CN110992330A - Multi-level integral relaxation matching high-resolution ortho-image shadow detection under artificial shadow drive - Google Patents
Multi-level integral relaxation matching high-resolution ortho-image shadow detection under artificial shadow drive Download PDFInfo
- Publication number
- CN110992330A CN110992330A CN201911195527.3A CN201911195527A CN110992330A CN 110992330 A CN110992330 A CN 110992330A CN 201911195527 A CN201911195527 A CN 201911195527A CN 110992330 A CN110992330 A CN 110992330A
- Authority
- CN
- China
- Prior art keywords
- shadow
- building
- matching
- image
- artificial
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 18
- 238000000034 method Methods 0.000 claims abstract description 23
- 239000011159 matrix material Substances 0.000 claims description 22
- 238000012216 screening Methods 0.000 claims description 11
- 230000005484 gravity Effects 0.000 claims description 5
- 230000011218 segmentation Effects 0.000 claims description 4
- 230000000903 blocking effect Effects 0.000 claims description 3
- 238000003708 edge detection Methods 0.000 claims description 2
- 238000000605 extraction Methods 0.000 abstract 1
- 238000004364 calculation method Methods 0.000 description 10
- 238000010586 diagram Methods 0.000 description 3
- 230000003595 spectral effect Effects 0.000 description 3
- 239000013598 vector Substances 0.000 description 3
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/66—Analysis of geometric attributes of image moments or centre of gravity
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Geometry (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Quality & Reliability (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a high-resolution ortho-image shadow detection method for multilayer integral relaxation matching under the drive of artificial shadows. Firstly, acquiring a shadow contour of a single building in an orthoimage and an artificial shadow image by using an edge extraction algorithm, and establishing a shadow data set; then, layering the two sets of data sets according to the building height, and performing overall relaxation matching on the multi-level parallel surface characteristics of the shadow data set to obtain an initial matching result; and finally, performing surface feature integral matching on the building shadows in the unmatched set to obtain a final matching result, thereby completing the shadow detection of the high-resolution orthoimage. By using the method and the device for detecting the building shadow based on the artificial shadow, the position of the building shadow can be determined more directly, and the boundary of the building shadow can be obtained more accurately.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a high-resolution ortho image building shadow detection method.
Technical Field
The artificial shadow image is as follows: the sun altitude and azimuth during aerial image shooting are combined with a Digital Building Model (DBM) of the image, a shadow modeling algorithm is utilized to calculate the geometric range of a shadow area on the ground plane caused by blocking of solar rays by a building at a certain aerial image shooting moment, the geometric area is the required shadow area, and the gray value of the geometric area is assigned to be 0.
There are two general categories of shadow detection methods. The first is a model-based method, which builds a shadow statistical model according to the information in the image, the geometric shape of the scene, the solar altitude, DSM or sensor and other parameters, and distinguishes pixel by pixel. The second type is a shadow detection algorithm based on the property of the shadow, and the shadow region and the non-shadow region are divided by combining methods such as a characteristic threshold, a statistical mixture model, a D-S evidence theory and the like according to the difference of information such as the spectrum, the texture, the edge and the like of the shadow region and the non-shadow region. Through analyzing the current research situation at home and abroad, the following defects of the existing shadow detection algorithm are found:
(1) the detection method based on shadow texture and edge features has many factors influencing shadow formation, and depends on texture features and edge characteristics of a shadow area to realize shadow detection, so that not only a large amount of calculation is introduced, but also the definition of the boundary of the shadow area has considerable difficulty.
(2) The spectral characteristic detection method based on the spectral characteristic interferes with the identification of the shadow region because the spectral characteristics of the water body, the blue-biased ground object and the like on each wave band are very close to the shadow and the peak value of the threshold segmentation histogram is not obvious.
(3) The selection of the threshold requires strong prior knowledge, and the threshold is continuously adjusted according to human experience to achieve a satisfactory effect, so that the applicability of the detection method is reduced.
Disclosure of Invention
The invention provides a method for detecting the shadow of a multi-layer integral loose matched high-resolution ortho-image under the drive of an artificial shadow, which can detect the shadow of a building on the basis of the artificial shadow, can more directly determine the position of the shadow of the building and more accurately acquire the boundary of the shadow of the building.
In order to achieve the technical purpose, the technical scheme of the invention is as follows:
the method for detecting the shadow of the multilayer integrally relaxed matched high-resolution ortho-image under the drive of the artificial shadow comprises the following specific steps:
1. the solar altitude and azimuth angles during aerial image shooting are combined with DBM data of the images, and an artificial shadow image is generated by utilizing a shadow modeling algorithm.
Combining the solar altitude and azimuth angle when the aerial image is shot with DBM data of the image, generating a shadow image by using a shadow modeling algorithm, reclassifying a shadow area and a non-shadow area by using a space reclassification method, classifying the areas with the gray value of 0 into one class, assigning the other gray values to 255 as the same class, and classifying the other gray values into another class to obtain the artificial shadow image.
2. Building shadow outline data sets of the two images are created, and the characteristic attribute of the shadow surface matching unit is calculated.
And extracting the shadow contour in the image by adopting a Canny edge detection algorithm, and carrying out segmentation screening by combining DBM data to obtain a contour data set of the single building shadow. Determining the outline range and the size of the shadow polygons of the building in the two images, and calculating the barycentric coordinate of each shadow polygon; and extracting shadow contour feature points, and calculating the shape center distance value of each shadow surface.
3. Two sets of building shadow data sets are layered.
By comparing the area and shape similarity of the building shadows in the artificial image and the orthographic image, it is found that the higher the building is in the image, the larger the difference in area is. Therefore, each shadow data set is divided into 3 layers according to the height of the building.
4. And performing surface feature integral relaxation matching on the shadow data set of each layer to obtain the accurate position of the building shadow in the orthoimage.
(1) Probability matrix initialization
And calculating the position and shape similarity between corresponding surfaces in the building shadow polygon data set to be matched by taking the building shadow surfaces as matching units. And (4) obtaining a comprehensive initial probability by setting a weight and carrying out weighted average to obtain a similarity matrix between shadow surfaces in the two groups of data. Coarse screening the initial candidate matching pairs, eliminating obviously wrong candidate matching pairs, deleting the candidate matching pairs with similarity smaller than a given threshold value and assigning the similarity to be 0 to obtain a final similarity matrix; and then estimating the initial probability of matching, namely adding the characteristic correlation coefficients of each candidate surface, and dividing the sum by the correlation coefficients of all candidate surfaces to obtain an initial probability matrix.
(2) Probability matrix iterative update
And determining the number of target surface neighborhoods of the shadow surface of the building by taking the gravity center point of the shadow area as a search target through a breadth-first search strategy. After the neighborhood relationship is determined, calculating the feature difference value and the support degree between adjacent candidate matching pairs, and updating the initial matching matrix.
(3) Matching pair selection
And adding three correlation coefficients of shadow neighborhood position similarity, shape similarity and correlation coefficient, and performing iterative updating on the matching probability matrix. And when the matching probability variation of each candidate matching pair is smaller than a given threshold value in the two iterations, stopping the iteration.
And detecting candidate matching pairs of 1: M, N:1 and M: N. And extracting the characteristic points in the image according to a Harris method, and then screening out a characteristic point set of the shadow contour by combining the edge contour of the shadow surface. And taking the average value of curvature values of all the characteristic points contained in each shadow contour as the curvature value of the surface, calculating the similarity value of the curvatures between the candidate matching pairs, and determining the initial matching pair.
5. And (3) a matching process of multilayer integral relaxation matching of the high-resolution ortho-image under the drive of the artificial shadow.
(1) And (3) performing multi-level parallel surface feature overall relaxation matching on the building shadow data set in the step (3), and setting different neighborhood similarity thresholds and total similarity thresholds for each layer according to the shadow feature values to determine the matching pairs of the layer.
(2) And dividing the unmatched building shadows in each matching layer into an artificial shadow unmatched set and an orthoimage unmatched set. And performing surface feature integral matching on the building shadow polygons in the unmatched set according to the same method to obtain a final matching result.
(3) And finally, carrying out global traversal check on a small number of unmatched building shadows, and serving as an effective supplement for improving the matching accuracy.
The method provided by the invention performs integral loose matching on the artificially generated building shadow and the shadow in the ortho-image to obtain the accurate position of the building shadow in the ortho-image. The method can more directly determine the position of the building shadow and more accurately acquire the boundary of the building shadow.
Drawings
FIG. 1 is a flow chart of shadow detection according to an embodiment of the present invention
FIG. 2 is a schematic diagram of the overall matching of surface features according to an embodiment of the present invention
FIG. 3 is an artificial shadow map of an embodiment of the present invention
FIG. 4 is a DBM plane model according to an embodiment of the present invention
FIG. 5 is a shadow matching diagram of a first floor building according to an embodiment of the present invention
FIG. 6 is a diagram of shadow detection results according to an embodiment of the present invention
Detailed Description
The following detailed description of embodiments of the invention refers to the accompanying drawings. It is to be understood that the embodiments described are only some of the embodiments of the invention, and not all of them. All other embodiments, which can be made by those skilled in the art without any inventive presupposition based on the embodiments of the present invention, belong to the scope of the present invention.
Example (b):
in this embodiment, a high-resolution aerial image of a foreign area is selected as experimental data, and the average altitude of the aircraft flight when the data is shot is 1650m, wherein the course overlapping rate is 65% and the side overlapping rate is 30%.
In specific implementation, the technical scheme of the invention can be implemented in a mode of automatic operation of computer programming.
The specific operation steps of the high-resolution ortho-image shadow detection program with multilayer integral relaxation matching under the drive of artificial shadow provided by the invention can refer to a flow chart (figure 1):
step 1, generating an artificial shadow image.
In this example, the DBM is used as a research model, the solar altitude and azimuth angle when the aerial image is shot are combined with the DBM data of the image, a shadow modeling algorithm is used to calculate the geometric range of a shadow area on the ground plane caused by the blocking of the solar rays by the building at a certain aerial shooting moment, and the gray value of the shadow area is assigned to 0. The resulting artificial shadow is shown in figure 3.
And 2, extracting the edge outline of the building shadow, creating a building shadow data set, and calculating the characteristic attribute of the shadow surface matching unit.
In the example, the orthoimage is converted into a binary image, and then a Canny operator is used for extracting the rough contour of the architectural shadow in the image. Combining the two-dimensional DBM data of the images (as shown in figure 3), carrying out segmentation screening on the extracted data, eliminating edge contours in a DBM area, closing the shadow area, and obtaining a building shadow data set data of two images1＝{qi|i＝1…m}，data2＝{tj1 … n, where siAnd tjAre respectively data1、data2M and n are the number of the corresponding surface entities of each data set respectively. Calculating the characteristic attribute of each shadow surface as: si＝{Ai,Ci}，tj＝{Aj,CjWhere d represents a set of barycentric coordinate points of the polygon and C represents a set of perimeters of a minimum bounding rectangle of the polygon.
And 3, layering the shadows in the image according to the height of the building.
In this example, by analyzing the similarity of the shadow planes of the two images corresponding to a single building, it is found that the area difference is larger as the building is higher in the images. Therefore, the shadow data set is divided into 3 layers according to the height of the building: the first-layer shadow data set is a shadow with the building height larger than 200 m; the second-layer shadow data set is a shadow with the building height of more than 100m and less than 200 m; the layer iii shadow dataset is a shadow of building height less than 100 m.
And 4, calculating the difference index of the matching pair according to the characteristic value between the position and the shape of the matching pair.
In this example, the shadow polygon s in the two images is calculated separatelyiAnd tjThe calculation method for the similarity of the positions and the shapes of the corresponding polygons comprises the following steps:
wherein s isi、tjRespectively representing arbitrary polygons in two sets of data, pd、ρCAre respectively candidate matching pairs(s)i,tj) Position, area and shape similarity of (c);
is a polygon si、tjThe coordinates of the center of gravity of the body,
the two features are combined to take the shadow plane as a matching unit. By calculating the data1Each polygon s in (1)iAnd data2And calculating the correlation coefficient of all the candidate matching shadow planes. The calculation method is as follows:
ρ(si,tj)＝ω1·ρd+ω2·ρC(2)
wherein, ω is1，ω2The position and shape similarity, respectively, are weighted with a value of 1/2. And calculating a similarity matrix R of each shadow surface of the two groups of data according to the characteristic values of the shadow surfaces.
And 5, initializing a probability matrix.
Coarse screening the initial candidate matching pairs calculated in the step 3 by setting corresponding position and shape similarity thresholds, removing obviously wrong candidate matching pairs, deleting and assigning the candidate matching pairs with similarity smaller than a given thresholdIs 0, obtaining a final similarity matrix R'M×N. Then, for the relaxation optimization calculation, the data is processed1The initial probability of matching of m polygons is estimated, i.e. the feature correlation coefficients of each candidate surface are added, and divided by the sum of the correlation coefficients of all candidate surfaces to obtain the initial probability. The calculation method is as follows:
calculate matrix R'M×NTo obtain an initial matching probability matrix P(0). Wherein each row represents data1Middle polygon siAnd data2Of n polygons.
And 6, iteratively updating the initial probability matrix.
And determining the number of target surface neighborhoods of the building shadow polygon data S and the data T by using the shadow region gravity center point as a search target through a breadth-first search strategy. And then, calculating the feature difference value and the support degree between the adjacent candidate matching pairs, and updating the initial matching matrix. Order to
Then shAnd tkRespectively represents siAnd tjArea targets within the neighborhood,(s)h,tk) Is(s)i,tj) Is determined. Then, the components are combined and respectively calculated(s)i,tj) And(s)h,tk) Relative position and relative shape relationship therebetween. The calculation method is as follows:
wherein r is1(si,tj；sh,tk)、r2(si,tj；sh,tk) Respectively represent(s)h,tk) And(s)i,tj) Relative position and relative character relationship between them; d (a, b) represents the gravity center distance between two surfaces a and b; c' (a, b) represents the similarity of the shape center distance values of the two faces a and b. Combining the above 2 relative geometric relationships to obtain(s)h,tk) And(s)i,tj) Coefficient r(s) of compatibility betweeni,tj；sh,tk) The calculation method is as follows:
r(si,tj；sh,tk)＝W1·r1(si,tj；sh,tk)+W2·r2(si,tj；sh,tk)+W3·ρ(si,tj) (5)
calculating the matching support degree of each building shadow area according to the formula, and performing iterative updating, wherein the calculation method comprises the following steps:
and 7, selecting a building shadow matching pair.
Extracting characteristic points in the image according to a Harris method from the detected candidate matching pairs of 1: M, N:1 and M: N; screening out a characteristic point set of the shadow profile by combining the edge profile of the shadow surface; and taking the average value of curvature values of all the feature points contained in the profile of each surface as the curvature value of the surface, calculating the similarity value of the curvatures between the candidate matching pairs, and determining the final matching pair.
The calculation method of the shadow surface curvature similarity value is as follows: three feature points P in each set of pointsi-1、Pi、Pi+1Combine to obtain a vector
The set point P of the characteristic points is obtained by calculating the included angle and the approximate radian value between the two vectorsiThe calculation formula is as follows:
ki＝sgn(i)Δαi/Δsi(8)
wherein, Delta αiIs a vector
And
the angle of,
for approximate arc length sgn (i) is a sign function, if point viSgn (i) ═ 1 in the case of convex, and sgn (i) ═ 1 in the case of concave.
Taking the average value of curvature values of all the feature points contained in each polygon outline as the curvature value of the polygon, further calculating similarity values E (A, B) of curvatures between candidate matching pairs, and defining:
where A, B denote the polygon pair to be matched, k0(i) Is a point PiCurvature on the shadow profile in the ortho image, k (i) being the point PiCurvature on the shadow contours in the artificial shadow image.
And 8, performing overall relaxation matching on the multi-level parallel surface characteristics of the building shadow to obtain a final matching result.
In this example, the multilevel parallel surface feature overall relaxation matching is performed on the building shadow data set (step 4-7), and different neighborhood similarity thresholds and total similarity thresholds are set for each layer according to the shadow feature values to determine the matching pairs of the layer. Through experiments, three layers of matching thresholds are respectively set as follows: the similarity value of the first layer matching neighborhood is greater than 0.7, and the total similarity value is greater than 0.65; the similarity value of the second-layer matching neighborhood is greater than 0.7, and the total similarity value is greater than 0.7; the similarity value of the III-level matching neighborhood is greater than 0.8, and the total similarity value is greater than 0.75.
And dividing the unmatched building shadows in each matching layer into an artificial shadow unmatched set and an orthoimage unmatched set, and performing surface feature integral matching on the building shadow polygons in the unmatched sets according to the same method. Wherein, the similarity value of the matching neighborhood is larger than 0.6, and the total similarity value is larger than 0.55.
And finally, carrying out global traversal inspection on a small number of unmatched building shadows to obtain a final matching result. The method comprises the steps of carrying out multi-layer integral loose matching on a building shadow surface in an image, and obtaining a matching result of an artificial shadow image and an orthoimage, so that shadow detection of the orthoimage with high resolution is completed.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.
Claims (8)
1. The method comprises the following specific steps of (1) multi-level integral relaxation-matched high-resolution ortho-image shadow detection under the drive of an artificial shadow:
step 1) generating an artificial shadow image by using a modeling tool;
step 2) building shadow contour data sets of the two images are created, and the characteristic attributes of the shadow surface matching units are calculated and added;
step 3) layering the two groups of shadow data sets according to the height of the building;
step 4) calculating initial probability values of shadow surface features in two groups of image building shadow data sets, and establishing an initial probability matrix;
step 5) calculating a support coefficient of each shadow face neighborhood in the image, and carrying out iterative updating on the initial probability matrix;
step 6) screening candidate matching pairs of 1: M, N:1 and M: N according to the corner point characteristics of the shadow surface to determine the matching pairs;
and 7) carrying out overall relaxation matching on the multi-level features of the building shadow data set in the image to obtain a matching result of the artificial shadow image and the ortho-image, thereby completing the shadow detection of the high-resolution ortho-image.
2. The method for detecting the shadow of the multi-level integral relaxation-matched high-resolution ortho image under the drive of the artificial shadow as claimed in claim 1, wherein: the matching unit feature attributes in the step (2) comprise the position and shape features of the shadow surface:
extracting the shadow contours in the images by adopting a Canny edge detection algorithm, and carrying out segmentation screening on the shadow contours by combining DBM data to determine the contour range and size of the building shadow polygons in the two images; according to the coordinate values corresponding to the contour, the barycentric coordinates of each shadow surface are calculated; and extracting shadow contour feature points, and calculating the shape center distance value of each shadow surface.
3. The method for detecting the shadow of the multi-level integral relaxation-matched high-resolution ortho image under the drive of the artificial shadow as claimed in claim 1, wherein: performing hierarchical matching on the two groups of shadow data sets in the step (3), and performing hierarchical matching according to the building height corresponding to each shadow surface:
and comparing the difference of the shadow surfaces in the two images, summarizing the law of the difference of the shadow surfaces, layering the building shadows in the images by using the height of the building, and dividing each image into 3 layers.
4. The method for detecting the shadow of the multi-level integral relaxation-matched high-resolution ortho image under the drive of the artificial shadow as claimed in claim 1, wherein: initializing two sets of shadow dataset probability matrices as described in step (4):
calculating the position and shape similarity value between the shadow surfaces of the two images to obtain the initial probability value of each shadow surface in the orthoimage; calculating the position and shape similarity between corresponding surfaces in two sets of planar building shadow surface data sets to be matched by taking the building shadow surfaces as matching units; setting a weight value, and carrying out weighted average to obtain a similarity matrix between shadow surfaces in the two groups of data;
coarse screening the initial candidate matching pairs, eliminating obviously wrong candidate matching pairs, deleting the candidate matching pairs with similarity smaller than a given threshold value and assigning the similarity to be 0 to obtain a final similarity matrix; and estimating the initial probability of matching, namely adding the characteristic correlation coefficients of each candidate surface, and dividing the sum by the correlation coefficients of all candidate surfaces to obtain an initial probability matrix.
5. The method for detecting the shadow of the multi-level integral relaxation-matched high-resolution ortho image under the drive of the artificial shadow as claimed in claim 1, wherein: in step (5), the support coefficient of each shadow neighborhood in the image is calculated, and the initial probability matrix is updated iteratively:
determining neighborhood candidate matching pairs of the building shadow surface, calculating the support coefficient of each neighborhood element, continuously iterating the initial probability value, and acquiring the position of the building shadow area in the orthoimage;
determining the number of target surface neighborhoods of the shadow surface of the building by taking the gravity center point of the shadow surface as a search target through a breadth-first search strategy; and calculating the feature difference value and the support degree between the adjacent candidate matching pairs, and updating the initial matching matrix.
6. The method for detecting the shadow of the multi-level integral relaxation-matched high-resolution ortho image under the drive of the artificial shadow as claimed in claim 1, wherein: in the step (6), screening candidate matching pairs 1: M, N:1 and M: N according to the corner feature pairs of the shadow surface to determine a final matching pair:
respectively calculating the relative position and the relative shape relationship between the adjacent candidate matching pairs; calculating the support coefficient of each neighborhood element; performing iterative updating on the matching probability matrix by using three correlation coefficients of the position, the shape similarity and the correlation coefficient;
extracting contour feature points of candidate matching pairs of 1: M, N:1 and M: N; screening out a characteristic point set of the shadow profile by combining the edge profile of the shadow surface; and taking the average value of curvature values of all the feature points contained in the profile of each surface as the curvature value of the surface, calculating the similarity value of the curvatures between the candidate matching pairs, and determining the initial matching pair.
7. The method for detecting the shadow of the multi-level integral relaxation-matched high-resolution ortho image under the drive of the artificial shadow as claimed in claim 1, wherein: in the step (7), performing multi-level parallel surface feature overall relaxation matching on the building shadow data set according to the claims 4-6, and setting different neighborhood similarity threshold values and total similarity threshold values for each layer according to the shadow feature values to determine the matching pairs of the layer; and dividing the unmatched building shadows in each matching layer into an artificial shadow unmatched set and an orthoimage unmatched set, and performing surface feature integral matching on the building shadows in the unmatched sets according to the same method to obtain a final matching result, so that the shadow detection of the orthoimage with high resolution is completed.
8. The method for detecting the shadow of the multilayer integral relaxation-matched high-resolution ortho-image under the drive of the artificial shadow according to claim 1, wherein the artificial shadow is characterized in that: the sun altitude and azimuth during aerial image shooting are combined with a Digital Building Model (DBM) of the image, a shadow modeling algorithm is utilized to calculate the geometric range of a shadow area on the ground plane caused by blocking of solar rays by a building at a certain aerial image shooting moment, the geometric area is the required shadow area, and the gray value of the geometric area is assigned to be 0.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911195527.3A CN110992330B (en) | 2019-11-28 | 2019-11-28 | High-resolution orthophoto shadow detection based on multi-level global relaxation matching driven by artificial shadows |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911195527.3A CN110992330B (en) | 2019-11-28 | 2019-11-28 | High-resolution orthophoto shadow detection based on multi-level global relaxation matching driven by artificial shadows |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110992330A true CN110992330A (en) | 2020-04-10 |
| CN110992330B CN110992330B (en) | 2021-05-25 |
Family
ID=70088025
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911195527.3A Active CN110992330B (en) | 2019-11-28 | 2019-11-28 | High-resolution orthophoto shadow detection based on multi-level global relaxation matching driven by artificial shadows |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110992330B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113222918A (en) * | 2021-04-29 | 2021-08-06 | 桂林理工大学 | Method for detecting high-resolution orthographic image shadow by polygonal semantic matching |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080226145A1 (en) * | 2007-03-05 | 2008-09-18 | Fujifilm Corporation | Image processing apparatus and computer readable media containing image processing program |
| CN101604018A (en) * | 2009-07-24 | 2009-12-16 | 中国测绘科学研究院 | High resolution remote sensing image data processing method and system |
| US9514381B1 (en) * | 2013-03-15 | 2016-12-06 | Pandoodle Corporation | Method of identifying and replacing an object or area in a digital image with another object or area |
| CN107564017A (en) * | 2017-08-29 | 2018-01-09 | 南京信息工程大学 | A kind of city high score remote sensing image shadow Detection and dividing method |
| CN108629742A (en) * | 2018-03-29 | 2018-10-09 | 深圳飞马机器人科技有限公司 | True orthophoto shadow Detection and compensation method, device and storage medium |
| CN110298211A (en) * | 2018-03-21 | 2019-10-01 | 北京大学 | A kind of Methods Deriving Drainage Network based on deep learning and high-resolution remote sensing image |
| CN110309780A (en) * | 2019-07-01 | 2019-10-08 | 中国科学院遥感与数字地球研究所 | Rapid Supervision and Recognition of House Information in High Resolution Images Based on BFD-IGA-SVM Model |
-
2019
- 2019-11-28 CN CN201911195527.3A patent/CN110992330B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080226145A1 (en) * | 2007-03-05 | 2008-09-18 | Fujifilm Corporation | Image processing apparatus and computer readable media containing image processing program |
| CN101604018A (en) * | 2009-07-24 | 2009-12-16 | 中国测绘科学研究院 | High resolution remote sensing image data processing method and system |
| US9514381B1 (en) * | 2013-03-15 | 2016-12-06 | Pandoodle Corporation | Method of identifying and replacing an object or area in a digital image with another object or area |
| CN107564017A (en) * | 2017-08-29 | 2018-01-09 | 南京信息工程大学 | A kind of city high score remote sensing image shadow Detection and dividing method |
| CN110298211A (en) * | 2018-03-21 | 2019-10-01 | 北京大学 | A kind of Methods Deriving Drainage Network based on deep learning and high-resolution remote sensing image |
| CN108629742A (en) * | 2018-03-29 | 2018-10-09 | 深圳飞马机器人科技有限公司 | True orthophoto shadow Detection and compensation method, device and storage medium |
| CN110309780A (en) * | 2019-07-01 | 2019-10-08 | 中国科学院遥感与数字地球研究所 | Rapid Supervision and Recognition of House Information in High Resolution Images Based on BFD-IGA-SVM Model |
Non-Patent Citations (2)
| Title |
|---|
| RONGJUN QIN ET AL: "Object-Based 3-D Building Change Detection on Multitemporal Stereo Images", 《IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING》 * |
| 李健等: "基于影像特征点、线整体匹配的DSM自动生成", 《测绘科学》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113222918A (en) * | 2021-04-29 | 2021-08-06 | 桂林理工大学 | Method for detecting high-resolution orthographic image shadow by polygonal semantic matching |
| CN113222918B (en) * | 2021-04-29 | 2022-07-01 | 桂林理工大学 | Method for detecting high-resolution orthographic image shadow through polygonal semantic matching |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110992330B (en) | 2021-05-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhou et al. | Seamless fusion of LiDAR and aerial imagery for building extraction | |
| CN110287898B (en) | Optical satellite remote sensing image cloud detection method | |
| US11804025B2 (en) | Methods and systems for identifying topographic features | |
| KR101549155B1 (en) | Method of automatic extraction of building boundary from lidar data | |
| AU2020103470A4 (en) | Shadow Detection for High-resolution Orthorectificed Imagery through Multi-level Integral Relaxation Matching Driven by Artificial Shadows | |
| Zakharov et al. | Automatic building detection from satellite images using spectral graph theory | |
| CN116805356A (en) | Building model construction method, building model construction equipment and computer readable storage medium | |
| CN103971377A (en) | Building extraction method based on prior shape level set segmentation | |
| CN115909099A (en) | Side slope dangerous rock identification and monitoring method based on unmanned aerial vehicle inspection | |
| Wang | Automatic extraction of building outline from high resolution aerial imagery | |
| CN117876984A (en) | Road sign detection and recognition method based on MLS point cloud | |
| CN112734816A (en) | Heterogeneous image registration method based on CSS-Delaunay | |
| CN107122782B (en) | Balanced semi-dense stereo matching method | |
| CN110992330B (en) | High-resolution orthophoto shadow detection based on multi-level global relaxation matching driven by artificial shadows | |
| Shahzad et al. | Reconstructing 2-D/3-D building shapes from spaceborne tomographic synthetic aperture radar data | |
| Omidalizarandi et al. | Segmentation and classification of point clouds from dense aerial image matching | |
| Bulatov et al. | On Applications of Sequential Multi-view Dense Reconstruction from Aerial Images. | |
| Novacheva | Building roof reconstruction from LiDAR data and aerial images through plane extraction and colour edge detection | |
| Di et al. | Multi-scale segmentation of high resolution remote sensing images by integrating multiple features | |
| CN113705582B (en) | Method for extracting edge feature key points of building facade | |
| CN115880325A (en) | Building outline automatic extraction method based on point cloud dimension and spatial distance clustering | |
| Ferreira et al. | Acquisition of 3D regular prismatic models in urban environments from DSM and orthoimages | |
| Haefner et al. | Photometric segmentation: Simultaneous photometric stereo and masking | |
| Li et al. | A high-resolution satellite DEM filtering method assisted with building segmentation | |
| Chen et al. | Extraction of orthogonal building boundary from airborne lidar data based on feature dimension reduction |
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 |