CN110660089A - Satellite image registration method and device - Google Patents
Satellite image registration method and device Download PDFInfo
- Publication number
- CN110660089A CN110660089A CN201910908166.6A CN201910908166A CN110660089A CN 110660089 A CN110660089 A CN 110660089A CN 201910908166 A CN201910908166 A CN 201910908166A CN 110660089 A CN110660089 A CN 110660089A
- Authority
- CN
- China
- Prior art keywords
- panchromatic
- feature point
- multispectral
- affine transformation
- satellite image
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 49
- 230000009466 transformation Effects 0.000 claims abstract description 140
- 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 abstract description 118
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 15
- 230000001131 transforming effect Effects 0.000 claims abstract description 11
- 238000005516 engineering process Methods 0.000 claims abstract description 9
- 239000013598 vector Substances 0.000 claims description 88
- 238000004364 calculation method Methods 0.000 claims description 22
- 238000006243 chemical reaction Methods 0.000 claims description 10
- 238000010276 construction Methods 0.000 claims description 10
- 238000000605 extraction Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 238000010586 diagram Methods 0.000 description 8
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 6
- 229910052799 carbon Inorganic materials 0.000 description 6
- 238000012544 monitoring process Methods 0.000 description 6
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 239000000779 smoke Substances 0.000 description 2
- 230000007480 spreading Effects 0.000 description 2
- 238000003708 edge detection Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000003252 repetitive effect Effects 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
-
- G06T3/02—
-
- 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
Abstract
The application discloses a satellite image registration method and a satellite image registration device, wherein the method comprises the following steps: acquiring a panchromatic satellite image and a multispectral satellite image of a target area by using a satellite remote sensing technology; carrying out multi-scale decomposition on the panchromatic satellite image to obtain a low-resolution panchromatic satellite image; extracting panchromatic feature points from the low-resolution panchromatic satellite image, and extracting multispectral feature points from the multispectral satellite image; calculating the Euclidean distance between each panchromatic characteristic point and each multispectral characteristic point; performing characteristic point matching on each panchromatic characteristic point and each multispectral characteristic point to obtain a plurality of characteristic point pairs, wherein the Euclidean distance between the panchromatic characteristic points and the multispectral characteristic points in the characteristic point pairs is the shortest; establishing an affine transformation model according to the feature point coordinates in the feature point pairs; and transforming the coordinates of all pixels of the low-resolution panchromatic satellite image by using an affine transformation model to complete the registration of the satellite image. The problem that the accuracy of the existing method is low is solved.
Description
Technical Field
The present application relates to a method and an apparatus for registering images, and in particular, to a method and an apparatus for registering satellite images.
Background
Forest fires are important influence factors of carbon circulation of a land ecosystem, can change the carbon circulation process and the carbon distribution pattern of the whole land ecosystem, and are one of main disasters which harm the life and property safety of people. The early warning capability of the fire can be improved by acquiring the fire information in time. Because forest fires usually have the characteristics of strong outburst, violent fire behavior and high spreading speed, and the fire places usually have rare smoke and are not easy to perceive, a satellite remote sensing monitoring method becomes an important mode for monitoring forest fires. The satellite remote sensing monitoring can obtain images with two resolutions, namely a panchromatic satellite image and a multispectral satellite image, wherein the panchromatic satellite image generally has higher spatial resolution and cannot display the color of a ground object, and the multispectral satellite image has lower spatial resolution and can display the color of the ground object. And registering the panchromatic satellite image and the multispectral satellite image of the monitored area, and judging whether the monitored area has a fire or not according to the registered images. The accuracy of the images obtained by different registration methods after registration on cloud and water judgment directly influences the accuracy of fire judgment.
In the prior art, registration of a satellite image is realized by reducing the resolution of a panchromatic satellite image, however, the image registration is realized by simply aligning pixels of two images with the same resolution, so that the accuracy of judging cloud and water by the registered image is low, and further the accuracy of judging whether a fire disaster occurs is low.
Therefore, how to provide a high-accuracy satellite image registration method to improve the accuracy of fire judgment has become a technical problem to be solved by those skilled in the art.
Disclosure of Invention
The application provides a satellite image registration method and a satellite image registration device, which are used for solving the problems of low accuracy and the like of the existing satellite image registration method.
In one aspect, the present application provides a method for registering satellite images, including:
acquiring a panchromatic satellite image and a multispectral satellite image of a target area by utilizing a satellite remote sensing technology, wherein the resolution of the panchromatic satellite image is higher than that of the multispectral satellite image;
carrying out multi-scale decomposition on the panchromatic satellite image to obtain a low-resolution panchromatic satellite image;
extracting panchromatic feature points from the low-resolution panchromatic satellite image and extracting multispectral feature points from the multispectral satellite image by adopting a scale invariant feature conversion algorithm;
calculating Euclidean distance between each panchromatic feature point and each multispectral feature point;
according to the Euclidean distance between each panchromatic characteristic point and each multispectral characteristic point, carrying out characteristic point matching on each panchromatic characteristic point and each multispectral characteristic point to obtain a plurality of characteristic point pairs, wherein the Euclidean distance between the panchromatic characteristic points and the multispectral characteristic points in the characteristic point pairs is the shortest;
establishing an affine transformation model according to the panchromatic feature point coordinates and the multispectral feature point coordinates in the feature point pairs;
and transforming the coordinates of all pixels of the low-resolution panchromatic satellite image by using the affine transformation model to obtain transformed pixel coordinates, and finishing the registration of the satellite image.
Optionally, the performing feature point matching on each panchromatic feature point and each multispectral feature point according to the euclidean distance between each panchromatic feature point and each multispectral feature point to obtain a plurality of feature point pairs, where the euclidean distance between the panchromatic feature point and the multispectral feature point in the feature point pairs is the shortest, and the method includes:
extracting the multispectral feature point corresponding to the shortest Euclidean distance of each panchromatic feature point to obtain a first corresponding relation according to the multispectral feature point corresponding to the shortest Euclidean distance of each panchromatic feature point, wherein the first corresponding relation is the corresponding relation between the panchromatic feature point and the multispectral feature point corresponding to the shortest Euclidean distance of the panchromatic feature point;
extracting the panchromatic feature point corresponding to the shortest Euclidean distance of each multispectral feature point to obtain a second corresponding relation according to the panchromatic feature point corresponding to the shortest Euclidean distance of each multispectral feature point, wherein the second corresponding relation is the corresponding relation between the multispectral feature point and the panchromatic feature point corresponding to the shortest Euclidean distance of the multispectral feature point;
and matching feature points of the first corresponding relation and the second corresponding relation to obtain a plurality of feature point pairs, wherein the Euclidean distance between the panchromatic feature points and the multispectral feature points in the feature point pairs is the shortest.
Optionally, the constructing an affine transformation model according to the panchromatic feature point coordinates and the multispectral feature point coordinates in the feature point pair includes:
step one, randomly selecting 3 pairs of the feature point pairs, and calculating affine transformation parameters according to the panchromatic feature point coordinates and the multispectral feature point coordinates of the 3 pairs of the feature point pairs and an affine transformation formula, wherein the affine transformation formula is as follows:
x′＝ax+by+m，y′＝cx+dy+n，
wherein (x, y) is the panchromatic feature point coordinate, (x ', y') is the multispectral feature point coordinate, and a, b, c, d, m and n are the affine transformation parameters;
secondly, performing affine transformation on the matched panchromatic feature point coordinates according to the affine transformation formula of the known affine transformation parameters to obtain transformed feature point coordinates;
calculating a first Euclidean distance, wherein the first Euclidean distance is the Euclidean distance between the transformation characteristic point coordinate and the matched multispectral characteristic point coordinate;
step four, normalizing the first Euclidean distance to obtain an estimation error of the characteristic point pair;
step five, obtaining an estimation error vector according to the estimation errors of all the characteristic point pairs;
step six, determining an error threshold value according to the estimation error vector;
step seven, calculating the number of interior point values in the estimation error vector according to the error threshold, wherein the interior point values are estimation errors smaller than the error threshold in the estimation error vector;
repeating the first step to the seventh step for multiple times until all the characteristic point pairs are selected, so as to obtain a plurality of estimation error vectors;
determining the estimation error vector with the maximum number corresponding to the inner value points as an optimal estimation error vector according to the number of the inner value points corresponding to the plurality of estimation error vectors;
and according to the optimal estimation error vector, using affine transformation parameters corresponding to the optimal estimation error vector to construct an affine transformation model.
Optionally, the constructing an affine transformation model by using the affine transformation parameters corresponding to the optimal estimation error vector according to the optimal estimation error vector includes:
establishing an optimal affine transformation formula by using the affine transformation parameters corresponding to the optimal estimation error vector according to the optimal estimation error vector;
and constructing the affine transformation model by using the optimal affine transformation formula.
Optionally, the transforming coordinates of all pixels of the low-resolution panchromatic satellite image by using the affine transformation model to obtain transformed pixel coordinates, and completing image registration includes:
and inputting the coordinates of all pixels of the low-resolution panchromatic satellite image into the affine transformation model to obtain transformation pixel coordinates, and finishing image registration.
In another aspect, the present application provides a registration apparatus for satellite images, including:
an image acquisition module: the system comprises a panchromatic satellite image acquisition module, a multispectral satellite image acquisition module, a data acquisition module and a data processing module, wherein the panchromatic satellite image acquisition module is used for acquiring a panchromatic satellite image and a multispectral satellite image of a target area by utilizing a satellite remote sensing technology, and the resolution of the panchromatic satellite image is higher than that;
a scale decomposition module: the system is used for carrying out multi-scale decomposition on the high-resolution panchromatic satellite image to obtain a low-resolution panchromatic satellite image;
a feature point extraction module: the system comprises a low-resolution panchromatic satellite image acquisition module, a scale invariant feature conversion module and a scale invariant feature conversion module, wherein the low-resolution panchromatic satellite image acquisition module is used for acquiring a low-resolution panchromatic satellite image;
a distance calculation module: the system is used for respectively calculating the Euclidean distance between each panchromatic characteristic point and each multispectral characteristic point according to the panchromatic characteristic point coordinates and the multispectral characteristic point coordinates;
a feature point matching module: the characteristic point matching module is used for matching characteristic points of each panchromatic characteristic point and each multispectral characteristic point according to the Euclidean distance between each panchromatic characteristic point and each multispectral characteristic point to obtain a plurality of characteristic point pairs, wherein the Euclidean distances between the panchromatic characteristic points and the multispectral characteristic points in the characteristic point pairs are the shortest;
a model construction module: the system is used for constructing an affine transformation model according to the panchromatic feature point coordinates and the multispectral feature point coordinates in the feature point pairs;
an image registration module: and the affine transformation model is used for transforming the coordinates of all pixels of the low-resolution panchromatic satellite image to obtain transformed pixel coordinates, and the registration of the satellite image is completed.
Optionally, the feature point matching module includes:
a first correspondence extraction submodule, configured to extract the multispectral feature point corresponding to the shortest euclidean distance of each panchromatic feature point, so as to obtain a first correspondence according to the multispectral feature point corresponding to the shortest euclidean distance of each panchromatic feature point, where the first correspondence is a correspondence between the panchromatic feature point and the multispectral feature point corresponding to the shortest euclidean distance thereof;
a second correspondence extracting sub-module, configured to extract the panchromatic feature point corresponding to the shortest euclidean distance of each multispectral feature point, so as to obtain a second correspondence according to the panchromatic feature point corresponding to the shortest euclidean distance of each multispectral feature point, where the second correspondence is a correspondence between the multispectral feature point and the panchromatic feature point corresponding to the shortest euclidean distance thereof:
and the characteristic point matching submodule is used for matching characteristic points of the first corresponding relation and the second corresponding relation to obtain a plurality of characteristic point pairs, wherein the Euclidean distance between the panchromatic characteristic points and the multispectral characteristic points in the characteristic point pairs is the shortest.
Optionally, the model building module includes:
a parameter calculation submodule: the method is used for randomly selecting 3 pairs of the feature point pairs, and calculating affine transformation parameters according to the panchromatic feature point coordinates and the multispectral feature point coordinates of the 3 pairs of the feature point pairs and an affine transformation formula, wherein the affine transformation formula is as follows:
x′＝ax+by+m，y′＝cx+dy+n，
wherein (x, y) is the panchromatic feature point coordinate, (x ', y') is the multispectral feature point coordinate, and a, b, c, d, m and n are the affine transformation parameters;
a coordinate transformation submodule: the panchromatic feature point coordinate matching module is used for matching the panchromatic feature point coordinate according to the affine transformation formula of the known affine transformation parameter to obtain a transformed feature point coordinate;
a first distance calculation submodule: the first Euclidean distance is used for calculating a first Euclidean distance, and the first Euclidean distance is the Euclidean distance between the transformation characteristic point coordinate and the matched multispectral characteristic point coordinate;
a normalization submodule: the Euclidean distance is normalized to obtain an estimation error of the characteristic point pair;
constructing a vector submodule: the estimation error vector is obtained according to the estimation errors of all the characteristic point pairs;
setting a threshold submodule: for determining an error threshold based on said estimated error vector;
a quantity calculation submodule: the error threshold value is used for calculating the number of interior point values in the estimation error vector according to the error threshold value, wherein the interior point values are estimation errors which are smaller than the error threshold value in the estimation error vector;
a repetition control submodule: the method is used for controlling repeated operation of a parameter calculation submodule, a coordinate transformation submodule, a first distance calculation submodule, a normalization submodule, a vector construction submodule, a threshold setting submodule and a quantity calculation submodule until all characteristic point pairs are selected, and a plurality of estimation error vectors are obtained;
an optimal vector selection submodule, configured to determine, according to the number of the interior value points corresponding to the multiple estimation error vectors, the estimation error vector with the largest number of the interior value points as an optimal estimation error vector;
and the model construction sub-module is used for constructing an affine transformation model by using the affine transformation parameters corresponding to the optimal estimation error vector according to the optimal estimation error vector.
Optionally, the model building submodule includes:
an optimal formula establishing unit, configured to establish an optimal affine transformation formula using the affine transformation parameters corresponding to the optimal estimation error vector according to the optimal estimation error vector;
and the model building unit is used for building the affine transformation model by using the optimal affine transformation formula.
According to the technical scheme, the registration method of the satellite image, provided by the application, is characterized in that a panchromatic satellite image and a multispectral satellite image of a target area are obtained by utilizing a satellite remote sensing technology, and the resolution of the panchromatic satellite image is higher than that of the multispectral satellite image; carrying out multi-scale decomposition on the panchromatic satellite image to obtain a low-resolution panchromatic satellite image; extracting panchromatic feature points from the low-resolution panchromatic satellite image and extracting multispectral feature points from the multispectral satellite image by adopting a scale invariant feature conversion algorithm; calculating the Euclidean distance between each panchromatic characteristic point and each multispectral characteristic point; according to the Euclidean distance between each panchromatic feature point and each multispectral feature point, performing feature point matching on each panchromatic feature point and each multispectral feature point to obtain a plurality of feature point pairs, wherein the Euclidean distances between the panchromatic feature points and the multispectral feature points in the feature point pairs are the shortest; establishing an affine transformation model according to the panchromatic feature point coordinates and the multispectral feature point coordinates in the feature point pairs; and transforming the coordinates of all pixels of the low-resolution panchromatic satellite image by using an affine transformation model to obtain transformed pixel coordinates, and finishing the registration of the satellite image. Compared with the prior art, the method and the device have the advantages that the satellite images are more accurately registered, and the accuracy of judging the fire according to the satellite image registration result can be improved.
Drawings
In order to more clearly explain the technical solution of the present application, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious to those skilled in the art that other drawings can be obtained according to the drawings without any creative effort.
Fig. 1 is a flowchart of a method for registering satellite images according to the present disclosure;
FIG. 2 is a detailed flowchart of step S5 in FIG. 1;
FIG. 3 is a detailed flowchart of step S6 in FIG. 1;
fig. 4 is a detailed flowchart of step S69 in fig. 3;
FIG. 5 is a schematic structural diagram of a satellite image registration device;
FIG. 6 is a schematic structural diagram of a feature point matching module;
FIG. 7 is a schematic diagram of a model building block;
FIG. 8 is a schematic diagram of the structure of the model building submodule.
Detailed Description
The technical solutions in the embodiments of the present application will be described clearly and completely with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
Forest fires are important influence factors of carbon circulation of a land ecosystem, can change the carbon circulation process and the carbon distribution pattern of the whole land ecosystem, and are one of main disasters which harm the life and property safety of people. The early warning capability of the fire can be improved by acquiring the fire information in time. Because forest fires usually have the characteristics of strong outburst, violent fire behavior and high spreading speed, and the fire places usually have rare smoke and are not easy to perceive, a satellite remote sensing monitoring method becomes an important mode for monitoring forest fires. The satellite remote sensing monitoring can obtain images with two resolutions, namely a panchromatic satellite image and a multispectral satellite image, wherein the panchromatic satellite image generally has higher spatial resolution and cannot display the color of a ground object, and the multispectral satellite image has lower spatial resolution and can display the color of the ground object. And registering the panchromatic satellite image and the multispectral satellite image of the monitored area, and judging whether the monitored area has a fire or not according to the registered images. The accuracy of the images obtained by different registration methods after registration on cloud and water judgment can directly influence the accuracy of fire judgment.
In the prior art, registration of a satellite image is realized by reducing the resolution of a panchromatic satellite image, however, the image registration is realized by simply aligning pixels of two images with the same resolution, so that the accuracy of judging cloud and water by the registered image is low, and further the accuracy of judging whether a fire disaster occurs is low.
In view of the above, in one aspect, fig. 1 is a flowchart of a method for registering a satellite image provided by the present application, and as shown in fig. 1, the present application provides a method for registering a satellite image, including:
s1: and acquiring a panchromatic satellite image and a multispectral satellite image of the target area by using a satellite remote sensing technology, wherein the resolution of the panchromatic satellite image is higher than that of the multispectral satellite image.
The panchromatic satellite image and the multispectral satellite image of the target area can be obtained by using a himawari-8 satellite, but the type and the variety of the used satellite are not limited in particular.
S2: and performing multi-scale decomposition on the panchromatic satellite image to obtain a low-resolution panchromatic satellite image.
It should be noted that the satellite images contain a large amount of information of different scales, which appears simultaneously in one image. The multi-scale decomposition can eliminate the influence of other useless scale information on the processing result, simplifies the processing difficulty and complexity, and is also one of preprocessing methods in the processing processes of image target identification, edge detection and the like. The method changes the panchromatic satellite image with high resolution into the panchromatic satellite image with low resolution through a multi-scale decomposition method.
S3: and extracting panchromatic feature points from the low-resolution panchromatic satellite image and extracting multispectral feature points from the multispectral satellite image by adopting a scale invariant feature conversion algorithm.
Scale-invariant feature transform (Scale-invariant feature transform or SIFT) is an algorithm used in machine vision to detect and describe local features in an image, and it finds extreme points in the spatial Scale and extracts its position coordinates, Scale and invariant rotation. In the step, a scale invariant feature conversion algorithm is adopted, the feature points extracted from the low-resolution panchromatic satellite image are panchromatic feature points, and the feature points extracted from the multispectral satellite image are multispectral feature points. The panchromatic feature points and the multispectral feature points represent local feature information of the low-resolution panchromatic satellite image and the multispectral satellite image.
S4: and calculating the Euclidean distance between each panchromatic characteristic point and each multispectral characteristic point.
Each panchromatic feature point and each multispectral feature point contain information such as respective position coordinates of the panchromatic feature point and each multispectral feature point, and the Euclidean distance between each panchromatic feature point and each multispectral feature point is calculated according to the coordinates of each panchromatic feature point and the coordinates of each multispectral feature point.
Euclidean metric (also known as euclidean distance) is a commonly used definition of distance, referring to the true distance between two points in an m-dimensional space, or the natural length of a vector (i.e., the distance of the point from the origin). The euclidean distance in two and three dimensions is the actual distance between two points.
S5: and matching the characteristic points of each panchromatic characteristic point and each multispectral characteristic point according to the Euclidean distance between each panchromatic characteristic point and each multispectral characteristic point to obtain a plurality of characteristic point pairs, wherein the Euclidean distances between the panchromatic characteristic points and the multispectral characteristic points in the characteristic point pairs are the shortest.
Optionally, fig. 2 is a detailed flowchart of step S5 in fig. 1, and as shown in fig. 2, the performing feature point matching on each panchromatic feature point and each multispectral feature point according to the euclidean distance between each panchromatic feature point and each multispectral feature point to obtain a plurality of feature point pairs, where the euclidean distance between the panchromatic feature point and the multispectral feature point in the feature point pairs is the shortest, and the step includes:
s51: and extracting the multispectral feature point corresponding to the shortest Euclidean distance of each panchromatic feature point to obtain a first corresponding relation according to the multispectral feature point corresponding to the shortest Euclidean distance of each panchromatic feature point, wherein the first corresponding relation is the corresponding relation between the panchromatic feature point and the multispectral feature point corresponding to the shortest Euclidean distance of the panchromatic feature point.
And extracting the multispectral feature point corresponding to the shortest Euclidean distance by taking the panchromatic feature point as a reference, and establishing a first corresponding relation between the panchromatic feature point and the multispectral feature point corresponding to the shortest Euclidean distance of the panchromatic feature point.
S52: extracting a panchromatic feature point corresponding to the shortest Euclidean distance of each multispectral feature point to obtain a second corresponding relation according to the panchromatic feature point corresponding to the shortest Euclidean distance of each multispectral feature point, wherein the second corresponding relation is the corresponding relation between the multispectral feature point and the panchromatic feature point corresponding to the shortest Euclidean distance of the multispectral feature point;
and extracting panchromatic feature points corresponding to the shortest Euclidean distance by taking the multispectral feature points as the reference, and establishing a second corresponding relation between the multispectral feature points and the panchromatic feature points corresponding to the shortest Euclidean distance.
S53: and matching the first corresponding relation with the second corresponding relation to obtain a plurality of characteristic point pairs, wherein the Euclidean distance between the panchromatic characteristic points and the multispectral characteristic points in the characteristic point pairs is the shortest.
It should be noted that each feature point pair includes a panchromatic feature point and a multispectral feature point; panchromatic feature points in all the feature point pairs do not repeatedly appear, namely, each panchromatic feature point is not repeatedly paired to form a feature point pair, and similarly, multispectral feature points in all the feature point pairs do not repeatedly appear, namely, each multispectral feature point is not repeatedly paired to form a feature point pair.
S6: and constructing an affine transformation model according to the panchromatic feature point coordinates and the multispectral feature point coordinates in the feature point pairs.
Optionally, fig. 3 is a detailed flowchart of step S6 in fig. 1, and as shown in fig. 3, the constructing an affine transformation model according to the panchromatic feature point coordinates and the multispectral feature point coordinates in the feature point pair includes:
s60: step one, randomly selecting 3 pairs of feature point pairs, and calculating affine transformation parameters according to panchromatic feature point coordinates and multispectral feature point coordinates of the 3 pairs of feature point pairs and an affine transformation formula, wherein the affine transformation formula is as follows:
x′＝ax+by+m，y′＝cx+dy+n，
wherein, (x, y) is panchromatic characteristic point coordinates, (x ', y') is multispectral characteristic point coordinates, and a, b, c, d, m and n are affine transformation parameters.
S61: and secondly, carrying out affine transformation on the matched panchromatic feature point coordinates according to an affine transformation formula of the known affine transformation parameters to obtain transformed feature point coordinates.
Substituting the affine transformation parameters obtained in the step S60 into an affine transformation formula to obtain an affine transformation formula of known affine transformation parameters; and inputting all the matched panchromatic feature point coordinates into an affine transformation formula of known affine transformation parameters, and carrying out affine transformation on the matched panchromatic feature point coordinates to obtain transformed feature point coordinates.
S62: and step three, calculating a first Euclidean distance, wherein the first Euclidean distance is the Euclidean distance between the transformation characteristic point coordinates and the matched multispectral characteristic point coordinates.
And calculating the Euclidean distance of the multispectral feature points corresponding to the transformed feature points and the panchromatic feature points before transformation.
S63: step four, normalizing the first Euclidean distance to obtain an estimation error of the characteristic point pair;
each pair of characteristic points corresponds to a normalized estimation error.
S64: and step five, obtaining an estimation error vector according to the estimation errors of all the characteristic point pairs.
S65: and step six, determining an error threshold according to the estimated error vector.
The error threshold is determined according to actual conditions, and the application is not particularly limited.
S66: and step seven, calculating the number of interior point values in the estimation error vector according to the error threshold, wherein the interior point values are estimation errors smaller than the error threshold in the estimation error vector.
S67: and repeating the steps one to seven for multiple times until all the characteristic point pairs are selected, so as to obtain a plurality of estimation error vectors.
S68: and determining the estimation error vector with the maximum number of corresponding inner value points as the optimal estimation error vector according to the number of the inner value points corresponding to the plurality of estimation error vectors.
S69: and according to the optimal estimation error vector, using the affine transformation parameters corresponding to the optimal estimation error vector to construct an affine transformation model.
Optionally, fig. 4 is a detailed flowchart of step S69 in fig. 3, and as shown in fig. 4, the constructing an affine transformation model according to the optimal estimation error vector by using the affine transformation parameters corresponding to the optimal estimation error vector includes:
s691: establishing an optimal affine transformation formula by using affine transformation parameters corresponding to the optimal estimation error vector according to the optimal estimation error vector;
s692: and (5) constructing an affine transformation model by using the optimal affine transformation formula.
S7: and transforming the coordinates of all pixels of the low-resolution panchromatic satellite image by using an affine transformation model to obtain transformed pixel coordinates, and finishing the registration of the satellite image.
Optionally, the affine transformation model is used to transform coordinates of all pixels of the low-resolution panchromatic satellite image to obtain transformed pixel coordinates, and image registration is completed, including:
and inputting the coordinates of all pixels of the low-resolution panchromatic satellite image into the affine transformation model to obtain transformation pixel coordinates, and finishing image registration.
And the transformed pixel coordinates obtained after image registration form a registered image, and fire disaster judgment can be performed according to the registered image.
On the other hand, fig. 5 is a schematic structural diagram of a registration apparatus for satellite images, and as shown in fig. 5, the present application provides a registration apparatus for satellite images, including:
the image acquisition module 1: the system comprises a satellite remote sensing device, a panchromatic satellite image acquisition device, a multispectral satellite image acquisition device and a multispectral satellite image acquisition device, wherein the panchromatic satellite image acquisition device is used for acquiring a panchromatic satellite image and a multispectral satellite image of a target area by utilizing a satellite remote sensing technology, and the resolution of the panchromat;
the scale decomposition module 2: the system is used for carrying out multi-scale decomposition on the high-resolution panchromatic satellite image to obtain a low-resolution panchromatic satellite image;
the feature point extracting module 3: the device comprises a scale invariant feature transformation algorithm, a low-resolution panchromatic satellite image extraction module, a scale invariant feature transformation algorithm and a multi-spectral feature transformation algorithm, wherein the scale invariant feature transformation algorithm is used for extracting panchromatic feature points from the low-resolution panchromatic satellite image and extracting multi-spectral feature points from the multi-spectral satellite image;
the distance calculation module 4: the system comprises a panchromatic characteristic point coordinate system, a multispectral characteristic point coordinate system and a multispectral characteristic point coordinate system, wherein the panchromatic characteristic point coordinate system is used for calculating the coordinate system of each panchromatic characteristic point;
the feature point matching module 5: the characteristic point matching device is used for matching characteristic points of each panchromatic characteristic point and each multispectral characteristic point according to the Euclidean distance between each panchromatic characteristic point and each multispectral characteristic point to obtain a plurality of characteristic point pairs, and the Euclidean distances between the panchromatic characteristic points and the multispectral characteristic points in the characteristic point pairs are the shortest;
the model construction module 6: the method comprises the steps of establishing an affine transformation model according to panchromatic feature point coordinates and multispectral feature point coordinates in feature point pairs;
the image registration module 7: and the method is used for transforming the coordinates of all pixels of the low-resolution panchromatic satellite image by using an affine transformation model to obtain transformed pixel coordinates, and finishing the registration of the satellite image.
Optionally, fig. 6 is a schematic structural diagram of the feature point matching module, and as shown in fig. 6, the feature point matching module 5 includes:
a first correspondence extracting submodule 51, configured to extract a multispectral feature point corresponding to the shortest euclidean distance of each panchromatic feature point, so as to obtain a first correspondence according to the multispectral feature point corresponding to the shortest euclidean distance of each panchromatic feature point, where the first correspondence is a correspondence between a panchromatic feature point and a multispectral feature point corresponding to the shortest euclidean distance thereof;
a second correspondence extracting sub-module 52, configured to extract a panchromatic feature point corresponding to the shortest euclidean distance of each multispectral feature point, so as to obtain a second correspondence according to the panchromatic feature point corresponding to the shortest euclidean distance of each multispectral feature point, where the second correspondence is a correspondence between a multispectral feature point and a panchromatic feature point corresponding to the shortest euclidean distance thereof;
and the feature point matching submodule 53 is configured to perform feature point matching on the first corresponding relationship and the second corresponding relationship to obtain a plurality of feature point pairs, where euclidean distances between the panchromatic feature points and the multispectral feature points in the feature point pairs are the shortest.
Optionally, fig. 7 is a schematic structural diagram of the model building module, and as shown in fig. 7, the model building module 6 includes:
parameter calculation submodule 60: the method is used for randomly selecting 3 pairs of the feature point pairs, and calculating affine transformation parameters according to the panchromatic feature point coordinates and the multispectral feature point coordinates of the 3 pairs of the feature point pairs and an affine transformation formula, wherein the affine transformation formula is as follows:
x′＝ax+by+m，y′＝cx+dy+n，
wherein, (x, y) is panchromatic characteristic point coordinates, (x ', y') is multispectral characteristic point coordinates, and a, b, c, d, m and n are affine transformation parameters;
coordinate transformation submodule 61: the system comprises a matching panchromatic feature point coordinate processing unit, a matching panchromatic feature point coordinate processing unit and a matching panchromatic feature point coordinate processing unit, wherein the matching panchromatic feature point coordinate processing unit is used for carrying out affine transformation on the panchromatic feature point coordinate according to an affine transformation formula of a known affine transformation parameter to obtain a transformation feature point coordinate;
the first distance calculation submodule 62: the first Euclidean distance is used for calculating a first Euclidean distance, and the first Euclidean distance is the Euclidean distance between the transformation characteristic point coordinates and the matched multispectral characteristic point coordinates;
normalization submodule 63: the first Euclidean distance is normalized to obtain an estimation error of the characteristic point pair;
construct vector submodule 64: the method comprises the steps of obtaining an estimation error vector according to estimation errors of all feature point pairs;
set threshold submodule 65: for determining an error threshold based on the estimated error vector;
quantity calculation submodule 66: the method comprises the steps of calculating the number of interior point values in an estimation error vector according to an error threshold, wherein the interior point values are estimation errors smaller than the error threshold in the estimation error vector;
the repetitive control sub-module 67: the method is used for controlling repeated operation of a parameter calculation submodule, a coordinate transformation submodule, a first distance calculation submodule, a normalization submodule, a vector construction submodule, a threshold setting submodule and a quantity calculation submodule until all characteristic point pairs are selected, and obtaining a plurality of estimation error vectors;
an optimal vector selection submodule 68, configured to determine, according to the number of the interior value points corresponding to the multiple estimation error vectors, the estimation error vector with the largest number of interior value points as an optimal estimation error vector;
a model construction sub-module 69 for constructing an affine transformation model from the optimal estimated error vector using the affine transformation parameters corresponding to the optimal estimated error vector.
Optionally, fig. 8 is a schematic structural diagram of the model building submodule, and as shown in fig. 8, the model building submodule 69 includes:
an optimal formula establishing unit 691, configured to establish an optimal affine transformation formula using the affine transformation parameters corresponding to the optimal estimation error vector according to the optimal estimation error vector;
a model construction unit 692 for constructing an affine transformation model using the optimal affine transformation formula.
According to the registration method of the satellite images, a panchromatic satellite image and a multispectral satellite image of a target area are obtained by utilizing a satellite remote sensing technology, and the resolution of the panchromatic satellite image is higher than that of the multispectral satellite image; carrying out multi-scale decomposition on the panchromatic satellite image to obtain a low-resolution panchromatic satellite image; extracting panchromatic feature points from the low-resolution panchromatic satellite image and extracting multispectral feature points from the multispectral satellite image by adopting a scale invariant feature conversion algorithm; calculating the Euclidean distance between each panchromatic characteristic point and each multispectral characteristic point; according to the Euclidean distance between each panchromatic feature point and each multispectral feature point, performing feature point matching on each panchromatic feature point and each multispectral feature point to obtain a plurality of feature point pairs, wherein the Euclidean distances between the panchromatic feature points and the multispectral feature points in the feature point pairs are the shortest; establishing an affine transformation model according to the panchromatic feature point coordinates and the multispectral feature point coordinates in the feature point pairs; and transforming the coordinates of all pixels of the low-resolution panchromatic satellite image by using an affine transformation model to obtain transformed pixel coordinates, and finishing the registration of the satellite image. Compared with the prior art, the method and the device have the advantages that the satellite images are more accurately registered, so that the accuracy of judging the fire according to the satellite image registration result is improved.
According to the satellite image registration method and device, the satellite image is subjected to multi-scale decomposition, feature point extraction and Euclidean distance calculation between feature points, an affine transformation model is established, and the satellite image is registered through the affine transformation model. The accuracy of registration is higher, and the accuracy of fire judgment is higher by utilizing the registered images formed by the registered transformation coordinates.
Those skilled in the art will readily appreciate that the techniques of the embodiments of the present invention may be implemented as software plus a required general purpose hardware platform. Based on such understanding, the technical solutions in the embodiments of the present invention may be essentially or partially implemented in the form of a software product, which may be stored in a storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method according to the embodiments or some parts of the embodiments.
The same and similar parts in the various embodiments in this specification may be referred to each other. In particular, for the embodiments, since they are substantially similar to the method embodiments, the description is simple, and the relevant points can be referred to the description in the method embodiments.
Claims (9)
1. A method for registration of satellite images, comprising:
acquiring a panchromatic satellite image and a multispectral satellite image of a target area by utilizing a satellite remote sensing technology, wherein the resolution of the panchromatic satellite image is higher than that of the multispectral satellite image;
carrying out multi-scale decomposition on the panchromatic satellite image to obtain a low-resolution panchromatic satellite image;
extracting panchromatic feature points from the low-resolution panchromatic satellite image and extracting multispectral feature points from the multispectral satellite image by adopting a scale invariant feature conversion algorithm;
calculating Euclidean distance between each panchromatic feature point and each multispectral feature point;
according to the Euclidean distance between each panchromatic characteristic point and each multispectral characteristic point, carrying out characteristic point matching on each panchromatic characteristic point and each multispectral characteristic point to obtain a plurality of characteristic point pairs, wherein the Euclidean distance between the panchromatic characteristic points and the multispectral characteristic points in the characteristic point pairs is the shortest;
establishing an affine transformation model according to the panchromatic feature point coordinates and the multispectral feature point coordinates in the feature point pairs;
and transforming the coordinates of all pixels of the low-resolution panchromatic satellite image by using the affine transformation model to obtain transformed pixel coordinates, and finishing the registration of the satellite image.
2. The method according to claim 1, wherein the performing feature point matching on each panchromatic feature point and each multispectral feature point according to the euclidean distance between each panchromatic feature point and each multispectral feature point to obtain a plurality of feature point pairs, wherein the euclidean distances between the panchromatic feature points and the multispectral feature points in the feature point pairs are the shortest, comprises:
extracting the multispectral feature point corresponding to the shortest Euclidean distance of each panchromatic feature point to obtain a first corresponding relation according to the multispectral feature point corresponding to the shortest Euclidean distance of each panchromatic feature point, wherein the first corresponding relation is the corresponding relation between the panchromatic feature point and the multispectral feature point corresponding to the shortest Euclidean distance of the panchromatic feature point;
extracting the panchromatic feature point corresponding to the shortest Euclidean distance of each multispectral feature point to obtain a second corresponding relation according to the panchromatic feature point corresponding to the shortest Euclidean distance of each multispectral feature point, wherein the second corresponding relation is the corresponding relation between the multispectral feature point and the panchromatic feature point corresponding to the shortest Euclidean distance of the multispectral feature point;
and matching feature points of the first corresponding relation and the second corresponding relation to obtain a plurality of feature point pairs, wherein the Euclidean distance between the panchromatic feature points and the multispectral feature points in the feature point pairs is the shortest.
3. The method of claim 1, wherein constructing an affine transformation model from panchromatic feature point coordinates and multispectral feature point coordinates in the feature point pairs comprises:
step one, randomly selecting 3 pairs of the feature point pairs, and calculating affine transformation parameters according to the panchromatic feature point coordinates and the multispectral feature point coordinates of the 3 pairs of the feature point pairs and an affine transformation formula, wherein the affine transformation formula is as follows:
x′＝ax+by+m，y′＝cx+dy+n，
wherein (x, y) is the panchromatic feature point coordinate, (x ', y') is the multispectral feature point coordinate, and a, b, c, d, m and n are the affine transformation parameters;
secondly, performing affine transformation on the matched panchromatic feature point coordinates according to the affine transformation formula of the known affine transformation parameters to obtain transformed feature point coordinates;
calculating a first Euclidean distance, wherein the first Euclidean distance is the Euclidean distance between the transformation characteristic point coordinate and the matched multispectral characteristic point coordinate;
step four, normalizing the first Euclidean distance to obtain an estimation error of the characteristic point pair;
step five, obtaining an estimation error vector according to the estimation errors of all the characteristic point pairs;
step six, determining an error threshold value according to the estimation error vector;
step seven, calculating the number of interior point values in the estimation error vector according to the error threshold, wherein the interior point values are estimation errors smaller than the error threshold in the estimation error vector;
repeating the first step to the seventh step for multiple times until all the characteristic point pairs are selected, so as to obtain a plurality of estimation error vectors;
determining the estimation error vector with the maximum number corresponding to the inner value points as an optimal estimation error vector according to the number of the inner value points corresponding to the plurality of estimation error vectors;
and according to the optimal estimation error vector, using affine transformation parameters corresponding to the optimal estimation error vector to construct an affine transformation model.
4. The method according to claim 3, wherein the constructing an affine transformation model using affine transformation parameters corresponding to the optimal estimation error vector according to the optimal estimation error vector comprises:
establishing an optimal affine transformation formula by using the affine transformation parameters corresponding to the optimal estimation error vector according to the optimal estimation error vector;
and constructing the affine transformation model by using the optimal affine transformation formula.
5. The method of claim 4, wherein transforming coordinates of all pixels of the low-resolution panchromatic satellite image using the affine transformation model to obtain transformed pixel coordinates to complete image registration comprises:
and inputting the coordinates of all pixels of the low-resolution panchromatic satellite image into the affine transformation model to obtain transformation pixel coordinates, and finishing image registration.
6. An apparatus for registering satellite images, comprising:
an image acquisition module: the system comprises a panchromatic satellite image acquisition module, a multispectral satellite image acquisition module, a data acquisition module and a data processing module, wherein the panchromatic satellite image acquisition module is used for acquiring a panchromatic satellite image and a multispectral satellite image of a target area by utilizing a satellite remote sensing technology, and the resolution of the panchromatic satellite image is higher than that;
a scale decomposition module: the system is used for carrying out multi-scale decomposition on the high-resolution panchromatic satellite image to obtain a low-resolution panchromatic satellite image;
a feature point extraction module: the system comprises a low-resolution panchromatic satellite image acquisition module, a scale invariant feature conversion module and a scale invariant feature conversion module, wherein the low-resolution panchromatic satellite image acquisition module is used for acquiring a low-resolution panchromatic satellite image;
a distance calculation module: the system is used for respectively calculating the Euclidean distance between each panchromatic characteristic point and each multispectral characteristic point according to the panchromatic characteristic point coordinates and the multispectral characteristic point coordinates;
a feature point matching module: the characteristic point matching module is used for matching characteristic points of each panchromatic characteristic point and each multispectral characteristic point according to the Euclidean distance between each panchromatic characteristic point and each multispectral characteristic point to obtain a plurality of characteristic point pairs, wherein the Euclidean distances between the panchromatic characteristic points and the multispectral characteristic points in the characteristic point pairs are the shortest;
a model construction module: the system is used for constructing an affine transformation model according to the panchromatic feature point coordinates and the multispectral feature point coordinates in the feature point pairs;
an image registration module: and the affine transformation model is used for transforming the coordinates of all pixels of the low-resolution panchromatic satellite image to obtain transformed pixel coordinates, and the registration of the satellite image is completed.
7. The apparatus of claim 6, wherein the feature point matching module comprises:
a first correspondence extraction submodule, configured to extract the multispectral feature point corresponding to the shortest euclidean distance of each panchromatic feature point, so as to obtain a first correspondence according to the multispectral feature point corresponding to the shortest euclidean distance of each panchromatic feature point, where the first correspondence is a correspondence between the panchromatic feature point and the multispectral feature point corresponding to the shortest euclidean distance thereof;
a second correspondence extracting sub-module, configured to extract the panchromatic feature point corresponding to the shortest euclidean distance of each multispectral feature point, so as to obtain a second correspondence according to the panchromatic feature point corresponding to the shortest euclidean distance of each multispectral feature point, where the second correspondence is a correspondence between the multispectral feature point and the panchromatic feature point corresponding to the shortest euclidean distance thereof:
and the characteristic point matching submodule is used for matching characteristic points of the first corresponding relation and the second corresponding relation to obtain a plurality of characteristic point pairs, wherein the Euclidean distance between the panchromatic characteristic points and the multispectral characteristic points in the characteristic point pairs is the shortest.
8. The apparatus of claim 7, wherein the model building module comprises:
a parameter calculation submodule: the method is used for randomly selecting 3 pairs of the feature point pairs, and calculating affine transformation parameters according to the panchromatic feature point coordinates and the multispectral feature point coordinates of the 3 pairs of the feature point pairs and an affine transformation formula, wherein the affine transformation formula is as follows:
x′＝ax+by+m，y′＝cx+dy+n，
wherein (x, y) is the panchromatic feature point coordinate, (x ', y') is the multispectral feature point coordinate, and a, b, c, d, m and n are the affine transformation parameters;
a coordinate transformation submodule: the panchromatic feature point coordinate matching module is used for matching the panchromatic feature point coordinate according to the affine transformation formula of the known affine transformation parameter to obtain a transformed feature point coordinate;
a first distance calculation submodule: the first Euclidean distance is used for calculating a first Euclidean distance, and the first Euclidean distance is the Euclidean distance between the transformation characteristic point coordinate and the matched multispectral characteristic point coordinate;
a normalization submodule: the Euclidean distance is normalized to obtain an estimation error of the characteristic point pair;
constructing a vector submodule: the estimation error vector is obtained according to the estimation errors of all the characteristic point pairs;
setting a threshold submodule: for determining an error threshold based on said estimated error vector;
a quantity calculation submodule: the error threshold value is used for calculating the number of interior point values in the estimation error vector according to the error threshold value, wherein the interior point values are estimation errors which are smaller than the error threshold value in the estimation error vector;
a repetition control submodule: the method is used for controlling repeated operation of a parameter calculation submodule, a coordinate transformation submodule, a first distance calculation submodule, a normalization submodule, a vector construction submodule, a threshold setting submodule and a quantity calculation submodule until all characteristic point pairs are selected, and a plurality of estimation error vectors are obtained;
an optimal vector selection submodule, configured to determine, according to the number of the interior value points corresponding to the multiple estimation error vectors, the estimation error vector with the largest number of the interior value points as an optimal estimation error vector;
and the model construction sub-module is used for constructing an affine transformation model by using the affine transformation parameters corresponding to the optimal estimation error vector according to the optimal estimation error vector.
9. The apparatus of claim 8, wherein the model building sub-module comprises:
an optimal formula establishing unit, configured to establish an optimal affine transformation formula using the affine transformation parameters corresponding to the optimal estimation error vector according to the optimal estimation error vector;
and the model building unit is used for building the affine transformation model by using the optimal affine transformation formula.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910908166.6A CN110660089A (en) | 2019-09-25 | 2019-09-25 | Satellite image registration method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910908166.6A CN110660089A (en) | 2019-09-25 | 2019-09-25 | Satellite image registration method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110660089A true CN110660089A (en) | 2020-01-07 |
Family
ID=69038970
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910908166.6A Pending CN110660089A (en) | 2019-09-25 | 2019-09-25 | Satellite image registration method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110660089A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112330700A (en) * | 2020-11-16 | 2021-02-05 | 四川航天神坤科技有限公司 | Cultivated land plot extraction method of satellite image |
| CN112598717A (en) * | 2020-12-14 | 2021-04-02 | 珠海欧比特宇航科技股份有限公司 | Full-spectrum registration method and medium for hyperspectral satellite images |
| CN113393499A (en) * | 2021-07-12 | 2021-09-14 | 自然资源部国土卫星遥感应用中心 | Automatic registration method for panchromatic image and multispectral image of high-resolution seven-satellite |
| CN113838108A (en) * | 2021-09-29 | 2021-12-24 | 中国科学院长春光学精密机械与物理研究所 | Charge accumulation direction testing system for multi-scale spectrum TDICMOS point-line registration |
| CN114820738A (en) * | 2022-06-30 | 2022-07-29 | 中国人民解放军国防科技大学 | Accurate registration method and device for star atlas, computer equipment and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100048242A1 (en) * | 2008-08-19 | 2010-02-25 | Rhoads Geoffrey B | Methods and systems for content processing |
| CN101996396A (en) * | 2010-09-16 | 2011-03-30 | 湖南大学 | Compressive sensing theory-based satellite remote sensing image fusion method |
| CN102800098A (en) * | 2012-07-19 | 2012-11-28 | 中国科学院自动化研究所 | Multi-characteristic multi-level visible light full-color and multi-spectrum high-precision registering method |
| CN103218796A (en) * | 2013-05-14 | 2013-07-24 | 中国科学院自动化研究所 | Fusion method of full color-multispectral remote sensing images |
| WO2014183259A1 (en) * | 2013-05-14 | 2014-11-20 | 中国科学院自动化研究所 | Full-color and multi-spectral remote sensing image fusion method |
| CN106651827A (en) * | 2016-09-09 | 2017-05-10 | 浙江大学 | Fundus image registering method based on SIFT characteristics |
| CN107631728A (en) * | 2017-09-13 | 2018-01-26 | 哈尔滨工业大学 | A kind of spaceborne visual aids air navigation aid |
-
2019
- 2019-09-25 CN CN201910908166.6A patent/CN110660089A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100048242A1 (en) * | 2008-08-19 | 2010-02-25 | Rhoads Geoffrey B | Methods and systems for content processing |
| CN101996396A (en) * | 2010-09-16 | 2011-03-30 | 湖南大学 | Compressive sensing theory-based satellite remote sensing image fusion method |
| CN102800098A (en) * | 2012-07-19 | 2012-11-28 | 中国科学院自动化研究所 | Multi-characteristic multi-level visible light full-color and multi-spectrum high-precision registering method |
| CN103218796A (en) * | 2013-05-14 | 2013-07-24 | 中国科学院自动化研究所 | Fusion method of full color-multispectral remote sensing images |
| WO2014183259A1 (en) * | 2013-05-14 | 2014-11-20 | 中国科学院自动化研究所 | Full-color and multi-spectral remote sensing image fusion method |
| CN106651827A (en) * | 2016-09-09 | 2017-05-10 | 浙江大学 | Fundus image registering method based on SIFT characteristics |
| CN107631728A (en) * | 2017-09-13 | 2018-01-26 | 哈尔滨工业大学 | A kind of spaceborne visual aids air navigation aid |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112330700A (en) * | 2020-11-16 | 2021-02-05 | 四川航天神坤科技有限公司 | Cultivated land plot extraction method of satellite image |
| CN112598717A (en) * | 2020-12-14 | 2021-04-02 | 珠海欧比特宇航科技股份有限公司 | Full-spectrum registration method and medium for hyperspectral satellite images |
| CN113393499A (en) * | 2021-07-12 | 2021-09-14 | 自然资源部国土卫星遥感应用中心 | Automatic registration method for panchromatic image and multispectral image of high-resolution seven-satellite |
| CN113838108A (en) * | 2021-09-29 | 2021-12-24 | 中国科学院长春光学精密机械与物理研究所 | Charge accumulation direction testing system for multi-scale spectrum TDICMOS point-line registration |
| CN113838108B (en) * | 2021-09-29 | 2023-07-21 | 中国科学院长春光学精密机械与物理研究所 | Multi-scale spectrum TDICMOS point-line registration charge accumulation direction test system |
| CN114820738A (en) * | 2022-06-30 | 2022-07-29 | 中国人民解放军国防科技大学 | Accurate registration method and device for star atlas, computer equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110660089A (en) | Satellite image registration method and device | |
| US11244197B2 (en) | Fast and robust multimodal remote sensing image matching method and system | |
| CN109615611B (en) | Inspection image-based insulator self-explosion defect detection method | |
| CN109348731B (en) | Image matching method and device | |
| Bouchiha et al. | Automatic remote-sensing image registration using SURF | |
| EP2534612B1 (en) | Efficient scale-space extraction and description of interest points | |
| Wang et al. | Recognition and location of the internal corners of planar checkerboard calibration pattern image | |
| Liu et al. | A contrario comparison of local descriptors for change detection in very high spatial resolution satellite images of urban areas | |
| CN108986152B (en) | Foreign matter detection method and device based on difference image | |
| CN110033484B (en) | High canopy density forest sample plot tree height extraction method combining UAV image and TLS point cloud | |
| KR101460313B1 (en) | Apparatus and method for robot localization using visual feature and geometric constraints | |
| CN110084743B (en) | Image splicing and positioning method based on multi-flight-zone initial flight path constraint | |
| CN104318559A (en) | Quick feature point detecting method for video image matching | |
| CN105678720A (en) | Image matching judging method and image matching judging device for panoramic stitching | |
| CN106529548A (en) | Sub-pixel level multi-scale Harris corner point detection algorithm | |
| Essa et al. | Volumetric directional pattern for spatial feature extraction in hyperspectral imagery | |
| CN112435283A (en) | Image registration method, electronic device and computer-readable storage medium | |
| CN112150522A (en) | Remote sensing image registration method, device, equipment, storage medium and system | |
| CN116363583A (en) | Human body identification method, device, equipment and medium for top view angle | |
| CN116128919A (en) | Multi-temporal image abnormal target detection method and system based on polar constraint | |
| CN113223033A (en) | Poultry body temperature detection method, device and medium based on image fusion | |
| CN114255398A (en) | Method and device for extracting and matching features of satellite video image | |
| Wu et al. | An accurate feature point matching algorithm for automatic remote sensing image registration | |
| Shen et al. | SAR and optical image registration based on edge features | |
| CN112907626A (en) | Moving object extraction method based on satellite time-exceeding phase data multi-source information |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200107 |