After determining radius, characteristic featureThe parameter d iff of the some gradient can be by formula diff=AverageLum[0]-AverageLum[radius] try to achieve.
3. the extracting method at the extracting method at brightness center and shade center is similar, difference be brightness center be line direction andBrightness maximum point on column direction.
4. the extracting method at the line center at shade center and brightness center is:
1) find out the brightness center paired with shade center possibility. Set a range threshold R, taking shade center as the center of circle,Radius is the interior brightness center of finding of the circle of R.
2) in the brightness of finding in previous step,, determine a brightness center paired with shade center according to radiation direction in the heart.If shade centre coordinate is (is，js), brightness centre coordinate is (ib，jb), sunshine incident vector is (dx, dy, 0), establishdi＝is-ib，dj＝js-jb，Cosa=(the d if satisfy conditioni×dx+dj×dy)/r >=cos (π/6),Can think that shade center and this candidate's brightness center are a pair of. If exist multiple candidate's brightness center to meet above-mentioned condition, in view ofPaired shade center and brightness center have uniqueness, for avoiding mismatching, should abandon extracting this in shade center and brightnessThe line center of the heart.
3) calculated characteristics property parameters. The characteristic parameter of line is determined by the shade center of being correlated with it and brightness center, is wherein satI=(i in mark (i, j)s+ib)/2，j＝(js+jb)/2, the parameter radius=radius of sign scopes+radiusb, the sign gradientParameter d iff=(diffs+diffb)/2。
The characteristic point of the lunar surface image that above-mentioned case realizes is selected and extracting method test environment is: CPUPIV2.8G, and in 1GDeposit, 120G hard disk, operating system WindowsXP, monitor resolution 1024 × 768, (128M is aobvious for GEFORCE6200Deposit) video card. Image size is 512 × 32554, and extracting all kinds of characteristic point numbers is 120,536.
Claims (2)
1. the characteristic point of lunar surface image is selected and an extracting method, it is characterized in that, the method, according to the feature of menology lunar crater, is establishedCount and selected 5 category feature points: center, large area shadow region, large area highlight regions center, shade center, brightness center,The line center at shade center and brightness center; The characteristic point of lunar surface image is selected and the realization of extracting method comprises the following steps:
Step 1: determine respectively shade and highlighted luminance threshold, according to the magnitude relationship of the each pixel intensity of lunar surface image and threshold value,Extract center, large area shadow region and large area highlight regions center, and the parameter of computational representation feature sizes---radius;
Step 2: lunar surface image is carried out respectively to line scanning and column scan, extract brightness on column direction and line direction and be all minimumPixel, be shade center; On extraction column direction and line direction, brightness is all the pixel of maximum, is brightness center;Then the parameter of computational representation feature sizes---parameter of radius and characteristic feature intensity of variation---gradient;
Step 3: according to sunshine incident direction, shade center and brightness center are matched, be extracted into right shade center andThe parameter of the line center at brightness center, and the parameter of computational representation feature sizes---radius and characteristic feature intensity of variation---gradient;
The wherein center (x, y) of the center, large area shadow region described in step 1 and large area highlight regions and its spyLevy parameter radius r adiusmDetermined by following formula:
In formula, n is the pixel number in large area shade or large area highlight regions, (ci，ri) be i shadow spots or highlightedThe pixel coordinate of point, lum[i] be the brightness of i shadow spots or high bright spot, weight[lum] be that gray scale is the pixel of lumWeights;
In step 2, characterize the characteristic parameter radius r adius of shade center or brightness center PoCalculating be divided into two steps: first withShade center or brightness center in the row direction with column direction on characteristic length smaller len as initial radium radiuso; SoAfter taking P as the center of circle, radiusoFor radius, set up circular window, according to from distance of center circle from trying to achieve successively AverageLum[i],I represents the distance from the center of circle, AverageLum[i] represent to be the mean flow rate of all pixels of i apart from the center of circle, i=len ..., 1,0,If AverageLum[i is worked as at shade center] < AverageLum[i-1], radiusoSubtract 1, if in brightnessThe heart, works as AverageLum[i] > AverageLum[i-1], radiusoSubtract 1, the like, final radius obtainedo；
The characteristic parameter gradient diff of shade center or brightness center PoComputing formula is as follows:
diffo＝(AverageLum[0]-AverageLum[radiuso])/radiuso；
Line central feature parameter radius r adius in step 3cAnd gradient diffcBy its corresponding shade center and brightness centerParameter determine:
radiusc＝radiuss+radiusb
diffc＝(diffs+diffb)/2
Wherein, radiussThe radius at shade center, radiusbIt is the radius at brightness center; DiffsIt is shade centerGradient, diffbIt is the gradient at brightness center.
2. a kind of characteristic point of lunar surface image is selected and extracting method according to claim 1, it is characterized in that: described stepThe establishment at three line centers is according to sun almanac data, calculates the incident direction of sunray, be specified to right shade center andBrightness center, and paired shade center and brightness center have uniqueness.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010160706.6A CN102236792B (en) | 2010-04-30 | 2010-04-30 | The characteristic point of lunar surface image is selected and extracting method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010160706.6A CN102236792B (en) | 2010-04-30 | 2010-04-30 | The characteristic point of lunar surface image is selected and extracting method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102236792A CN102236792A (en) | 2011-11-09 |
| CN102236792B true CN102236792B (en) | 2016-05-18 |
Family
ID=44887430
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201010160706.6A Expired - Fee Related CN102236792B (en) | 2010-04-30 | 2010-04-30 | The characteristic point of lunar surface image is selected and extracting method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102236792B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104913784B (en) * | 2015-06-19 | 2017-10-10 | 北京理工大学 | A kind of autonomous extracting method of planetary surface navigation characteristic |
| CN113837095B (en) * | 2021-09-24 | 2023-08-08 | 福州大学 | Terrain correction effect evaluation method based on three types of shadows |
-
2010
- 2010-04-30 CN CN201010160706.6A patent/CN102236792B/en not_active Expired - Fee Related
Non-Patent Citations (1)
| Title |
|---|
| 《嫦娥一号月面影像图自动生成软件的设计与实现》;李学军等;《装备指挥技术学院学报》;20090228;第20卷(第1期);85-90 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102236792A (en) | 2011-11-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20220148213A1 (en) | Method for fully automatically detecting chessboard corner points | |
| CN106650640B (en) | Negative obstacle detection method based on laser radar point cloud local structure characteristics | |
| Xu et al. | Zoomnet: Part-aware adaptive zooming neural network for 3d object detection | |
| CN104134234B (en) | A kind of full automatic three-dimensional scene construction method based on single image | |
| CN108537782B (en) | Building image matching and fusing method based on contour extraction | |
| CN104200461B (en) | The remote sensing image registration method of block and sift features is selected based on mutual information image | |
| CN106803286A (en) | Mutual occlusion real-time processing method based on multi-view image | |
| CN107451982A (en) | A kind of high canopy density standing forest tree crown area acquisition methods based on unmanned plane image | |
| CN111028271B (en) | Multi-camera personnel three-dimensional positioning and tracking system based on human skeleton detection | |
| CN106504284A (en) | A kind of depth picture capturing method combined with structure light based on Stereo matching | |
| CN111369495B (en) | Panoramic image change detection method based on video | |
| CN111222511B (en) | Infrared unmanned aerial vehicle target detection method and system | |
| CN101976436B (en) | Pixel-level multi-focus image fusion method based on correction of differential image | |
| CN112669280B (en) | Unmanned aerial vehicle inclination aerial photography right-angle image control point target detection method based on LSD algorithm | |
| CN117036641A (en) | Road scene three-dimensional reconstruction and defect detection method based on binocular vision | |
| CN111354047B (en) | Computer vision-based camera module positioning method and system | |
| CN110189375A (en) | A kind of images steganalysis method based on monocular vision measurement | |
| Zhang et al. | Lidar-guided stereo matching with a spatial consistency constraint | |
| CN102236792B (en) | The characteristic point of lunar surface image is selected and extracting method | |
| CN111080712B (en) | Multi-camera personnel positioning, tracking and displaying method based on human body skeleton detection | |
| CN116402904A (en) | Combined calibration method based on laser radar inter-camera and monocular camera | |
| CN116245872A (en) | Aircraft structure crack detection method based on convolutional neural network | |
| US11699303B2 (en) | System and method of acquiring coordinates of pupil center point | |
| Liu et al. | Color component–based road feature extraction from airborne lidar and imaging data sets | |
| Jorissen et al. | Multi-view wide baseline depth estimation robust to sparse input sampling |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20160518 Termination date: 20170430 |