CN113096168A - Optical remote sensing image registration method and system combining SIFT points and control line pairs - Google Patents
Optical remote sensing image registration method and system combining SIFT points and control line pairs Download PDFInfo
- Publication number
- CN113096168A CN113096168A CN202110287621.2A CN202110287621A CN113096168A CN 113096168 A CN113096168 A CN 113096168A CN 202110287621 A CN202110287621 A CN 202110287621A CN 113096168 A CN113096168 A CN 113096168A
- Authority
- CN
- China
- Prior art keywords
- image
- straight line
- registration
- matching
- feature set
- 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
- 238000000034 method Methods 0.000 title claims abstract description 62
- 230000003287 optical effect Effects 0.000 title claims abstract description 30
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 55
- 230000009466 transformation Effects 0.000 claims description 16
- 230000000007 visual effect Effects 0.000 claims description 16
- 239000011159 matrix material Substances 0.000 claims description 12
- 238000012216 screening Methods 0.000 claims description 10
- 230000002457 bidirectional effect Effects 0.000 claims description 9
- 238000013519 translation Methods 0.000 claims description 6
- 238000011524 similarity measure Methods 0.000 claims description 4
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 claims description 3
- 238000005314 correlation function Methods 0.000 claims description 3
- 238000011426 transformation method Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 description 38
- 238000010586 diagram Methods 0.000 description 14
- 238000004590 computer program Methods 0.000 description 7
- 230000006872 improvement Effects 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 230000000694 effects Effects 0.000 description 6
- 230000004927 fusion Effects 0.000 description 6
- 230000006870 function Effects 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 238000000605 extraction Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 230000000052 comparative effect Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 235000017166 Bambusa arundinacea Nutrition 0.000 description 1
- 235000017491 Bambusa tulda Nutrition 0.000 description 1
- 241001330002 Bambuseae Species 0.000 description 1
- 241000282326 Felis catus Species 0.000 description 1
- 235000015334 Phyllostachys viridis Nutrition 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 239000011425 bamboo Substances 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000019771 cognition Effects 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012549 training Methods 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
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20048—Transform domain processing
- G06T2207/20061—Hough transform
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an optical remote sensing image registration method and system combining SIFT points and control line pairs, wherein the method comprises the following steps: respectively acquiring linear feature sets of a reference image and an image to be registered; acquiring an initial registration image based on the linear feature set of the reference image and the linear feature set of the image to be registered; acquiring a linear feature set of the initial registration image; respectively counting the intersecting straight line pairs in the straight line feature set of the reference image and the straight line feature set of the initial registration image to obtain a statistical result, and obtaining a candidate matching set based on the statistical result; obtaining a coarse registration result based on the candidate matching set; and obtaining an accurate registration result of the reference image and the coarse registration image by utilizing an SIFT key point registration algorithm, and finishing the registration of the optical remote sensing image. The method is high in speed and good in robustness, and can realize registration of the visible light remote sensing image with the typical linear target.
Description
Technical Field
The invention belongs to the technical field of aerospace and aviation, relates to the field of optical remote sensing image registration, and particularly relates to an optical remote sensing image registration method and system combining SIFT points and control line pairs.
Background
In aerospace research, abundant ground remote sensing data can be obtained by using aerospace vehicles with different heights. The remote sensing data is widely applied to a plurality of fields such as battlefield detection, aircraft navigation, unmanned aerial vehicle landing, horizon detection and the like with the advantages of timeliness and practicability. The rapid development of the satellite remote sensing technology enables the amount of space remote sensing data to increase rapidly, and how to utilize massive space remote sensing data to meet different production and living requirements becomes one of the outstanding problems in remote sensing application. The information fusion can comprehensively utilize different remote sensing data of a single sensor or a plurality of sensors, reduce or inhibit the possible ambiguity, incompleteness, uncertainty and error of a single signal source in the sensed object or environment interpretation through the organic integration of complementary information, greatly improve the application efficiency of the remote sensing data, and improve the reliability in the aspects of feature extraction, classification, target identification and the like.
The image fusion is one of important research directions of information fusion, and two or more images obtained by imaging the same scene are cooperatively utilized to realize information complementation, so that more detailed, comprehensive and reliable description of the scene is obtained, and the cognition and subsequent interpretation processing are facilitated. The image registration is used as the necessary basic work of image fusion, and the geometric inconsistency of the multi-source images is eliminated through a space conversion mechanism, so that the foundation is laid for subsequent fusion application such as target identification and classification. Image registration algorithms currently fall broadly into two broad categories: a method of image registration based on gray scale information and a method of image registration based on features.
The registration method based on the gray scale information mainly utilizes the gray scale statistical information of the image to search the region with the closest similarity between the reference image and the image to be registered so as to determine the optimal geometric transformation parameters. Kern JP provides a registration algorithm for multispectral remote sensing images based on a mutual information method, and the model can control and optimize the selection of image transformation parameters by associating the curved surface shape of mutual information with the frequency domain characteristics of the images. The plum is beautiful and combines the cat swarm algorithm and the normalized cross-correlation matching algorithm, so that the sensitivity of the original normalized cross-correlation matching algorithm to image rotation and scaling transformation is reduced, and the accuracy of image registration is effectively improved. Generally, the registration algorithm based on the gray information directly utilizes the image gray information, so that the robustness of the registration method is improved, but the gray relevance is low due to the difference of image data acquisition environments, and the method is difficult to obtain a satisfactory effect.
The characteristic-based image registration algorithm mainly utilizes salient characteristics of points, lines, edges or areas and the like of image targets to establish a mapping relation between images and overcome the defect of sensitivity of a gray information algorithm. Among the point-Feature-based image registration algorithms, Scale-invariant Feature Transform (SIFT), Speeded Up Robust Features (SURF), and their improved algorithms are most widely used. The Lowe D.G. perfects a classic algorithm-SIFT algorithm based on image features, the algorithm extracts interest points which are kept unchanged in the image transformation process as feature description by comparing extreme points of the image in different scale spaces, and finally image thought measurement is carried out on the determined local features to realize accurate registration of the image. The strong stability and high efficiency of the SIFT algorithm play a crucial role in the field of image registration. Bay H. et al propose a new stable feature acceleration algorithm, SURF algorithm, which combines the advantages of hessian matrix feature detection and distribution-based feature descriptors, and significantly improves the matching accuracy and speed of images. In order to further reduce the event complexity of the SIFT algorithm, the von Jia changes the rectangular region of the key points into the concentric circle region, so that the Gaussian blur process of the SIFT algorithm is simplified. Shiya bamboo proposes an improved SURF algorithm that confines the sign points to a certain area of the image edge and secondarily constrains the registration points. The improved algorithm does not improve the precision of image registration to the utmost extent, and has stronger robustness to image training with strong noise. The image similarity measurement function with excellent performance is determined by the Chengdlog and the like according to the feature vector generated by the SIFT algorithm, so that the feature matching accuracy of the SIFT algorithm is effectively improved, the registration time is greatly shortened, and a new method is provided for real-time engineering tasks. And the like apply the improved SURF algorithm to the registration process of the remote sensing images, and improve the accuracy of the SURF algorithm in the registration of the remote sensing images by iteratively removing mismatching points existing in the SURF algorithm. However, the improved SURF algorithm is slow in registration. Yanli et al, in combination with Haar-like features and SURF features, provide a new feature descriptor for registration of remote sensing images, the new feature descriptor increases the description of features, and multi-level feature extraction achieves a more accurate registration result.
The line features have higher-level semantic information than the point features, and parameters such as length and direction of the line features can be used for establishing mathematical constraint or similarity measurement when the matching relation is calculated. Therefore, a large number of image matching algorithms based on linear features have been proposed in recent years. Stamos I. et al propose to use the slope of the characteristic line and the distribution of the edge profile to obtain the registration parameters, but when there are multiple lines with approximate slopes, it is not easy to make one-to-one correspondence, and it is difficult to avoid generating mismatching. Kim Y.S. et al, which uses the point feature and edge feature direction information for registration, but both the original remote sensing image and the visible light image adopt Canny operators, and do not consider the characteristic of the remote sensing image that the remote sensing image is susceptible to noise. Sujuan and the like construct control points by utilizing image linear features and design a function based on the matching degree of the control points, realize automatic registration from coarse to fine, obtain good effect, but are long in time consumption and difficult to be applied to application with strong real-time property. Li Ying et al adopt three straight line edges as reference models, construct candidate matching sets under certain constraint conditions, and determine the corresponding relationship between the straight line edges on the basis, and experimental results show that the operation time is short, but the selection of the constraint conditions is complex, a reference with universal applicability is not obtained, and when the straight line edges are similar in position, direction and length, mismatching easily occurs, and more constraint conditions need to be introduced.
In summary, most feature-based image registration algorithms at the present stage adopt a single feature, and although the running time of the algorithms is shortened, the robustness of the algorithms is not good for remote sensing images with more noise.
Disclosure of Invention
The invention aims to provide an optical remote sensing image registration method and system combining SIFT points and control line pairs, so as to solve one or more technical problems. The method is high in speed and good in robustness, and can realize registration of the visible light remote sensing image with the typical linear target.
In order to achieve the purpose, the invention adopts the following technical scheme:
the invention discloses an optical remote sensing image registration method combining SIFT points and control line pairs, which comprises the following steps:
respectively acquiring linear feature sets of a reference image and an image to be registered;
acquiring an initial registration image based on the linear feature set of the reference image and the linear feature set of the image to be registered; acquiring a linear feature set of the initial registration image;
respectively counting the intersecting straight line pairs in the straight line feature set of the reference image and the straight line feature set of the initial registration image to obtain a statistical result, and obtaining a candidate matching set based on the statistical result;
obtaining a coarse registration result based on the candidate matching set;
and obtaining an accurate registration result of the reference image and the coarse registration image by utilizing an SIFT key point registration algorithm, and finishing the registration of the optical remote sensing image.
The further improvement of the present invention is that the step of respectively obtaining the linear feature sets of the reference image and the image to be registered specifically includes:
extracting and obtaining a reference image I by using an LSD algorithm1And image I to be registered2Set of alternative straight line segments L in1And L2；
According to visual saliency theory, collecting L for alternative straight line segments1And L2Screening to obtain effective linear characteristic set C1And C2；
Connecting effective straight line feature set C by Hough transformation1And C2Obtaining a straight line feature set Q by using a broken long straight line1And Q2。
In a further development of the invention, the method comprisesAccording to visual saliency theory, collecting L for alternative straight line segments1And L2Screening to obtain effective linear characteristic set C1And C2(ii) a Connecting effective straight line feature set C by Hough transformation1And C2Obtaining a straight line feature set Q by using a broken long straight line1And Q2The method specifically comprises the following steps:
judging a set L of alternative straight line segments1Or L2Whether the difference between the slope of a certain line segment C and the slope of the assumed straight line meets the significance criterion represented by the formula (1) or not is judged to be an available characteristic straight line segment if the difference meets the significance criterion, the visual significance characteristic value of the straight line is determined to be length, and the effective straight line characteristic set C of the reference image and the image to be registered is obtained1And C2；
Effective straight line feature set C by using Hough transformation method1And C2The straight line segments with the same slope are mapped to the peak point of another coordinate space, and the straight line feature set Q is obtained by statistically connecting the feature straight line segments of the peak points1And Q2。
The invention has the further improvement that the initial registration image is obtained based on the linear feature set of the reference image and the linear feature set of the image to be registered; the step of obtaining the linear feature set of the initial registration image specifically includes:
from a set of straight line features Q1And Q2Obtaining an initial registration image I by using an initial registration method based on feature consistency3Extracting and obtaining an initial registration image I by using an LSD algorithm3Set of alternative straight line segments L in3According to the visual saliency theory, a set L of alternative straight line segments3Screening to obtain effective linear characteristic set C3Connecting effective linear feature set C by Hough transformation3Obtaining a straight line feature set Q by using a broken long straight line3。
Further aspects of the inventionThe improvement is that the set Q of features from straight lines1And Q2Obtaining an initial registration image I by using an initial registration method based on feature consistency3The method specifically comprises the following steps:
the initial registration method based on feature consistency utilizes a four-parameter affine transformation model which consists of a scaling factor s, a rotation angle beta and a translation parameter txAnd tyAccording to this transformation model, the point p ═ x, y in the reference image and its corresponding point in the image to be registered
The following relationship is satisfied:
wherein the scaling factor s is obtained from the resolution of the image; suppose that
And
is a straight line with the same name,
and
included angles with the horizontal axis are respectively
And
then satisfy
Scaling and rotating transformation are carried out on the image to be registered according to the scaling factor s and the rotating angle beta, the cross-correlation function of the reference image and the image to be registered is calculated, and the x-axis coordinate and the y-axis coordinate corresponding to the peak value are translation distance txAnd ty。
The further improvement of the present invention is that the step of respectively counting the intersecting straight line pairs in the straight line feature set of the reference image and the straight line feature set of the initial registration image to obtain a statistical result, and the step of obtaining a candidate matching set based on the statistical result specifically includes:
respectively counting a linear feature set Q1And Q3The intersecting straight line pairs in (1) are respectively stored in sets LB and LW; and if the absolute value of the difference between the included angle of a certain orthogonal straight line pair LB in the LB and the included angle of a certain orthogonal straight line pair LW in the LW is smaller than a preset threshold value, the straight line pair LB and the straight line pair LW are considered to be same-name straight line pairs, and all the same-name straight line pairs meeting the relation form a candidate matching set G.
In a further development of the invention, the separately statistically linear feature sets Q1And Q3The specific step of storing the intersecting straight line pairs in the sets LB and LW respectively comprises the following steps:
to straight line feature set Q1And Q3And the line segments with the middle included angle more than or equal to 30 degrees form a straight line pair, and an intersection point is generated and recorded as:
LB＝{LBij|0＜i,j≤N,i≠j}，LW＝{LWij|0＜i,j≤N,i≠j}；
wherein, LBij＝{(li,lj,Pij,θij)|0＜i,j≤N,i≠j}；LWij＝{(li,lj,Pij,θij)|0＜i,j≤N,i≠j}；
In the formula, N represents the number of straight line features in the image; liAnd ljAre two straight lines that make up a line pair; pijIs the intersection point of the straight line pair; thetaijRepresenting the included angle between the straight line pairs;
if the absolute value of the difference between the included angle of a certain intersecting straight line pair LB in the LB and the included angle of a certain intersecting straight line pair LW in the LW is smaller than the preset threshold, the straight line pair LB and the straight line pair LW are considered to be the same-name straight line pair, and the step of forming the candidate matching set G by all the same-name straight line pairs satisfying the relationship specifically comprises the following steps:
suppose that
And
is a reference image I1The center point is crossed on a point p straight line pair, and the included angle is thetab；
And
is an initial registered image I3The center is crossed with a straight line pair at a point p' and the included angle is thetaw；
Constructing the following similarity measure criteria, if the straight line pair
And a straight line pair
If the following formula is satisfied, the points p and p' are considered as matching control point pairs, and the expression is:
dθ(lb,lw)＝|θb-θw|,0＜dθ(lb,lw)＜dθmax，
wherein,
representing the angular relationship between the pairs of straight lines, dθmaxTo characterize the straight line pairA threshold value of angular relationship;
the pairs of straight lines satisfying the condition constitute a candidate matching set, which is expressed as:
wherein,
and piThe straight line pairs and the intersection points thereof in the reference image;
and pi' are pairs of lines in the image to be registered and their intersections.
A further improvement of the present invention is that the step of obtaining a coarse registration result based on the candidate matching set specifically comprises:
defining homonymy line segment matching matrixes MB and MW for representing the matching relation of homonymy line pairs; constructing a bidirectional matching relation according to the homonymous line segment matching matrixes MB and MW, and removing repeated matching to obtain a one-to-one matching point set; obtaining a reference image I by a least square method1And initial registration image I3The optimal transformation matrix is used, and then the coarse registration result I is obtained by utilizing the reverse interpolation4。
The invention is further improved in that the defined homonymous line segment matching matrixes MB and MW are used for representing the matching relation of homonymous straight line pairs; constructing a bidirectional matching relationship according to the homonymous line segment matching matrixes MB and MW, and removing repeated matching to obtain a one-to-one matching point set specifically comprises the following steps:
suppose there is N in the candidate matching set G1The lines of the bars are from the reference image, N2The straight lines come from the image to be registered; mb if the ith element in the combined image contains the jth line in the reference image ij1 is ═ 1; otherwise, mbij0; mw if the ith element in G contains the jth line in the image to be registeredij1 is ═ 1; otherwise, mwij＝0；
The matrix MB represents the one-way matching of the image to be registered to the reference image; the matrix MW represents the one-way matching of the reference image to the image to be registered;
and counting the same rows in the matrixes MB and MW, wherein the same rows represent repeated matching, calculating the distance between the same straight lines corresponding to the same rows, and only keeping the straight line pair with the shortest distance.
The invention discloses an optical remote sensing image registration method combining SIFT points and control line pairs, which comprises the following steps:
the linear feature set acquisition module is used for respectively acquiring linear feature sets of the reference image and the image to be registered;
the linear feature set acquisition module of the initial registration image is used for acquiring the initial registration image according to the linear feature set of the reference image and the linear feature set of the image to be registered; acquiring a linear feature set of the initial registration image;
the candidate matching set acquisition module is used for respectively counting the intersecting straight line pairs in the straight line feature set of the reference image and the straight line feature set of the initial registration image to obtain a statistical result, and acquiring a candidate matching set based on the statistical result;
a coarse registration result obtaining module, configured to obtain a coarse registration result according to the candidate matching set;
and the accurate registration result acquisition module is used for obtaining an accurate registration result of the reference image and the coarse registration image by utilizing an SIFT key point registration algorithm to complete the registration of the optical remote sensing image.
Compared with the prior art, the invention has the following beneficial effects:
the invention provides an image registration algorithm based on feature fusion, which can realize advantage complementation and overcome the defect of low robustness of the traditional registration method. Specifically, the invention provides an optical remote sensing image registration algorithm combining SIFT points and control Line pairs, which not only optimizes a Line Segment Detection algorithm (LSD) by using a visual saliency theory to accurately position Line features and reduce additional Line interference in a coarse registration process, but also extracts point features with rotation and scale invariance by using the SIFT algorithm to realize an accurate registration process. The invention provides a coarse-to-fine matching method, which expands the idea of image matching and improves the matching precision.
Optical remote sensing images with typical linear objects (including airports, buildings, bridges and roads) all contain a large number of linear features, so that efficient and accurate linear feature extraction is the key for registration of the optical remote sensing images. The LSD algorithm can realize line segment detection with sub-pixel level precision, can effectively extract linear characteristics in the image, but directly utilizes the LSD algorithm to detect the phenomena of breakage and discontinuity of the image due to the influence of noise, contrast and brightness, so that the invention combines a visual saliency model and Hough change to remove the interference of miscellaneous line segments, obtains long straight lines, reduces the matching control line pairs and improves the matching efficiency.
The line features are ideal matching primitives, but in the actual extraction and matching process, the obtained line features are often broken, and accurate image registration is difficult to realize by directly utilizing the line features. But intersecting lines tend to have more stable geometric properties, with the most common and invariant properties being the directional characteristics of the line. Therefore, the invention searches the intersection point of the straight line pairs with the same name which are possibly matched as the registration control point from the direction information between the straight line pairs, and realizes more accurate registration.
In the invention, the homonymy straight line pairs of the reference image and the image to be registered are obtained to form a candidate matching set, and a bidirectional matching relation is established by utilizing a homonymy line segment matching matrix to remove error matching.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art are briefly introduced below; it is obvious that the drawings in the following description are some embodiments of the invention, and that for a person skilled in the art, other drawings can be derived from them without inventive effort.
FIG. 1 is a schematic flow chart of a visible light remote sensing image registration method combining SIFT points and control line pairs according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of distances between homonymous straight line pairs in an embodiment of the present invention.
FIG. 3 is a schematic diagram showing comparison of a first set of effects in an experimental chart, wherein the image is a visible light remote sensing image with a size of 993 × 993; wherein (a) in fig. 3 is a reference image, (b) in fig. 3 is an image to be registered, (c) in fig. 3 is an initial registration image of (b) in fig. 3, (d) in fig. 3 is an LSD straight-line segment detection map (LSD _ a) in fig. 3 (a), fig. 3 (e) is a Hough transform detection control line map (HF _ d) in fig. 3 (d), fig. 3 (f) is an LSD straight-line segment detection map (VSLSD _ a) in fig. 3 (a) improved based on the present invention, (g) in fig. 3 is a Hough transform detection control line map (HF _ f) in fig. 3 (f), fig. 3 (h) is an LSD straight-line segment detection map (LSD _ c) in fig. 3 (c), fig. 3 (i) is an Hough transform detection control line map (HF _ h) in fig. 3 (h), fig. 3 (j) is an improved straight-line segment detection map (vsd _ a) in fig. 3 (vsd _ c), fig. 3 (k) is the Hough transform detection control line graph (HF _ j) of fig. 3 (j), fig. 3 (l) is a checkerboard graph showing SIFT registration result, and fig. 3 (m) is a checkerboard graph showing inventive registration result;
FIG. 4 is a schematic diagram of comparison of effects of a second group of experimental diagrams, where the images are visible light remote sensing images of 220 × 291 size; wherein (a) in fig. 4 is a reference image, (b) in fig. 4 is an image to be registered, (c) in fig. 4 is an initial registration image of (b) in fig. 4, (d) in fig. 4 is (a) an LSD straight-line segment detection map (LSD _ a) in fig. 4, (e) in fig. 4 is a Hough transform detection control line map (HF _ d) in fig. 4, (f) in fig. 4 is (a) an LSD straight-line segment detection map (VSLSD _ a) in fig. 4 improved based on the present invention, (g) in fig. 4 is a Hough transform detection control line map (HF _ f) in fig. 4, (h) in fig. 4 is (c) an LSD straight-line segment detection map (LSD _ c) in fig. 4, (i) in fig. 4 is a Hough transform detection control line map (HF _ h) in fig. 4, and (j) in fig. 4 is (vsd _ c) an improved straight-line segment detection map (vsd _ a) in fig. 4, fig. 4 (k) is the Hough transform detection control line graph (HF _ j) of fig. 4 (j), fig. 4 (l) is a checkerboard graph showing SIFT registration result, and fig. 4 (m) is a checkerboard graph showing inventive registration result;
FIG. 5 is a schematic diagram showing comparison of effects of the third group of experimental diagrams in the embodiment of the present invention, where the image is a 196 × 491 size visible light remote sensing image; wherein (a) in fig. 5 is a reference image, (b) in fig. 5 is an image to be registered, (c) in fig. 5 is an initial registration image of (b) in fig. 5, (d) in fig. 5 is (a) an LSD straight-line segment detection map (LSD _ a) in fig. 5, (e) in fig. 5 is a Hough transform detection control line map (HF _ d) in fig. 5, (f) in fig. 5 is (a) an LSD straight-line segment detection map (VSLSD _ a) in fig. 5 improved based on the present invention, (g) in fig. 5 is a Hough transform detection control line map (HF _ f) in fig. 5, (h) in fig. 5 is (c) an LSD detection map (LSD _ c) in fig. 5, (i) in fig. 5 is a Hough transform detection control line map (HF _ h) in fig. 5, (h), and (vsd _ j) in fig. 5 is (vsd _ c) an improved straight-line segment detection map (vsd _ a) in fig. 5, fig. 5 (k) is the Hough transform detection control line graph (HF _ j) of fig. 5 (j), fig. 5 (l) is a checkerboard graph showing SIFT registration result, and fig. 5 (m) is a checkerboard graph showing inventive registration result.
Detailed Description
In order to make the purpose, technical effect and technical solution of the embodiments of the present invention clearer, the following clearly and completely describes the technical solution of the embodiments of the present invention with reference to the drawings in the embodiments of the present invention; it is to be understood that the described embodiments are only some of the embodiments of the present invention. Other embodiments, which can be derived by one of ordinary skill in the art from the disclosed embodiments without inventive faculty, are intended to be within the scope of the invention.
Referring to fig. 1, a visible light remote sensing image registration method combining SIFT points and control line pairs according to an embodiment of the present invention includes the following steps:
step 1: extracting reference image I by using LSD algorithm1And image I to be registered2Set of alternative straight line segments L in1And L2And outputting the alternative straight line segment set L to the LSD according to the visual saliency theory1And L2Screening, eliminating the interference of miscellaneous lines, and extracting effective linear characteristic set C1And C2Connecting the screened linear characteristic set C by Hough transformation1And C2Obtaining a corresponding linear feature set Q by using the long straight line of the medium fracture1And Q2。
Step 2: from a set of straight line features Q1And Q2Obtaining an initial registration image I by using an initial registration method based on feature consistency3Extracting an initial registration image I by using an LSD algorithm3Set of alternative straight line segments L in3And outputting the alternative straight line segment set L to the LSD according to the visual saliency theory3Screening, eliminating the interference of miscellaneous lines, and extracting the effective linear characteristic set C3Connecting the screened linear characteristic set C by Hough transformation3Obtaining a corresponding linear feature set Q by using the long straight line of the medium fracture3。
And step 3: statistical linear feature set Q1And Q3The intersecting straight line pairs in (1) are stored in sets LB and LW, respectively. And if the absolute value of the difference between the included angle of a certain orthogonal straight line pair LB in the LB and the included angle of a certain orthogonal straight line pair LW in the LW is smaller than the threshold value, the straight line pair LB and the straight line pair LW are considered to be the same-name straight line pair, and all the same-name straight line pairs meeting the relationship form a candidate matching set G.
And 4, step 4: if the candidate matching set G is composed only by using the angle relationship between the same-name straight line pairs, it is difficult to satisfy the "one-to-one" matching relationship, and there may also be mismatching. In order to improve matching precision and remove repeated matching, the method utilizes the matching matrix of the same-name line segments to establish a bidirectional matching relation, and realizes accurate matching between the same-name line segments. And identical line segment matching matrixes MB and MW are defined and used for representing the matching relation of identical straight line pairs. Constructing a bidirectional matching relation according to the matching matrixes MB and MW of the homonymous line segments, and removing repeated matchingAnd obtaining a one-to-one matching point set. Obtaining a reference image I by a least square method1And initial registration image I3The optimal transformation matrix is used, and then the coarse registration result I is obtained by utilizing the reverse interpolation4。
And 5: obtaining a reference image I by utilizing an SIFT key point registration algorithm1And coarse registration of image I4Accurate registration result I5。
In step 1 of the embodiment of the invention, an alternative straight-line segment set L output to the LSD according to the visual saliency theory1And L2Screening, eliminating the interference of miscellaneous lines, and extracting effective linear characteristic set C1And C2Connecting the screened linear characteristic set C by Hough transformation1And C2Obtaining a corresponding linear feature set Q by using the long straight line of the medium fracture1And Q2The method specifically comprises the following steps: judging alternative straight line segment set L output by LSD1Or L2If the difference between the slope of a certain line segment C and the slope of the assumed straight line meets the significance criterion represented by the following formula, if the difference meets the significance criterion, the length of the line segment C is at least length, the line segment C meets the visual characteristics of human eyes, the line segment C is judged to be an available characteristic straight line segment, the visual significance characteristic value of the straight line is determined to be length, and the straight line characteristic set C of the reference image and the image to be registered is obtained1And C2。
Linear feature set C by using Hough transformation method1，C2The straight line segments with the same slope are mapped to the peak point of another coordinate space, and the registration control line set Q is obtained by counting the peak points and connecting the characteristic straight line segments1And Q2。
In step 2 of the embodiment of the invention, a set Q is set according to the linear characteristics1And Q2Obtaining an initial registration image I by using an initial registration method based on feature consistency3The method specifically comprises the following steps: initial registration method based on feature consistency by utilizing four-parameter affine transformationThe model is composed of a scaling factor s, a rotation angle beta and a translation parameter txAnd tyAccording to this transformation model, the point p ═ x, y in the reference image and its corresponding point in the image to be registered
The following relationship is satisfied:
wherein,
in general, the scaling factor s can be directly determined from the resolution of the image. Suppose that
And
is a straight line with the same name,
and
included angles with the horizontal axis are respectively
And
then satisfy
Scaling and rotating transformation are carried out on the image to be registered according to the scaling factor s and the rotating angle beta, then the cross-correlation function of the reference image and the image to be registered is calculated, and the x-axis coordinate and the y-axis coordinate corresponding to the peak value are the translation distance txAnd ty。
In step 3 of the embodiment of the invention, a linear feature set Q is counted1And Q3The storing of the intersecting straight line pairs in the sets LB and LW respectively specifically includes: using a reference image I1And initial registration image I3Obtaining a linear feature set Q1And Q3For the linear feature set Q1And Q3And the line segments with the middle included angle not less than 30 degrees form a straight line pair, and an intersection point is generated and recorded as:
LB＝{LBij|0＜i,j≤N,i≠j}，LW＝{LWij|0＜i,j≤N,i≠j}
wherein, LBij＝{(li,lj,Pij,θij)|0＜i,j≤N,i≠j}；LWij＝{(li,lj,Pij,θij) I0 is more than i, j is less than or equal to N, i is not equal to j }; n represents the number of straight line features in the image; liAnd ljAre two straight lines that make up a line pair; pijIs the intersection point of the straight line pair; thetaijRepresenting the angle between the pair of straight lines.
In step 3 of the embodiment of the present invention, if an absolute value of a difference between an included angle of a certain orthogonal straight line pair LB in the LB and an included angle of a certain orthogonal straight line pair LW in the LW is smaller than a threshold, the straight line pair LB and the straight line pair LW are considered to be a same-name straight line pair, and all the same-name straight line pairs satisfying the above relationship form the candidate matching set G specifically including: suppose that
And
is a reference image I1The center point is crossed on a point p straight line pair, and the included angle is thetab；
And
is an initial registered image I3The center is crossed with a straight line pair at a point p' and the included angle is thetaw. Constructing the following similarity measure criteria, if the straight line pair
And a straight line pair
Points p and p' are considered to be a pair of matching control points if the following equation is satisfied, namely:
dθ(lb,lw)＝|θb-θw|,0＜dθ(lb,lw)＜dθmax
wherein,
representing the angular relationship between the pairs of straight lines, dθmaxIs a threshold value characterizing the straight-line diagonal relationship.
The pair of straight lines satisfying the above condition is made into a candidate matching set, which can be expressed as:
wherein,
and piThe straight line pairs and the intersection points thereof in the reference image;
and pi' are pairs of lines in the image to be registered and their intersections.
In step 3 of the embodiment of the present invention, generating a candidate matching set using a position relationship between the intersecting straight line pairs specifically includes: if I1And I2Individual watchShowing the reference image and the image to be registered,
and
is an image I1The center point is crossed on a point p straight line pair, and the included angle is thetab；
And
is an image I2The center is crossed with a straight line pair at a point p' and the included angle is thetaw. Constructing the following similarity measure criteria, if the straight line pair
And a straight line pair
Points p and p' are considered to be a pair of matching control points if the following equation is satisfied, namely:
dθ(lb,lw)＝|θb-θw|,0＜dθ(lb,lw)＜dθmax，
The pair of straight lines satisfying the above condition is made into a candidate matching set, which can be expressed as:
wherein,
and piThe straight line pairs and the intersection points thereof in the reference image;
and p'iThe straight line pairs and the intersection points thereof in the image to be registered.
In step 4 of the embodiment of the present invention, homonymy line segment matching matrices MB and MW are defined for representing the matching relationship of homonymy line pairs. Constructing a bidirectional matching relationship according to the homonymous line segment matching matrixes MB and MW, removing repeated matching, and obtaining a one-to-one matching point set specifically comprises: suppose there is N in the candidate matching set G1The lines of the bars are from the reference image, N2The bar lines come from the image to be registered. Mb if the ith element in the set G contains the jth line in the reference image ij1 is ═ 1; otherwise, mbij0. Similarly, mw is calculated if the ith element in the set G includes the jth line in the image to be registeredij1 is ═ 1; otherwise, mwij＝0。
The matrix MB represents the one-way matching of the image to be registered to the reference image; the matrix MW represents a one-way matching of the reference image to the image to be registered. And counting the same rows in the matrixes MB and MW, wherein the same rows represent repeated matching, calculating the distance between the same straight lines corresponding to the same rows, and only keeping the straight line pair with the shortest distance.
As shown in figure 2 of the drawings, in which,
and
is a pair of orthogonal lines in the reference image,
and
is a pair of AC and DC lines in the image to be registered, assuming
And
the line segments with the same name are selected,
and
being a line segment of the same name, d1To represent
And
distance between d2To represent
And
when d is1And d2The assumption is considered correct when all of the repetitions match the minimum distance between the same-name straight line pairs. Retention
And
deleting the remaining sums for correct matching of the straight line pairs
A pair of straight lines having a matching relationship. Storing the matched line pairs obtained after screening in the matrixes MB and MW in a set LPbAnd LPwIn (1), calculating LPbMatched line pair in (1) and corresponding LPwWhether the matched line pairs in (1) are consistent. If the two-way matching is consistent, the two-way matching is successful, and the matching line pair is reserved; otherwise, the bidirectional matching fails, and the matched line pair is deleted.
Referring to fig. 3 to 5, in the embodiment of the present invention, three groups of remote sensing images are selected for a simulation experiment, and are compared by an image registration algorithm based on SIFT key points. The experimental CPU is configured to be intel Core i7-8750H 2.20GHz, the experimental operating system is Windows 10, and the compiling environment is MATLAB2018 b. The experimental results are shown in the following comparative figures: graph (a) is a reference image, graph (b) is an image to be registered, graph (c) is an initial registration image of graph (b), graph (d) is a graph (a) LSD straight-line segment detection graph (LSD _ a), graph (e) is a Hough transform detection control line graph (HF _ d) of graph (d), graph (f) is a graph (a) LSD straight-line segment detection graph (VSLSD _ a) based on the improvement of the present invention, graph (g) is a Hough transform detection control line graph (HF _ f) of graph (f), graph (h) is a graph (c) LSD straight-line segment detection graph (LSD _ c), graph (i) is a Hough transform detection control line graph (HF _ h) of graph (h), graph (j) is a graph (c) based on the improvement of the present invention, graph (k) is a Hough transform detection control line graph (HF _ j) of graph (j), graph (sifl) is a checkerboard graph showing a registration result, graph (l) is a checkerboard graph showing the registration results of the present invention; from the above three sets of simulation comparative experiments, it can be seen that: the algorithm registration effect provided by the invention is superior to that of the traditional registration algorithm based on SIFT key points. The line segment detected only by using the LSD is easily influenced by noise, and the problems of incomplete identification line segment, false identification and the like are easily caused. Although the Hough transform detection method can overcome the influence of noise, the robustness is not strong due to the fact that the number of output line segments needs to be manually set, and the length of the identified line segment exceeds the length of an actual target line segment and is not easy to control. The invention provides a novel line segment detection algorithm based on an LSD algorithm, Hough transformation and a visual saliency model, the anti-noise capability of the LSD algorithm can be improved, the conditions of line segment breakage and the like are avoided, and the effectiveness of the algorithm is verified by the experimental results.
The embodiment of the invention provides an optical remote sensing image registration method combining SIFT points and control line pairs, which comprises the following steps:
the linear feature set acquisition module is used for respectively acquiring linear feature sets of the reference image and the image to be registered;
the linear feature set acquisition module of the initial registration image is used for acquiring the initial registration image according to the linear feature set of the reference image and the linear feature set of the image to be registered; acquiring a linear feature set of the initial registration image;
the candidate matching set acquisition module is used for respectively counting the intersecting straight line pairs in the straight line feature set of the reference image and the straight line feature set of the initial registration image to obtain a statistical result, and acquiring a candidate matching set based on the statistical result;
a coarse registration result obtaining module, configured to obtain a coarse registration result according to the candidate matching set;
and the accurate registration result acquisition module is used for obtaining an accurate registration result of the reference image and the coarse registration image by utilizing an SIFT key point registration algorithm to complete the registration of the optical remote sensing image.
The invention takes the optical remote sensing image containing typical linear objects (such as airports, buildings, bridges, roads and the like) as an object to carry out registration method research.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art can make modifications and equivalents to the embodiments of the present invention without departing from the spirit and scope of the present invention, which is set forth in the claims of the present application.
Claims (10)
1. An optical remote sensing image registration method combining SIFT points and control line pairs is characterized by comprising the following steps:
respectively acquiring linear feature sets of a reference image and an image to be registered;
acquiring an initial registration image based on the linear feature set of the reference image and the linear feature set of the image to be registered; acquiring a linear feature set of the initial registration image;
respectively counting the intersecting straight line pairs in the straight line feature set of the reference image and the straight line feature set of the initial registration image to obtain a statistical result, and obtaining a candidate matching set based on the statistical result;
obtaining a coarse registration result based on the candidate matching set;
and obtaining an accurate registration result of the reference image and the coarse registration image by utilizing an SIFT key point registration algorithm, and finishing the registration of the optical remote sensing image.
2. The optical remote sensing image registration method according to claim 1, wherein the step of respectively obtaining the linear feature sets of the reference image and the image to be registered specifically comprises:
extracting and obtaining a reference image I by using an LSD algorithm1And image I to be registered2Set of alternative straight line segments L in1And L2；
According to visual saliency theory, collecting L for alternative straight line segments1And L2Screening to obtain effective linear characteristic set C1And C2；
Connecting effective straight line feature set C by Hough transformation1And C2Obtaining a straight line feature set Q by using a broken long straight line1And Q2。
3. The method for registering optical remote sensing images as claimed in claim 2, wherein the set L of alternative straight line segments is set according to visual saliency theory1And L2Screening to obtain effective linear characteristic set C1And C2(ii) a Connecting effective straight line feature set C by Hough transformation1And C2Long straight line of medium fracture, obtainGet the linear feature set Q1And Q2The method specifically comprises the following steps:
judging a set L of alternative straight line segments1Or L2Whether the difference between the slope of a certain line segment C and the slope of the assumed straight line meets the significance criterion represented by the formula (1) or not is judged to be an available characteristic straight line segment if the difference meets the significance criterion, the visual significance characteristic value of the straight line is determined to be length, and the effective straight line characteristic set C of the reference image and the image to be registered is obtained1And C2；
Effective straight line feature set C by using Hough transformation method1And C2The straight line segments with the same slope are mapped to the peak point of another coordinate space, and the straight line feature set Q is obtained by statistically connecting the feature straight line segments of the peak points1And Q2。
4. The optical remote sensing image registration method according to claim 2, wherein an initial registration image is obtained based on the linear feature set of the reference image and the linear feature set of the image to be registered; the step of obtaining the linear feature set of the initial registration image specifically includes:
from a set of straight line features Q1And Q2Obtaining an initial registration image I by using an initial registration method based on feature consistency3Extracting and obtaining an initial registration image I by using an LSD algorithm3Set of alternative straight line segments L in3According to the visual saliency theory, a set L of alternative straight line segments3Screening to obtain effective linear characteristic set C3Connecting effective linear feature set C by Hough transformation3Obtaining a straight line feature set Q by using a broken long straight line3。
5. The optical remote sensing image registration method according to claim 4, wherein the set Q of features from straight lines is1And Q2Obtaining an initial registration image I by using an initial registration method based on feature consistency3The method specifically comprises the following steps:
the initial registration method based on feature consistency utilizes a four-parameter affine transformation model which consists of a scaling factor s, a rotation angle beta and a translation parameter txAnd tyAccording to this transformation model, the point p ═ x, y in the reference image and its corresponding point in the image to be registered
The following relationship is satisfied:
wherein the scaling factor s is obtained from the resolution of the image; suppose that
And
is a straight line with the same name,
and
included angles with the horizontal axis are respectively
And
then satisfy
Scaling and rotating transformation are carried out on the image to be registered according to the scaling factor s and the rotating angle beta, the cross-correlation function of the reference image and the image to be registered is calculated, and the x-axis coordinate and the y-axis coordinate corresponding to the peak value are translation distance txAnd ty。
6. The optical remote sensing image registration method according to claim 4, wherein the step of separately counting the intersecting straight line pairs in the straight line feature set of the reference image and the straight line feature set of the initial registration image to obtain a statistical result, and the step of obtaining the candidate matching set based on the statistical result specifically comprises:
respectively counting a linear feature set Q1And Q3The intersecting straight line pairs in (1) are respectively stored in sets LB and LW; and if the absolute value of the difference between the included angle of a certain orthogonal straight line pair LB in the LB and the included angle of a certain orthogonal straight line pair LW in the LW is smaller than a preset threshold value, the straight line pair LB and the straight line pair LW are considered to be same-name straight line pairs, and all the same-name straight line pairs meeting the relation form a candidate matching set G.
7. The optical remote sensing image registration method of claim 6, wherein the separately statistically linear feature set Q1And Q3The specific step of storing the intersecting straight line pairs in the sets LB and LW respectively comprises the following steps:
to straight line feature set Q1And Q3And the line segments with the middle included angle more than or equal to 30 degrees form a straight line pair, and an intersection point is generated and recorded as:
LB＝{LBij|0＜i,j≤N,i≠j}，LW＝{LWij|0＜i,j≤N,i≠j}；
wherein, LBij＝{(li,lj,Pij,θij)|0＜i,j≤N,i≠j}；LWij＝{(li,lj,Pij,θij)|0＜i,j≤N,i≠j}；
In the formula, N represents the number of straight line features in the image; liAnd ljAre two straight lines that make up a line pair; pijIs the intersection point of the straight line pair; thetaijRepresenting the included angle between the straight line pairs;
if the absolute value of the difference between the included angle of a certain intersecting straight line pair LB in the LB and the included angle of a certain intersecting straight line pair LW in the LW is smaller than the preset threshold, the straight line pair LB and the straight line pair LW are considered to be the same-name straight line pair, and the step of forming the candidate matching set G by all the same-name straight line pairs satisfying the relationship specifically comprises the following steps:
suppose that
And
is a reference image I1The center point is crossed on a point p straight line pair, and the included angle is thetab；
And
is an initial registered image I3The center is crossed with a straight line pair at a point p' and the included angle is thetaw；
Constructing the following similarity measure criteria, if the straight line pair
And a straight line pair
If the following formula is satisfied, the points p and p' are considered as matching control point pairs, and the expression is:
dθ(lb,lw)＝|θb-θw|,0＜dθ(lb,lw)＜dθmax，
wherein,
representing the angular relationship between the pairs of straight lines, dθmaxA threshold value representing the diagonal relation of the straight line;
the pairs of straight lines satisfying the condition constitute a candidate matching set, which is expressed as:
wherein,
and piThe straight line pairs and the intersection points thereof in the reference image;
and p'iThe straight line pairs and the intersection points thereof in the image to be registered.
8. The optical remote sensing image registration method according to claim 6, wherein the step of obtaining a coarse registration result based on the candidate matching set specifically comprises:
defining homonymy line segment matching matrixes MB and MW for representing the matching relation of homonymy line pairs; constructing a bidirectional matching relation according to the homonymous line segment matching matrixes MB and MW, and removing repeated matching to obtain a one-to-one matching point set; obtaining a reference image I by a least square method1And initial registration image I3The optimal transformation matrix is used, and then the coarse registration result I is obtained by utilizing the reverse interpolation4。
9. The optical remote sensing image registration method according to claim 8, wherein the defined homonymous line segment matching matrices MB and MW are used to represent matching relationships of homonymous straight line pairs; constructing a bidirectional matching relationship according to the homonymous line segment matching matrixes MB and MW, and removing repeated matching to obtain a one-to-one matching point set specifically comprises the following steps:
suppose there is N in the candidate matching set G1The lines of the bars are from the reference image, N2The straight lines come from the image to be registered; mb if the ith element in the combined image contains the jth line in the reference imageij1 is ═ 1; otherwise, mbij0; mw if the ith element in G contains the jth line in the image to be registeredij1 is ═ 1; otherwise, mwij＝0；
The matrix MB represents the one-way matching of the image to be registered to the reference image; the matrix MW represents the one-way matching of the reference image to the image to be registered;
and counting the same rows in the matrixes MB and MW, wherein the same rows represent repeated matching, calculating the distance between the same straight lines corresponding to the same rows, and only keeping the straight line pair with the shortest distance.
10. An optical remote sensing image registration method combining SIFT points and control line pairs is characterized by comprising the following steps:
the linear feature set acquisition module is used for respectively acquiring linear feature sets of the reference image and the image to be registered;
the linear feature set acquisition module of the initial registration image is used for acquiring the initial registration image according to the linear feature set of the reference image and the linear feature set of the image to be registered; acquiring a linear feature set of the initial registration image;
the candidate matching set acquisition module is used for respectively counting the intersecting straight line pairs in the straight line feature set of the reference image and the straight line feature set of the initial registration image to obtain a statistical result, and acquiring a candidate matching set based on the statistical result;
a coarse registration result obtaining module, configured to obtain a coarse registration result according to the candidate matching set;
and the accurate registration result acquisition module is used for obtaining an accurate registration result of the reference image and the coarse registration image by utilizing an SIFT key point registration algorithm to complete the registration of the optical remote sensing image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110287621.2A CN113096168B (en) | 2021-03-17 | 2021-03-17 | Optical remote sensing image registration method and system combining SIFT points and control line pairs |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110287621.2A CN113096168B (en) | 2021-03-17 | 2021-03-17 | Optical remote sensing image registration method and system combining SIFT points and control line pairs |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113096168A true CN113096168A (en) | 2021-07-09 |
| CN113096168B CN113096168B (en) | 2024-04-02 |
Family
ID=76668426
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110287621.2A Active CN113096168B (en) | 2021-03-17 | 2021-03-17 | Optical remote sensing image registration method and system combining SIFT points and control line pairs |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113096168B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103514606A (en) * | 2013-10-14 | 2014-01-15 | 武汉大学 | Heterology remote sensing image registration method |
| CN104992431A (en) * | 2015-06-19 | 2015-10-21 | 北京邮电大学 | Method and device for multispectral image registration |
| WO2020206903A1 (en) * | 2019-04-08 | 2020-10-15 | 平安科技（深圳）有限公司 | Image matching method and device, and computer readable storage medium |
-
2021
- 2021-03-17 CN CN202110287621.2A patent/CN113096168B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103514606A (en) * | 2013-10-14 | 2014-01-15 | 武汉大学 | Heterology remote sensing image registration method |
| CN104992431A (en) * | 2015-06-19 | 2015-10-21 | 北京邮电大学 | Method and device for multispectral image registration |
| WO2020206903A1 (en) * | 2019-04-08 | 2020-10-15 | 平安科技（深圳）有限公司 | Image matching method and device, and computer readable storage medium |
Non-Patent Citations (2)
| Title |
|---|
| 余婷;厉小润;: "基于SIFT的全自动遥感图像配准算法", 机电工程, no. 01 * |
| 戴激光;宋伟东;李建军;: "一种可靠的高分辨率光学卫星遥感影像匹配方法", 遥感信息, no. 01 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113096168B (en) | 2024-04-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2024077812A1 (en) | Single building three-dimensional reconstruction method based on point cloud semantic segmentation and structure fitting | |
| Toft et al. | Semantic match consistency for long-term visual localization | |
| Li et al. | Hierarchical line matching based on line–junction–line structure descriptor and local homography estimation | |
| CN102236794B (en) | Recognition and pose determination of 3D objects in 3D scenes | |
| Mousavi et al. | A two-step descriptor-based keypoint filtering algorithm for robust image matching | |
| Hu et al. | Efficient and automatic plane detection approach for 3-D rock mass point clouds | |
| Ge et al. | Image-guided registration of unordered terrestrial laser scanning point clouds for urban scenes | |
| Zhu et al. | A review of 6d object pose estimation | |
| CN116485856B (en) | Unmanned aerial vehicle image geographic registration method based on semantic segmentation and related equipment | |
| Suwanwimolkul et al. | Learning of low-level feature keypoints for accurate and robust detection | |
| CN110246165A (en) | It improves visible images and SAR image matches the method and system of Quasi velosity | |
| CN114332172A (en) | Improved laser point cloud registration method based on covariance matrix | |
| CN113704276A (en) | Map updating method and device, electronic equipment and computer readable storage medium | |
| Yousif et al. | 3D SLAM in texture-less environments using rank order statistics | |
| CN113096168B (en) | Optical remote sensing image registration method and system combining SIFT points and control line pairs | |
| Seo et al. | Learning 3D local surface descriptor for point cloud images of objects in the real-world | |
| CN111160433B (en) | High-speed matching method and system for high-resolution image feature points | |
| Yu et al. | Multimodal urban remote sensing image registration via roadcross triangular feature | |
| Wang et al. | Aerial-DEM geolocalization for GPS-denied UAS navigation | |
| Foresti et al. | A hierarchical approach to feature extraction and grouping | |
| Zhang et al. | 3d convolutional neural network for semantic scene segmentation based on unstructured point clouds | |
| Wang et al. | A Pointer Instrument Reading Approach Based On Mask R-CNN Key Points Detection | |
| Tychola et al. | Identifying Historic Buildings over Time through Image Matching. Technologies 2023, 11, 32 | |
| Xia et al. | Building-Specific Recognition Algorithms Based on Image Sequence | |
| Zeng et al. | The Gestalt Principle for Building Area Extraction |
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 |