CN115690460A - Remote sensing image fusion method considering image distortion and computer readable medium - Google Patents
Remote sensing image fusion method considering image distortion and computer readable medium Download PDFInfo
- Publication number
- CN115690460A CN115690460A CN202211434744.5A CN202211434744A CN115690460A CN 115690460 A CN115690460 A CN 115690460A CN 202211434744 A CN202211434744 A CN 202211434744A CN 115690460 A CN115690460 A CN 115690460A
- Authority
- CN
- China
- Prior art keywords
- image
- remote sensing
- block
- polygon
- resolution
- 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
Images
Landscapes
- Image Processing (AREA)
Abstract
The invention provides a remote sensing image fusion method considering image distortion and a computer readable medium. The method solves the intersected polygons participating in the fusion image, and constructs the minimum circumscribed rectangle of the intersected polygons according to the geographic range of the intersected polygons. And carrying out grid division on the minimum external rectangle to obtain a plurality of blocks, carrying out initial matching, and carrying out affine transformation processing on the corrected high-spectral-resolution image based on a matching result. Updating the intersected polygons participating in the fusion image and the corresponding minimum external rectangles, carrying out grid division to obtain a plurality of blocks, carrying out fine matching processing, and further carrying out polynomial correction on the high-spectral-resolution remote sensing image subjected to affine transformation processing based on a matching result. And respectively carrying out fusion processing on each block participating in the fusion image so as to obtain the fused remote sensing image. The method can effectively solve the problem of poor fusion effect caused by image internal distortion, and can realize the fusion between large-width remote sensing images.
Description
Technical Field
The invention belongs to the technical field of image processing, and particularly relates to a remote sensing image fusion method considering image distortion and a computer readable medium.
Background
The fusion between the high spatial resolution image and the high spectral resolution image is called remote sensing image space spectrum fusion, and the space spectrum fusion is a hot point problem in the field of remote sensing image fusion because the inherent contradiction between the imaging system space and the spectral resolution can be broken through. However, in the spatial spectrum fusion, due to the difference of the spatial spectrum image data sources, the registration error between the images will have a great influence on the fusion performance. In addition, the wide coverage range of the remote sensing image and the large internal topographic relief of the remote sensing image cause that the local geometric deformation of the image is different, namely, the image has internal distortion, and at the moment, the same registration model is adopted for the whole image, so that higher registration precision is difficult to obtain, and the fusion effect is further restricted. How to effectively improve the registration accuracy is a difficult problem which needs to be solved urgently in space spectrum fusion.
Disclosure of Invention
The invention provides a remote sensing image fusion method and a computer readable medium considering image distortion aiming at the defects of the prior art. In consideration of low registration accuracy caused by image internal distortion and further influence on a fusion effect, when fusion is carried out, a remote sensing image fusion method considering image distortion is provided by combining a coarse registration method and a block fine registration method.
In order to achieve the above object, the technical solution provided by the present invention is a remote sensing image fusion method considering image distortion, comprising the following steps:
step 1: acquiring an original high-spatial-resolution remote sensing image and a corresponding RPC positioning model thereof, and carrying out geometric correction to obtain a corrected high-spatial-resolution remote sensing image; acquiring a high-spectral-resolution remote sensing image and a corresponding RPC positioning model thereof, and performing geometric correction to obtain a corrected high-spectral-resolution remote sensing image;
step 2: calculating the intersection polygons of the corrected high-spatial-resolution images and the corrected high-spectral-resolution images, acquiring the geographic range of the intersection polygons, and constructing the minimum circumscribed rectangle of the intersection polygons according to the geographic range of the intersection polygons. Carrying out grid division processing on the minimum circumscribed rectangles of the intersected polygons to obtain a plurality of blocks of the minimum circumscribed rectangles;
and step 3: mapping each block of the minimum circumscribed rectangle to the corrected high-spatial-resolution remote sensing image to obtain each corrected high-spatial-resolution remote sensing block image, mapping each block of the minimum circumscribed rectangle to the corrected high-spectral-resolution remote sensing image to obtain each corrected high-spectral-resolution remote sensing block image, and performing initial matching processing on each corrected high-spatial-resolution remote sensing block image and each corresponding corrected high-spectral-resolution remote sensing block image to obtain multiple groups of homonymy point pairs of each block;
and 4, step 4: an affine transformation model is built according to a plurality of groups of homonymous point pairs of a plurality of blocks, and the corrected high-spectral-resolution image is subjected to affine transformation according to the affine transformation model to obtain an affine-transformed high-spectral-resolution image;
and 5: calculating the updated intersection polygon of the corrected high spatial resolution image and the affine-transformed high spectral resolution image, acquiring the geographical range of the updated intersection polygon, and constructing the minimum circumscribed rectangle of the updated intersection polygon according to the geographical range of the updated intersection polygon. Carrying out grid division processing on the minimum circumscribed rectangle of the updated intersected polygon to obtain a plurality of blocks of the minimum circumscribed rectangle;
step 6: mapping each block of the minimum external rectangle to the corrected high-spatial-resolution remote sensing image to obtain each corrected high-spatial-resolution remote sensing block image, mapping each block of the minimum external rectangle to the affine-transformed high-spectral-resolution remote sensing image to obtain each corrected high-spectral-resolution remote sensing block image, and performing fine matching processing on each corrected high-spatial-resolution remote sensing block image and each corresponding affine-transformed high-spectral-resolution remote sensing block image to obtain multiple groups of homonymy point pairs of each block;
and 7: constructing a polynomial transformation model of each block according to the multiple groups of homonymy point pairs of each block, and solving parameters of the polynomial transformation model of each sub-block by combining the multiple groups of homonymy points of each sub-block by adopting a least square method to obtain the polynomial transformation model of each block;
and 8: and carrying out quadratic polynomial correction on the affine-transformed high-spectral-resolution remote sensing image according to the quadratic polynomial transformation model of each block to obtain the high-spectral-resolution remote sensing block image corrected by each quadratic polynomial.
And step 9: fusing each corrected high-spatial-resolution remote sensing block image and the corresponding high-spectral-resolution remote sensing block image corrected by the quadratic polynomial to obtain a remote sensing image fused with high spatial resolution and high spectral resolution;
preferably, step 2 calculates an intersecting polygon of the corrected high spatial resolution image and the corrected high spectral resolution image, and the specific process is as follows:
constructing a corrected high-spatial-resolution image vector polygon according to the corrected four-corner geographical coordinates of the high-spatial-resolution image;
constructing a corrected high-spectral-resolution image vector polygon according to the corrected four-corner geographical coordinates of the high-spectral-resolution image;
and calculating by a space intersection calculation method to obtain an intersection polygon between the corrected high-spatial-resolution image vector polygon and the corrected high-spectral-resolution image vector polygon.
Step 2, acquiring the geographic range of the intersected polygon, and constructing the minimum circumscribed rectangle of the intersected area according to the geographic range of the intersected polygon, wherein the specific process is as follows:
sequentially acquiring the longitude of each point and the latitude of each point in the intersected polygon;
and (3) screening the longitude minimum value of each point in the intersected polygon and the longitude maximum value of each point in the intersected polygon from the longitudes of the points in the intersected polygon, wherein the longitude range for constructing the intersected polygon is defined as follows:
[FMIN_LON，FMAX_LON]
wherein FMIN _ LON represents a longitude minimum for each point in the intersecting polygon and FMAX _ LON represents a longitude maximum for each point in the intersecting polygon;
screening out the latitude minimum value of each point in the intersected polygon and the dimension maximum value of each point in the intersected polygon from the latitudes of the points in the intersected polygon, and constructing the latitude range of the intersected polygon, wherein the latitude range is defined as follows:
[FMIN_LAT，FMAX_LAT]
wherein FMIN _ LAT represents the latitude minimum value of each point in the intersecting polygon, and FMAX _ LAT represents the latitude maximum value of each point in the intersecting polygon;
(FMAX _ LON, FMlN _ LAT), (FMAX _ LON, FMAX _ LAT), (FMAX _ LON, FMIN _ LAT) are taken as the four vertices of the minimum bounding rectangle of the intersecting polygon to construct the minimum bounding rectangle of the intersecting polygon.
Step 2, grid division processing is performed on the minimum circumscribed rectangle of the intersected polygons to obtain a plurality of blocks of the minimum circumscribed rectangle, which specifically comprises the following steps:
setting the division number of a longitudinal direction grid network as nFBK _ lon and setting the division number of a latitudinal direction grid network as nFBK _ lat;
calculating the longitude width of each minimum circumscribed rectangle block and the latitude height of each minimum circumscribed rectangle block, specifically as follows:
wherein Δ F _ LON is the longitude width of each minimum circumscribed rectangle block, Δ F _ LAT is the latitude height of each minimum circumscribed rectangle block, nFBK _ LON is the longitude grid division number, nFBK _ LAT is the set latitude grid division number, FMIN _ LON represents the longitude minimum value of each point in the intersected polygon, FMAX _ LON represents the longitude maximum value of each point in the intersected polygon, FMIN _ LAT represents the latitude minimum value of each point in the intersected polygon, and FMAX _ LAT represents the latitude maximum value of each point in the intersected polygon;
the minimum circumscribed rectangle block of the jth row and the ith column is defined as:
FBK i，j
i∈[1，nFBK_lon]，j∈[1，nFBK_lat]
wherein i represents the serial number of the blocks in the longitude direction, j represents the serial number of the blocks in the latitude direction, nFBK _ lon is the division number of the directional grid network, and nFBK _ lat is the division number of the grid network in the latitude direction
Sequentially calculating the longitude minimum value, the longitude maximum value, the latitude minimum value and the latitude maximum value of the minimum circumscribed rectangle block of the jth row and the ith column, wherein the method specifically comprises the following steps:
wherein the content of the first and second substances,
represents the minimum longitude of the minimum bounding rectangle block of the jth row and ith column,
represents the maximum value of the longitude of the minimum bounding rectangle block of the jth row and ith column,
represents the minimum value of the latitude of the smallest circumscribed rectangular block of the jth row and ith column,
the maximum value of the latitude of the minimum circumscribed rectangle block of the jth row and the ith column is represented;
FMIN _ LON represents the longitude minimum for each point in the intersecting polygon, and FMAX _ LON represents the longitude maximum for each point in the intersecting polygon; wherein FMIN _ LAT represents the latitude minimum value of each point in the intersecting polygon, and FMAX _ LAT represents the latitude maximum value of each point in the intersecting polygon;
will be provided with
The geographic rectangular blocks are used as four vertexes to construct the geographic rectangular blocks corresponding to the minimum circumscribed rectangular blocks of the jth row and the ith column;
preferably, the mapping to the corrected high spatial resolution remote sensing image in step 3 obtains each corrected high spatial resolution remote sensing block image, and the specific process is as follows:
partitioning four vertexes according to the minimum circumscribed rectangle of the jth row and the ith column
And acquiring image content corresponding to the geographical range from the corrected high-spatial-resolution remote sensing image.
The corrected high spatial resolution remote sensing block image of the jth row and ith column is defined as:
HSPA_FBK i，j
i∈[1，nFBK_lon]，j∈[1，nFBK_lat]
wherein i represents the block number in the longitude direction, j represents the block number in the latitude direction, nFBK _ lon is the division number of the directional grid network, and nFBK _ lat is the division number of the latitudinal grid network.
And 3, mapping to the corrected high-spectral-resolution remote sensing image to obtain each high-spectral-resolution remote sensing block image, wherein the specific process is as follows:
partitioning four vertexes according to the minimum circumscribed rectangle of the jth row and the ith column
The geographical range of the composition is obtained from the corrected remote sensing image with high spectral resolutionAnd taking the image content corresponding to the geographic range.
The corrected high spectral resolution remote sensing block image of the jth row and the ith column is defined as:
HSPE_FBK i，j
i∈[1，nFBK_lon]，j∈[1，nFBK_lat]
wherein, i represents the serial number of the blocks in the longitude direction, j represents the serial number of the blocks in the latitude direction, nFBK _ lon is the division number of the grid network in the direction, and nFBK _ lat is the division number of the grid network in the latitude direction;
and 3, obtaining a plurality of groups of homonymy point pairs of each block through initial matching processing, wherein the specific process is as follows:
corrected high spatial resolution remote sensing blocked image HSPA _ FBK of jth row and ith column i，j And the corrected high-spectral-resolution remote sensing blocked image HSPE _ FBK of the ith row and the ith column in the corresponding j i，j Acquiring the corrected high spatial resolution remote sensing blocked image HSPA _ FBK of the jth row and ith column by adopting an SIFT + KNN matching algorithm i，j High-spectral-resolution remote sensing block image HSPE _ FBK corrected by pixel coordinates and ith row and ith column of jth row i，j Pixel coordinates.
Obtaining the minimum longitude value of the jth row and ith column sub-blocks
Minimum latitude
Correcting the jth row and ith column of the high-spatial-resolution remote sensing blocked image HSPA _ FBK i，j The pixel coordinates are mapped into high spatial resolution remote sensing block image geographical coordinates in a mode of the sum of the product of the pixel coordinates and the resolution and the minimum value;
obtaining the minimum longitude value of the jth row and ith column sub-block
Minimum latitude
Correcting the jth row and ith columnHigh-spectral-resolution remote sensing block image HSPE _ FBK i，j The pixel coordinates are mapped into geographical coordinates according to the mode of the sum of the product of the pixel coordinates and the resolution and the minimum value;
and combining the geographic coordinates of the high-spatial-resolution remote sensing block images and the geographic coordinates of the high-spectral-resolution remote sensing block images corresponding to the sub-blocks in the ith row and the ith column with a plurality of groups of homonymous point pairs in the jth row and the ith column.
Preferably, in step 4, the affine transformation model is constructed according to the plurality of groups of same-name point pairs of the plurality of partitions, and the specific process is as follows:
and constructing a global affine transformation model according to the geographic coordinates of the high-spatial-resolution remote sensing block images and the geographic coordinates of the high-spectral-resolution remote sensing block images corresponding to all block homonymous point pairs, and solving parameters of the affine transformation model by adopting a least square method.
And 4, carrying out affine transformation processing on the corrected high-spectral-resolution image according to an affine transformation model, wherein the specific process is as follows:
incorporating affine transformation model parameters (e) 0 ，e 1 ，e 2 ，f 0 ，f 1 ，f 2 ) And performing affine transformation correction on the corrected high-spectral-resolution image by adopting an indirect correction method to obtain the high-spectral-resolution image with affine transformation.
Preferably, the step 5 of calculating the intersection polygon after updating the corrected high spatial resolution image and the affine transformed high spectral resolution image includes the following specific steps:
and constructing a corrected high-spatial-resolution image vector polygon according to the corrected four-corner-point geographic coordinates of the high-spatial-resolution image. And constructing an affine-transformed high-spectral-resolution image vector polygon according to the geographical coordinates of the four corner points of the affine-transformed high-spectral-resolution image. And calculating an intersection polygon between the corrected high-spatial-resolution image vector polygon and the affine-transformed high-spectral-resolution image vector polygon by using a spatial intersection calculation method, namely the updated intersection polygon.
Step 5, obtaining the geographical range of the updated intersected polygon, and constructing the minimum circumscribed rectangle of the intersected area according to the geographical range of the updated intersected polygon, wherein the specific process is as follows:
sequentially acquiring the longitude of each point and the latitude of each point in the updated intersected polygon;
screening out the longitude minimum value of each point in the updated intersected polygon and the longitude maximum value of each point in the updated intersected polygon from the longitudes of the points in the updated intersected polygon, and constructing the longitude range of the updated intersected polygon as follows:
[SMIN_LON，SMAX_LON]
wherein SMIN _ LON represents the longitude minimum value of each point in the updated intersected polygon, and SMAX _ LON represents the longitude maximum value of each point in the updated intersected polygon;
screening out the latitude minimum value of each point in the updated intersected polygon and the latitude maximum value of each point in the updated intersected polygon from the latitudes of the points in the updated intersected polygon, and constructing the latitude range of the updated intersected polygon, wherein the latitude range is defined as follows:
[SMIN_LAT，SMAX_LAT]
wherein SMIN _ LAT represents the latitude minimum value of each point in the updated intersected polygon, and SMAX _ LAT represents the latitude maximum value of each point in the updated intersected polygon;
(SMAX _ LON, SMIN _ LAT), (SMAX _ LON, SMAX _ LAT), (SMAX _ LON, SMIN _ LAT) are taken as the four vertices of the minimum bounding rectangle of the updated intersected polygon to construct the minimum bounding rectangle of the updated intersected polygon.
Step 5, grid division processing is performed on the minimum external rectangle of the updated intersected polygon to obtain a plurality of blocks of the minimum external rectangle, which specifically includes:
setting the division number of the longitudinal direction grid as nSBK _ lon and setting the division number of the latitudinal direction grid as nSBK _ lat;
calculating the longitude width of each minimum circumscribed rectangle block and the latitude height of each minimum circumscribed rectangle block, specifically as follows:
wherein, Δ S _ LON is the longitude width of each minimum circumscribed rectangle block, Δ S _ lat is the latitude height of each minimum circumscribed rectangle block, nSBK _ LON is the division number of the grid in the longitude direction, nSBK _ lat is the division number of the grid in the set latitude direction, SMIN _ LON represents the longitude minimum value of each point in the intersection polygon after updating, SMAX _ LON represents the longitude maximum value of each point in the intersection polygon after updating; wherein SMIN _ LAT represents the latitude minimum value of each point in the updated intersected polygon, and SMAX _ LAT represents the latitude maximum value of each point in the updated intersected polygon;
the minimum circumscribed rectangle block of the qth row and the pth column is defined as:
SBK p，q
p∈[1，nSBK_lon]，q∈[1，nSBK_lat]
wherein p represents the block serial number in the longitude direction, q represents the block serial number in the latitude direction, nSBK _ lon is the division number of the directional grid, nSBK _ lat is the division number of the grid in the latitude direction
Sequentially calculating the longitude minimum value, the longitude maximum value, the latitude minimum value and the latitude maximum value of the minimum circumscribed rectangle block of the qth row and the pth column, specifically as follows:
wherein the content of the first and second substances,
represents the minimum longitude of the smallest circumscribed rectangular block of the qth row and the pth column,
represents the maximum value of the longitude of the minimum bounding rectangle block of the qth row and the pth column,
represents the minimum value of latitude of the minimum circumscribed rectangle block of the qth row and the pth column,
represents the latitude maximum of the minimum bounding rectangle block of the qth row and the pth column,
SMIN _ LON represents the longitude minimum value of each point in the updated intersected polygon, and SMAX _ LON represents the longitude maximum value of each point in the updated intersected polygon; wherein SMIN _ LAT represents the latitude minimum value of each point in the updated intersected polygon, and SMAX _ LAT represents the latitude maximum value of each point in the updated intersected polygon; will be provided with
The geographic rectangular blocks are used as four vertexes to construct the geographic rectangular blocks corresponding to the minimum circumscribed rectangular blocks of the qth row and the pth column;
preferably, the mapping to the corrected high spatial resolution remote sensing image in step 6 obtains each corrected high spatial resolution remote sensing block image, and the specific process is as follows:
partitioning four vertexes according to the minimum circumscribed rectangle of the q-th row and the p-th column
And acquiring image content corresponding to the geographical range from the corrected high-spatial-resolution remote sensing image.
The corrected high spatial resolution remote sensing blocked image of the q-th row and the p-th column is defined as follows:
HSPA_SBK p，q
p∈[1，nSBK_lon]，q∈[1，nSBK_lat]
wherein p represents the serial number of the blocks in the longitude direction, q represents the serial number of the blocks in the latitude direction, nSBK _ lon is the division number of the latticed network in the longitude direction, and nSBK _ lat is the division number of the latticed network in the latitude direction.
And 6, mapping the image to the affine transformed high-spectral-resolution remote sensing image to obtain each corrected high-spatial-resolution remote sensing block image, wherein the specific process is as follows:
partitioning four vertexes according to the minimum circumscribed rectangle of the q-th row and the p-th column
And acquiring image content corresponding to the geographical range from the corrected high-spatial-resolution remote sensing image.
The affine transformation high spatial resolution remote sensing block image of the q-th row and the p-th column is defined as follows:
HSPE_SBK p，q
p∈[1，nSBK_lon]，q∈[1，nSBK_lat]
wherein p represents the serial number of the blocks in the longitude direction, q represents the serial number of the blocks in the latitude direction, nSBK _ ion is the division number of the direction grid network, and nSBK lat is the division number of the latitude grid network.
And 6, obtaining a plurality of groups of same-name point pairs of each block through fine matching processing, wherein the specific process is as follows:
corrected high spatial resolution remote sensing blocked image HSPA _ SBK of q row and p column p，q The validity determination is specifically performed as follows:
if CNT _ HSPA _ SBK p，q Greater than or equal to alpha, then
If CNT _ HSPA _ SBK p，q < alpha, then
Wherein CNT _ HSPA _ SBK p，q Indicating corrected high spatial resolution of the qth row and the pth columnRemote sensing blocking image HSPA _ SBK p，q The number of pixel values greater than 0, alpha represents a block validity determination threshold,
and representing the validity attribute of the high-spatial-resolution remote sensing blocked image corrected by the qth row and the pth column.
If it is
The corrected high spatial resolution remote sensing blocked image HSPA _ SBK of the q-th row and the p-th column p，q And a high-spectral-resolution remote sensing blocked image HSPE-SBK of affine transformation of the p-th column of the corresponding q rows p，q No subsequent matching is performed.
If it is
The corrected high spatial resolution remote sensing blocked image HSPA _ SBK of the q-th row and the p-th column p，q And a high-spectral-resolution remote sensing blocked image HSPE _ SBK of affine transformation of the p-th column of the corresponding q rows p，q Matching is carried out, specifically as follows:
corrected high spatial resolution remote sensing blocked image HSPA _ SBK of p row and p column of q rows p，q And a high-spectral-resolution remote sensing blocked image HSPE _ SBK of affine transformation of the p-th column of the corresponding q rows p，q Acquiring a corrected high-spatial-resolution remote sensing blocked image HSPA _ SBK of the p-th row of the q-th line by adopting an SIFT + RANSAC matching algorithm p，q High-spectral-resolution remote sensing block image HSPE _ SBK corrected by pixel coordinates and p-th row and p-th column p，q Pixel coordinates.
And combining the pixel coordinates of the high-spatial-resolution remote sensing block images corresponding to the q-th row and p-th column blocks and the pixel coordinates of the affine-transformed high-spatial-resolution remote sensing block images into a plurality of groups of homonymy point pairs of the q-th row and p-th column blocks.
Preferably, in step 7, the polynomial transformation model for constructing each partition according to the plurality of groups of same-name point pairs of each partition is specifically as follows:
obtaining corrected high spatial resolution remote sensing blocks of the q-th row and p-th columnImage HSPA _ SBK p，q Properties
If it is
Directly adopting the corrected high spatial resolution remote sensing blocked image HSPA _ SBK of the q row and the p column p，q And a high-spectral-resolution remote sensing blocked image HSPE _ SBK of affine transformation of the p-th column of the corresponding q rows p，q The matched pairs of homonymous points are used for solving the polynomial model.
If it is
Sequentially judging corrected high-spatial-resolution remote sensing blocked images HSPA _ SBK of the q-th row and the p-1 th column p-1， q attribute
Corrected high-spatial-resolution remote sensing blocked image HSPA _ SBK of q row and p +1 column p+1，q Properties
Corrected high-spatial-resolution remote sensing blocked image HSPA _ SBK of q-1 row and p column p，q-1 Property, property
Corrected high-spatial-resolution remote sensing blocked image HSPA _ SBK of q +1 row and p column p，q+1 Properties
If the number of the sub-blocks is 1, matching a plurality of groups of homonymous point pairs with the corresponding sub-blocks to serve as a plurality of groups of homonymous point pairs of the current qth row and the pth column sub-blocks;
the polynomial transformation model of the block of the qth row and the block of the pth column is specifically constructed as follows:
obtaining corrected high-spatial-resolution remote sensing blocked image HSPA _ SBK of the q-th row and the p-th column p，q MatchingH group of pixel coordinates
h∈[1，CT p，q ]ph∈[1，nSBK_lon]，q h ∈[1，nSBK_lat]
Wherein, the first and the second end of the pipe are connected with each other,
corrected high spatial resolution remote sensing blocked image HSPA _ SBK representing q-th row and p-th column p，q And q rows and p columns of affine transformation high-spectral-resolution remote sensing block image HSPE _ SBK p，q Number of matched homonyms. h represents the serial number, nSBK _ lon is the division number of the grid in the longitude direction, and nSBK _ lat is the division number of the grid in the latitude direction.
Obtaining a high-spectral-resolution remote sensing blocked image HSPE _ SBK of affine transformation of a q-th row and a p-th column p，q Matched h-th group of pixel coordinates
h∈[1，CT p，q ]p h ∈[1，nSBK_lon]，q h ∈[1，nSBK_lat]
Wherein the content of the first and second substances,
corrected high spatial resolution remote sensing blocked image HSPA _ SBK for expressing q row and p column p，q And q rows and p columns of affine transformation high-spectral-resolution remote sensing blocked image HSPE _ SBK p，q Number of matched homonyms. h represents the serial number, nSBK _ lon is the division number of the grid in the longitude direction, and nSBK _ lat is the division number of the grid in the latitude direction.
Sequentially correcting the high-spatial-resolution remote sensing blocked image HSPA _ SBK according to the q-th row and the p-th column p，q Matched h-th group pixel coordinates
And affine transformation high-spectral resolution remote sensing blocked image HSPE _ SBK of the qth row and the pth column p，q Matched h-th group of pixel coordinates
And (3) constructing an h group of quadratic polynomials of the p column of the q row so as to establish a quadratic polynomial model of the whole block of the p column of the q row, and solving corresponding quadratic polynomial model parameters by adopting least square.
Preferably, the step 9 of performing fusion processing to obtain the remote sensing image with high spatial resolution and high spectral resolution fused specifically includes:
setting the weight of blue band as B w Green band weight of G w Red band weight of R w The weight of the near infrared band is IR w 。
Obtaining pixel values of each corrected high-spatial-resolution remote sensing block image and corresponding high-spectral-resolution remote sensing block images corrected by quadratic polynomial pixel by pixel, and calculating blue-green-red three-band adjustment values ADJ corresponding to each pixel value 1 Near infrared adjustment value ADJ corresponding to each pixel value 2 The calculation method is as follows:
wherein, HSPA DN Representing corresponding panchromatic pixel values, ADJ, on a corrected high spatial resolution remote sensing tile image 1 Adjust values for each pixel value for three bands of cyan and red, ADJ 2 For a near infrared adjustment value, AVGW, corresponding to each pixel value 1 AVGW being weighted average of red, green and blue bands corresponding to each pixel 2 The weighted average of the red band, the green band, the blue band and the near-infrared band corresponding to each pixel is calculated in the following way,
wherein, B DN High-spectral-resolution remote sensing block image with blue-band pixel value G representing quadratic polynomial correction DN Representing green band pixel value, R, in a quadratic polynomial corrected high spectral resolution remote sensing block image DN Representing red band pixel value, IR in high spectral resolution remote sensing block image corrected by quadratic polynomial DN And expressing the pixel value of the near infrared band in the high-spectral-resolution remote sensing block image corrected by the quadratic polynomial.
The adjustment value ADJ of blue, green and red three bands corresponding to each pixel value 1 Near infrared adjustment value ADJ corresponding to each pixel value 2 Calculating the fused image value pixel by pixel, specifically as follows:
wherein, B FUS Representing the blue band pixel value, G, of the fused image FUS Representing the green band pixel value, R, of the fused image FUS Indicating the red band pixel value, IR of the fused image FUS And representing the pixel value of the fused image near infrared band.
The invention also provides a computer readable medium storing a computer program for execution by an electronic device, which, when run on the electronic device, causes the electronic device to perform the steps of the method for remote sensing image fusion taking into account image distortion.
The method has the advantages that based on the initial registration and the block fine registration strategy, the polynomial correction is independently adopted for each block, the problem of poor fusion effect caused by distortion in the image can be effectively solved, and the fusion effect is improved. In addition, the blocking idea has small requirement on the memory overhead, and the fusion between the remote sensing images with large width can be realized.
Drawings
FIG. 1: a method flowchart of an embodiment of the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, 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 invention.
In specific implementation, a person skilled in the art can implement the automatic operation process by using a computer software technology, and a system device for implementing the method, such as a computer-readable storage medium storing a corresponding computer program according to the technical solution of the present invention and a computer device including a corresponding computer program for operating the computer program, should also be within the scope of the present invention.
The technical scheme of the embodiment of the invention is introduced in combination with fig. 1, which is a remote sensing image fusion method considering image distortion, and specifically comprises the following steps:
step 1: acquiring an original high-spatial-resolution remote sensing image and a corresponding RPC positioning model thereof, and carrying out geometric correction to obtain a corrected high-spatial-resolution remote sensing image; acquiring a high-spectral-resolution remote sensing image and a corresponding RPC positioning model thereof, and performing geometric correction to obtain a corrected high-spectral-resolution remote sensing image;
step 2: calculating the intersection polygons of the corrected high-spatial-resolution images and the corrected high-spectral-resolution images, acquiring the geographic range of the intersection polygons, and constructing the minimum circumscribed rectangle of the intersection polygons according to the geographic range of the intersection polygons. Carrying out grid division processing on the minimum external rectangle of the intersected polygon to obtain a plurality of blocks of the minimum external rectangle;
step 2, calculating an intersecting polygon of the corrected high-spatial-resolution image and the corrected high-spectral-resolution image, wherein the specific process is as follows:
constructing a corrected high-spatial-resolution image vector polygon according to the corrected four-corner geographical coordinates of the high-spatial-resolution image;
constructing a corrected high-spectral-resolution image vector polygon according to the corrected four-corner geographical coordinates of the high-spectral-resolution image;
and calculating by a space intersection calculation method to obtain an intersection polygon between the corrected high-spatial-resolution image vector polygon and the corrected high-spectral-resolution image vector polygon.
Step 2, acquiring the geographic range of the intersected polygon, and constructing the minimum circumscribed rectangle of the intersected area according to the geographic range of the intersected polygon, wherein the specific process is as follows:
sequentially acquiring the longitude of each point and the latitude of each point in the intersected polygon;
and (3) screening the longitude minimum value of each point in the intersected polygon and the longitude maximum value of each point in the intersected polygon from the longitudes of the points in the intersected polygon, wherein the longitude range for constructing the intersected polygon is defined as follows:
[FMIN_LON，FMAX_LON]
wherein FMIN _ LON represents the minimum longitude value of each point in the intersecting polygon, FMAX _ LON represents the maximum longitude value of each point in the intersecting polygon:
screening out the latitude minimum value of each point in the intersected polygon and the dimension maximum value of each point in the intersected polygon from the latitudes of the points in the intersected polygon, and constructing the latitude range of the intersected polygon, wherein the latitude range is defined as follows:
[FMIN_LAT，FMAX_LAT]
wherein FMIN _ LAT represents the latitude minimum value of each point in the intersecting polygon, and FMAX _ LAT represents the latitude maximum value of each point in the intersecting polygon;
(FMAX _ LON, FMIN _ LAT), (FMAX _ LON, FMAX _ LAT), (FMAX _ LON, FMIN _ LAT) are taken as the four vertices of the minimum bounding rectangle of the intersecting polygon to construct the minimum bounding rectangle of the intersecting polygon.
Step 2, grid division processing is performed on the minimum circumscribed rectangle of the intersected polygons to obtain a plurality of blocks of the minimum circumscribed rectangle, which specifically comprises the following steps:
setting the division number of the longitudinal direction grid mesh as nFBK _ lon and setting the division number of the latitudinal direction grid mesh as nFBK _ lat;
calculating the longitude width of each minimum circumscribed rectangle block and the latitude height of each minimum circumscribed rectangle block, specifically as follows:
where Δ F _ LON is the longitude width of each minimum bounding rectangle tile, Δ F _ LAT is the latitude height of each minimum bounding rectangle tile, nFBK _ LON =5 is the longitude grid division number, nFBK _ LAT =5 is the set latitude grid division number, FMIN _ LON represents the longitude minimum for each point in the intersecting polygon, FMAX _ LON represents the longitude maximum for each point in the intersecting polygon, where FMIN _ LAT represents the latitude minimum for each point in the intersecting polygon, and FMAX _ LAT represents the latitude maximum for each point in the intersecting polygon;
the minimum circumscribed rectangle block of the jth row and the ith column is defined as:
FBK i，j
i∈[1，nFBK_lon]，j∈[1，nFBK_lat]
wherein i represents the block number in the longitude direction, j represents the block number in the latitude direction, nFBK _ lon is the division number of the direction grid network, and nFBK _ lat is the division number of the grid network in the latitude direction
Sequentially calculating the longitude minimum value, the longitude maximum value, the latitude minimum value and the latitude maximum value of the minimum circumscribed rectangle block of the jth row and the ith column, wherein the method specifically comprises the following steps:
wherein the content of the first and second substances,
represents the minimum longitude of the smallest bounding rectangle block of the jth row and ith column,
representing blocks of the smallest circumscribed rectangle of the jth row and ith columnThe maximum value of the longitude is used,
represents the minimum value of the dimension of the minimum bounding rectangle block of the jth row and ith column,
the maximum value of the latitude of the minimum circumscribed rectangle block of the jth row and the ith column is represented;
FMIN _ LON represents the longitude minimum for each point in the intersecting polygon, and FMAX _ LON represents the longitude maximum for each point in the intersecting polygon; wherein FMIN _ LAT represents the latitude minimum value of each point in the intersecting polygon, and FMAX _ LAT represents the latitude maximum value of each point in the intersecting polygon;
will be provided with
The geographic rectangular blocks are used as four vertexes to construct the geographic rectangular blocks corresponding to the minimum circumscribed rectangular blocks of the jth row and the ith column;
and 3, step 3: mapping each block of the minimum circumscribed rectangle to the corrected high-spatial-resolution remote sensing image to obtain each corrected high-spatial-resolution remote sensing block image, mapping each block of the minimum circumscribed rectangle to the corrected high-spectral-resolution remote sensing image to obtain each corrected high-spectral-resolution remote sensing block image, and performing initial matching processing on each corrected high-spatial-resolution remote sensing block image and each corresponding corrected high-spectral-resolution remote sensing block image to obtain multiple groups of homonymy point pairs of each block;
and 3, mapping to the corrected high spatial resolution remote sensing images to obtain each corrected high spatial resolution remote sensing block image, wherein the specific process is as follows:
partitioning four vertexes according to the minimum circumscribed rectangle of the jth row and the ith column
And acquiring image content corresponding to the geographical range from the corrected high-spatial-resolution remote sensing image.
The corrected high spatial resolution remote sensing block image of the jth row and ith column is defined as:
HSPA_FBK i，j
i∈[1，nFBK_lon]，j∈[1，nFBK_lat]
wherein i represents a block number in the longitude direction, j represents a block number in the latitude direction, nFBK _ lon represents the number of division of the directional grid network, and nFBK _1at represents the number of division of the latitudinal grid network.
And 3, mapping to the corrected high-spectral-resolution remote sensing image to obtain each high-spectral-resolution remote sensing block image, wherein the specific process is as follows:
partitioning four vertexes according to the minimum circumscribed rectangle of the jth row and the ith column
And acquiring image content corresponding to the geographical range from the corrected remote sensing image with the high spectral resolution.
The corrected high spectral resolution remote sensing block image of the jth row and the ith column is defined as:
HSPE_FBK i，j
i∈[1，nFBK_lon]，j∈[1，nFBK_lat]
wherein i represents the serial number of the blocks in the longitude direction, j represents the serial number of the blocks in the latitude direction, nFBK _ lon is the division number of the directional grid network, and nFBK _ lat is the division number of the grid network in the latitude direction;
and 3, obtaining a plurality of groups of homonymy point pairs of each block through initial matching processing, wherein the specific process is as follows:
corrected high spatial resolution remote sensing blocked image HSPA _ FBK of jth row and ith column i，j And the corresponding secondCorrected high-spectral-resolution remote sensing block image HSPE _ FBK of j rows and i columns i，j Acquiring the corrected high-spatial-resolution remote sensing blocked image HSPA _ FBK of the jth row and ith column by adopting an SIFT + KNN matching algorithm i，j High-spectral-resolution remote sensing block image HSPE _ FBK corrected by pixel coordinates and ith row and ith column of jth row i，j Pixel coordinates.
Obtaining the minimum longitude value of the jth row and ith column sub-block
Minimum value of latitude
Correcting the jth row and ith column of the high-spatial-resolution remote sensing blocked image HSPA _ FBK i，j The pixel coordinates are mapped into high-spatial-resolution remote sensing block image geographical coordinates in a mode of the sum of the pixel coordinates, the resolution product and the minimum value;
obtaining the minimum longitude value of the jth row and ith column sub-block
Minimum latitude
High-spectral-resolution remote sensing blocked image HSPE _ FBK corrected in jth row and ith column i，j The pixel coordinates are mapped into geographical coordinates according to the mode of the sum of the product of the pixel coordinates and the resolution and the minimum value;
and combining the geographic coordinates of the high-spatial-resolution remote sensing block images and the geographic coordinates of the high-spectral-resolution remote sensing block images corresponding to the sub-blocks in the ith row and the ith column with a plurality of groups of homonymous point pairs in the jth row and the ith column.
And 4, step 4: an affine transformation model is built according to a plurality of groups of homonymous point pairs of a plurality of blocks, and the corrected high-spectral-resolution image is subjected to affine transformation according to the affine transformation model to obtain an affine-transformed high-spectral-resolution image;
step 4, establishing an affine transformation model according to the multiple groups of homonymous point pairs of the multiple blocks, which specifically comprises the following steps:
and constructing a global affine transformation model according to the geographic coordinates of the high-spatial-resolution remote sensing block images and the geographic coordinates of the high-spectral-resolution remote sensing block images corresponding to all block homonymous point pairs, and solving parameters of the affine transformation model by adopting a least square method.
And 4, carrying out affine transformation processing on the corrected high-spectral-resolution image according to an affine transformation model, wherein the specific process is as follows:
incorporating affine transformation model parameters (e) 0 ，e 1 ，e 2 ，f 0 ，f 1 ，f 2 ) And performing affine transformation correction on the corrected high-spectral-resolution image by adopting an indirect correction method to obtain the high-spectral-resolution image with affine transformation.
And 5: calculating the updated intersection polygon of the corrected high spatial resolution image and the affine-transformed high spectral resolution image, acquiring the geographical range of the updated intersection polygon, and constructing the minimum circumscribed rectangle of the updated intersection polygon according to the geographical range of the updated intersection polygon. Carrying out grid division processing on the minimum circumscribed rectangle of the updated intersected polygon to obtain a plurality of blocks of the minimum circumscribed rectangle;
step 5, calculating the intersection polygon after the corrected high spatial resolution image and the affine transformation high spectral resolution image are updated, and the specific process is as follows:
and constructing a corrected high-spatial-resolution image vector polygon according to the corrected four-corner-point geographic coordinates of the high-spatial-resolution image. And constructing an affine-transformed high-spectral-resolution image vector polygon according to the geographical coordinates of the four corner points of the affine-transformed high-spectral-resolution image. And calculating an intersection polygon between the corrected high-spatial-resolution image vector polygon and the affine-transformed high-spectral-resolution image vector polygon by using a spatial intersection calculation method, namely the updated intersection polygon.
Step 5, obtaining the updated geographic range of the intersected polygon, and constructing the minimum circumscribed rectangle of the intersected area according to the updated geographic range of the intersected polygon, wherein the specific process is as follows:
sequentially acquiring the longitude of each point and the latitude of each point in the updated intersected polygon;
screening out the longitude minimum value of each point in the updated intersected polygon and the longitude maximum value of each point in the updated intersected polygon from the longitudes of the points in the updated intersected polygon, and constructing the longitude range of the updated intersected polygon as follows:
[SMIN_LON，SMAX_LON]
wherein SMIN _ LON represents the longitude minimum value of each point in the updated intersected polygon, and SMAX _ LON represents the longitude maximum value of each point in the updated intersected polygon;
screening out the latitude minimum value of each point in the updated intersected polygon and the latitude maximum value of each point in the updated intersected polygon from the latitudes of the points in the updated intersected polygon, and constructing the latitude range of the updated intersected polygon, wherein the latitude range is defined as follows:
[SMIN_LAT，SMAX_LAT]
wherein SMIN _ LAT represents the latitude minimum value of each point in the updated intersected polygon, and SMAX _ LAT represents the latitude maximum value of each point in the updated intersected polygon;
(SMAX _ LON, SMIN _ LAT), (SMAX _ LON, SMAX _ LAT), (SMAX _ LON, SMIN _ LAT) are taken as the four vertices of the minimum bounding rectangle of the updated intersected polygon to construct the minimum bounding rectangle of the updated intersected polygon.
Step 5, performing grid division processing on the updated minimum circumscribed rectangle of the intersected polygon to obtain a plurality of blocks of the minimum circumscribed rectangle, which specifically includes:
setting the division number of the longitudinal direction grid as nSBK _ lon and setting the division number of the latitudinal direction grid as nSBK _ lat;
calculating the longitude width of each minimum circumscribed rectangle block and the latitude height of each minimum circumscribed rectangle block, specifically as follows:
wherein Δ S _ LON is the longitude width of each minimum circumscribed rectangle block, Δ S _ lat is the latitude height of each minimum circumscribed rectangle block, nSBK _ LON =5 is the division number of the grid in the longitude direction, nSBK _ lat =5 is the division number of the grid in the set latitude direction, SMIN _ LON represents the longitude minimum value of each point in the intersected polygon after updating, SMAX _ LON represents the longitude maximum value of each point in the intersected polygon after updating; wherein SMIN _ LAT represents the latitude minimum value of each point in the updated intersected polygon, and SMAX _ LAT represents the latitude maximum value of each point in the updated intersected polygon;
the minimum circumscribed rectangle block of the qth row and the pth column is defined as:
SBK p，q
p∈[1，nSBK_lon]，q∈[1，nSBK_lat]
wherein p represents the block serial number in the longitude direction, q represents the block serial number in the latitude direction, nSBK _ lon is the division number of the directional grid, nSBK _ lat is the division number of the grid in the latitude direction
Sequentially calculating the longitude minimum value, the longitude maximum value, the latitude minimum value and the latitude maximum value of the minimum circumscribed rectangle block of the qth row and the pth column, specifically as follows:
wherein the content of the first and second substances,
represents the minimum longitude of the smallest circumscribed rectangular block of the qth row and the pth column,
represents the maximum value of the longitude of the minimum bounding rectangle block of the qth row and the pth column,
represents the minimum value of latitude of the minimum circumscribed rectangle block of the qth row and the pth column,
represents the latitude maximum of the minimum bounding rectangle block of the qth row and the pth column,
SMIN _ LON represents the longitude minimum value of each point in the updated intersected polygon, and SMAX _ LON represents the longitude maximum value of each point in the updated intersected polygon; wherein SMIN _ LAT represents the latitude minimum value of each point in the updated intersected polygon, and SMAX _ LAT represents the latitude maximum value of each point in the updated intersected polygon; will be provided with
The geographic rectangular blocks are used as four vertexes to construct the geographic rectangular blocks corresponding to the minimum circumscribed rectangular blocks of the qth row and the pth column;
and 6: mapping each block of the minimum external rectangle to the corrected high-spatial-resolution remote sensing image to obtain each corrected high-spatial-resolution remote sensing block image, mapping each block of the minimum external rectangle to the affine-transformed high-spectral-resolution remote sensing image to obtain each corrected high-spectral-resolution remote sensing block image, and performing fine matching processing on each corrected high-spatial-resolution remote sensing block image and each corresponding affine-transformed high-spectral-resolution remote sensing block image to obtain multiple groups of homonymy point pairs of each block;
and 6, mapping to the corrected high-spatial-resolution remote sensing images to obtain each corrected high-spatial-resolution remote sensing block image, wherein the specific process is as follows:
partitioning four vertexes according to the minimum circumscribed rectangle of the q-th row and the p-th column
(ii) the geographical range of the composition, obtaining a map of the corresponding geographical range from the corrected high spatial resolution remote sensing imageLike the content.
The corrected high spatial resolution remote sensing block image of the q-th row and the p-th column is defined as follows:
HSPA_SBK p，q
p∈[1，nSBK_lon]，q∈[1，nSBK_lat]
wherein p represents the serial number of the blocks in the longitude direction, q represents the serial number of the blocks in the latitude direction, nSBK _ ion is the division number of the latticed network in the longitude direction, and nSBK _ lat is the division number of the latticed network in the latitude direction.
And 6, mapping the image to the affine transformed high-spectral-resolution remote sensing image to obtain each corrected high-spatial-resolution remote sensing block image, wherein the specific process is as follows:
partitioning four vertexes according to the minimum circumscribed rectangle of the q-th row and the p-th column
And acquiring image content corresponding to the geographical range from the corrected high-spatial-resolution remote sensing image.
The affine transformation high spatial resolution remote sensing block image of the q-th row and the p-th column is defined as follows:
HSPE_SBK p，q
p∈[1，nSBK_lon]，q∈[1，nSBK_lat]
wherein p represents the serial number of the blocks in the longitude direction, q represents the serial number of the blocks in the latitude direction, nSBK _ lon is the division number of the directional grid, and nSBK _ lat is the division number of the grid in the latitude direction.
And 6, obtaining a plurality of groups of same-name point pairs of each block through fine matching processing, wherein the specific process is as follows:
corrected high spatial resolution remote sensing blocked image HSPA _ SBK of q row and p column p，q The validity determination is specifically performed as follows:
if CNT _ HSPA _ SBK p，q Not less than alpha, then
If CNT _ HSPA _ SBK p，q < alpha, then
Wherein CNT _ HSPA _ SBK p，q Corrected high spatial resolution remote sensing blocked image HSPA _ SBK for expressing q row and p column p，q The number of pixel values greater than 0, α represents a block validity determination threshold value (α = 50),
and representing the validity attribute of the high-spatial-resolution remote sensing blocked image corrected by the q-th row and the p-th column.
If it is
The corrected high spatial resolution remote sensing blocked image HSPA _ SBK of the q-th row and the p-th column p，q And a high-spectral-resolution remote sensing blocked image HSPE _ SBK of affine transformation of the p-th column of the corresponding q rows p，q No subsequent matching is performed.
If it is
The corrected high spatial resolution remote sensing blocked image HSPA _ SBK of the q-th row and the p-th column p，q And a high-spectral-resolution remote sensing blocked image HSPE _ SBK of affine transformation of the p-th column of the corresponding q rows p，q Matching is carried out, specifically as follows:
corrected high spatial resolution remote sensing blocked image HSPA _ SBK of p row and p column of q rows p，q And a high-spectral-resolution remote sensing blocked image HSPE _ SBK of affine transformation of the p-th column of the corresponding q rows p，q Acquiring a corrected high-spatial-resolution remote sensing blocked image HSPA _ SBK of the p-th row of the q-th line by adopting an SIFT + RANSAC matching algorithm p，q High-spectral-resolution remote sensing block image HSPE _ SBK corrected by pixel coordinates and p-th row and p-th column p，q Pixel coordinates.
And combining the pixel coordinates of the high-spatial-resolution remote sensing block images corresponding to the q-th row and p-th column blocks and the pixel coordinates of the affine-transformed high-spectral-resolution remote sensing block images into a plurality of groups of homonymous point pairs of the q-th row and p-th column blocks.
And 7: constructing a polynomial transformation model of each block according to the multiple groups of homonymy point pairs of each block, and solving parameters of the polynomial transformation model of each sub-block by combining the multiple groups of homonymy points of each sub-block by adopting a least square method to obtain the polynomial transformation model of each block;
step 7, constructing a polynomial transformation model of each block according to the multiple groups of homonymy point pairs of each block as follows:
obtaining corrected high-spatial-resolution remote sensing blocked image HSPA _ SBK of the q-th row and the p-th column p，q Attribute
If it is
Directly adopting the corrected high-spatial-resolution remote sensing blocking image HSPA _ SBK of the q row and the p column p，q And a high-spectral-resolution remote sensing blocked image HSPE _ SBK of affine transformation of the p-th column of the corresponding q rows p，q The matched pairs of homonymous points are used for solving the polynomial model.
If it is
Sequentially judging corrected high-spatial-resolution remote sensing blocked images HSPA _ SBK of the q-th row and the p-1 th column p-1，q Properties
Corrected high spatial resolution remote sensing blocked image HSPA _ SBK of q row and p +1 column p+1，q Properties
Corrected high-spatial-resolution remote sensing blocked image HSPA _ SBK of q-1 row and p column p，q-1 Property, property
Corrected high-spatial-resolution remote sensing blocked image HSPA _ SBK of q +1 row and p column p，q+1 Attribute
If the number of the sub-blocks is 1, matching a plurality of groups of homonymous point pairs with the corresponding sub-blocks to serve as a plurality of groups of homonymous point pairs of the current qth row and the pth column sub-blocks;
the polynomial transformation model of the block of the qth row and the block of the pth column is specifically constructed as follows:
obtaining corrected high-spatial-resolution remote sensing blocked image HSPA _ SBK of the q-th row and the p-th column p，q Matched h-th group of pixel coordinates
h∈[1，CT p，q ]ph∈[1，nSBK_lon]，qh∈[1，nSBK_lat]
Wherein the content of the first and second substances,
corrected high spatial resolution remote sensing blocked image HSPA _ SBK representing q-th row and p-th column p，q And q rows and p columns of affine transformation high-spectral-resolution remote sensing block image HSPE _ SBK p，q Number of matched homonyms. h represents the serial number, nSBK _ ion is the number of divisions of the longitudinal grid, and nSBK _ lat is the number of divisions of the latitudinal grid.
Obtaining a high-spectral-resolution remote sensing blocked image HSPE _ SBK of affine transformation of a q-th row and a p-th column p，q Matched h-th group of pixel coordinates
h∈[1，CT p，q ]p h ∈[1，nSBK_lon]，q h ∈[1，nSBK_lat]
Wherein the content of the first and second substances,
corrected high spatial resolution remote sensing blocked image HSPA _ SBK for expressing q row and p column p，q And q rows and p columns of affine transformation high-spectral-resolution remote sensing block image HSPE _ SBK p，q Number of matched homonyms. h represents the serial number, nSBK _ lon is the division number of the grid in the longitude direction, and nSBK _ lat is the division number of the grid in the latitude direction.
Sequentially correcting the high-spatial-resolution remote sensing blocked image HSPA _ SBK according to the q-th row and the p-th column p，q Matched h-th group of pixel coordinates
And a q-th row and p-th column affine transformation high-spectral-resolution remote sensing blocked image HSPE _ SBK p，q Matched h-th group of pixel coordinates
And (3) constructing an h group of quadratic polynomials of the p column of the q row so as to establish a quadratic polynomial model of the whole block of the p column of the q row, and solving corresponding quadratic polynomial model parameters by adopting least square.
And 8: and performing quadratic polynomial correction on the affine-transformed high-spectral-resolution remote sensing image according to the quadratic polynomial transformation model of each block to obtain the high-spectral-resolution remote sensing block image corrected by each quadratic polynomial.
And step 9: and carrying out fusion processing on each corrected high-spatial-resolution remote sensing block image and the corresponding high-spectral-resolution remote sensing block image corrected by the quadratic polynomial to obtain a remote sensing image fused with high spatial resolution and high spectral resolution.
Step 9, performing fusion processing to obtain a remote sensing image fused with high spatial resolution and high spectral resolution as follows:
setting the weight of blue band as B w =0.75 and weight of green band is G w =0.85, red band weight is R w =0.6, weight of near infrared band is IR w And =0.3. Acquiring each correction pixel by pixelThe pixel values of the high-spatial-resolution remote sensing block image corrected by the corresponding quadratic polynomial are calculated, and the blue-green-red three-band adjustment value ADJ corresponding to each pixel value is calculated 1 Near infrared adjustment value ADJ corresponding to each pixel value 2 The calculation method is as follows:
wherein, HSPA DN Representing corresponding panchromatic pixel values, ADJ, on a corrected high spatial resolution remote sensing tile image 1 Adjust values for each pixel value for three bands of cyan and red, ADJ 2 For each pixel value corresponding near-infrared adjustment value, AVGw 1 AVGW being weighted average of red, green and blue bands corresponding to each pixel 2 The weighted average of the red band, the green band, the blue band and the near-infrared band corresponding to each pixel is calculated in the following way,
wherein, B DN Representing blue band pixel value G in quadratic polynomial corrected high spectral resolution remote sensing block image DN Representing green band pixel value, R, in a quadratic polynomial corrected high spectral resolution remote sensing block image DN Representing red band pixel value, IR in high spectral resolution remote sensing block image corrected by quadratic polynomial DN And expressing the pixel value of the near infrared band in the high-spectral-resolution remote sensing block image corrected by the quadratic polynomial.
A blue, green and red three-band adjustment value ADJ corresponding to each pixel value 1 Near infrared adjustment value ADJ corresponding to each pixel value 2 Calculating the fused image value pixel by pixel, specifically as follows:
wherein, B FUs Representing the blue band pixel value, G, of the fused image FUs Representing the green band pixel value, R, of the fused image FUs Indicating the red band pixel value, IR of the fused image FUS And representing the pixel value of the fused image near infrared band.
Embodiments of the present invention also provide a computer-readable medium.
The computer readable medium is a server workstation;
the server workstation stores a computer program executed by the electronic equipment, and when the computer program runs on the electronic equipment, the electronic equipment executes the steps of the remote sensing image fusion method considering image distortion.
It should be understood that parts of the specification not set forth in detail are well within the prior art.
It should be understood that the above description of the preferred embodiments is given for clarity and not for any purpose of limitation, and that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (10)
1. A remote sensing image fusion method considering image distortion is characterized by comprising the following steps:
step 1: acquiring an original high-spatial-resolution remote sensing image and a corresponding RPC positioning model thereof, and carrying out geometric correction to obtain a corrected high-spatial-resolution remote sensing image; acquiring a high-spectral-resolution remote sensing image and a corresponding RPC positioning model thereof, and performing geometric correction to obtain a corrected high-spectral-resolution remote sensing image;
step 2: calculating the intersection polygons of the corrected high-spatial-resolution images and the corrected high-spectral-resolution images, acquiring the geographic range of the intersection polygons, and constructing the minimum circumscribed rectangle of the intersection polygons according to the geographic range of the intersection polygons; carrying out grid division processing on the minimum external rectangle of the intersected polygon to obtain a plurality of blocks of the minimum external rectangle;
and 3, step 3: mapping each block of the minimum circumscribed rectangle to the corrected high-spatial-resolution remote sensing image to obtain each corrected high-spatial-resolution remote sensing block image, mapping each block of the minimum circumscribed rectangle to the corrected high-spectral-resolution remote sensing image to obtain each corrected high-spectral-resolution remote sensing block image, and performing initial matching processing on each corrected high-spatial-resolution remote sensing block image and each corresponding corrected high-spectral-resolution remote sensing block image to obtain multiple groups of homonymy point pairs of each block;
and 4, step 4: an affine transformation model is built according to a plurality of groups of homonymous point pairs of a plurality of blocks, and the corrected high-spectral-resolution image is subjected to affine transformation according to the affine transformation model to obtain an affine-transformed high-spectral-resolution image;
and 5: calculating the updated intersection polygon of the corrected high-spatial-resolution image and the affine-transformed high-spectral-resolution image, acquiring the geographical range of the updated intersection polygon, and constructing the minimum circumscribed rectangle of the updated intersection polygon according to the geographical range of the updated intersection polygon; carrying out grid division processing on the minimum external rectangle of the updated intersected polygon to obtain a plurality of blocks of the minimum external rectangle;
and 6: mapping each block of the minimum external rectangle to the corrected high-spatial-resolution remote sensing image to obtain each corrected high-spatial-resolution remote sensing block image, mapping each block of the minimum external rectangle to the affine-transformed high-spectral-resolution remote sensing image to obtain each corrected high-spectral-resolution remote sensing block image, and performing fine matching processing on each corrected high-spatial-resolution remote sensing block image and each corresponding affine-transformed high-spectral-resolution remote sensing block image to obtain multiple groups of homonymy point pairs of each block;
and 7: constructing a polynomial transformation model of each block according to the multiple groups of homonymy point pairs of each block, and solving by combining the multiple groups of homonymy points of each sub-block by adopting a least square method to obtain parameters of the polynomial transformation model of each sub-block to obtain the polynomial transformation model of each block;
and 8: performing quadratic polynomial correction on the affine-transformed high-spectral-resolution remote sensing image according to the quadratic polynomial transformation model of each block to obtain a high-spectral-resolution remote sensing block image corrected by each quadratic polynomial;
and step 9: and carrying out fusion processing on each corrected high-spatial-resolution remote sensing block image and the corresponding high-spectral-resolution remote sensing block image corrected by the quadratic polynomial to obtain a remote sensing image fused with high spatial resolution and high spectral resolution.
2. The remote sensing image fusion method taking image distortion into account as claimed in claim 1, wherein:
step 2, calculating an intersecting polygon of the corrected high-spatial-resolution image and the corrected high-spectral-resolution image, wherein the specific process is as follows:
constructing a corrected high-spatial-resolution image vector polygon according to the corrected four-corner geographical coordinates of the high-spatial-resolution image;
constructing a corrected high-spectral-resolution image vector polygon according to the corrected four-corner geographical coordinates of the high-spectral-resolution image;
calculating by a space intersection calculation method to obtain an intersection polygon between the corrected high-spatial-resolution image vector polygon and the corrected high-spectral-resolution image vector polygon;
step 2, acquiring the geographic range of the intersected polygon, and constructing the minimum circumscribed rectangle of the intersected area according to the geographic range of the intersected polygon, wherein the specific process is as follows:
sequentially acquiring the longitude of each point and the latitude of each point in the intersected polygon;
and (3) screening the longitude minimum value of each point in the intersected polygon and the longitude maximum value of each point in the intersected polygon from the longitudes of the points in the intersected polygon, wherein the longitude range for constructing the intersected polygon is defined as follows:
[FMIN_LON，FMAX_LON]
wherein FMIN _ LON represents a longitude minimum for each point in the intersecting polygon and FMAX _ LON represents a longitude maximum for each point in the intersecting polygon;
screening out the latitude minimum value of each point in the intersected polygon and the dimension maximum value of each point in the intersected polygon from the latitudes of the points in the intersected polygon, and constructing the latitude range of the intersected polygon, wherein the latitude range is defined as follows:
[FMIN_LAT，FMAX_LAT]
wherein FMIN _ LAT represents the latitude minimum value of each point in the intersecting polygon, and FMAX _ LAT represents the latitude maximum value of each point in the intersecting polygon;
taking (FMAX _ LON, FMIN _ LAT), (FMAX _ LON, FMAX _ LAT), (FMAX _ LON, FMIN _ LAT) as four vertices of a minimum bounding rectangle of the intersecting polygon to construct a minimum bounding rectangle of the intersecting polygon;
step 2, grid division processing is performed on the minimum circumscribed rectangle of the intersected polygons to obtain a plurality of blocks of the minimum circumscribed rectangle, which specifically comprises the following steps:
setting the division number of a longitudinal direction grid network as nFBK _ lon and setting the division number of a latitudinal direction grid network as nFBK _ lat;
calculating the longitude width of each minimum circumscribed rectangle block and the latitude height of each minimum circumscribed rectangle block, specifically as follows:
wherein, the delta F _ LON is the longitude width of each minimum external rectangular block, the delta F _ LAT is the latitude height of each minimum external rectangular block, the nFBK _ LON is the division number of the longitude direction grid, the nFBK _ LAT is the division number of the latitude direction grid, FMIN _ LON represents the longitude minimum for each point in the intersecting polygon, FMAX _ LON represents the longitude maximum for each point in the intersecting polygon, where FMIN _ LAT represents the latitude minimum for each point in the intersecting polygon, and FMAX _ LAT represents the latitude maximum for each point in the intersecting polygon;
the minimum circumscribed rectangle block of the jth row and the ith column is defined as:
FBK i,j
i∈[1,nFBK_lon]，j∈[1,nFBK_lat]
wherein i represents the block number in the longitude direction, j represents the block number in the latitude direction, nFBK _ lon is the division number of the direction grid network, and nFBK _ lat is the division number of the grid network in the latitude direction
Sequentially calculating the longitude minimum value, the longitude maximum value, the latitude minimum value and the latitude maximum value of the minimum circumscribed rectangle block of the jth row and the ith column, wherein the method specifically comprises the following steps:
wherein the content of the first and second substances,
represents the minimum longitude of the smallest bounding rectangle block of the jth row and ith column,
represents the maximum value of the longitude of the minimum bounding rectangle block of the jth row and ith column,
represents the minimum value of the latitude of the smallest circumscribed rectangular block of the jth row and ith column,
the maximum value of the latitude of the minimum circumscribed rectangle block of the jth row and the ith column is represented;
FMIN _ LON represents the longitude minimum for each point in the intersecting polygon, and FMAX _ LON represents the longitude maximum for each point in the intersecting polygon; wherein FMIN _ LAT represents the latitude minimum value of each point in the intersecting polygon, and FMAX _ LAT represents the latitude maximum value of each point in the intersecting polygon;
will be provided with
And the geographic rectangular blocks are used as four vertexes to construct the geographic rectangular blocks corresponding to the minimum circumscribed rectangular blocks of the jth row and the ith column.
3. The method for fusing remote sensing images considering image distortion according to claim 2, wherein:
and 3, mapping to the corrected high-spatial-resolution remote sensing images to obtain each corrected high-spatial-resolution remote sensing block image, wherein the specific process is as follows:
partitioning four vertexes according to the minimum circumscribed rectangle of the jth row and the ith column
The image content of the corresponding geographical range is obtained from the corrected remote sensing image with the high spatial resolution;
the corrected high spatial resolution remote sensing block image of the jth row and ith column is defined as:
HSPA_FBK i,j
i∈[1,nFBK_lon]，j∈[1,nFBK_lat]
wherein i represents the serial number of the blocks in the longitude direction, j represents the serial number of the blocks in the latitude direction, nFBK _ lon is the division number of the directional grid network, and nFBK _ lat is the division number of the grid network in the latitude direction;
and 3, mapping to the corrected high-spectral-resolution remote sensing image to obtain each high-spectral-resolution remote sensing block image, wherein the specific process is as follows:
partitioning four vertexes according to the minimum circumscribed rectangle of the jth row and the ith column
The image content of the corresponding geographical range is obtained from the corrected remote sensing image with the high spectral resolution;
the corrected high spectral resolution remote sensing block image of the jth row and the ith column is defined as:
HSPE_FBK i,j
i∈[1,nFBK_lon]，j∈[1,nFBK_lat]
wherein i represents the serial number of the blocks in the longitude direction, j represents the serial number of the blocks in the latitude direction, nFBK _ lon is the division number of the directional grid network, and nFBK _ lat is the division number of the grid network in the latitude direction;
and 3, obtaining a plurality of groups of homonymy point pairs of each block through initial matching processing, wherein the specific process is as follows:
corrected high spatial resolution remote sensing blocked image HSPA _ FBK of jth row and ith column i,j And the corrected high-spectral-resolution remote sensing blocked image HSPE _ FBK of the ith row and the ith column in the corresponding j i,j Acquiring the corrected high spatial resolution remote sensing blocked image HSPA _ FBK of the jth row and ith column by adopting an SIFT + KNN matching algorithm i,j Pixel coordinates and jth row and ith column corrected high-spectral-resolution remote sensing block image HSPE _ FBK i,j Pixel coordinates;
obtaining the minimum longitude value of the jth row and ith column sub-block
Minimum latitude
Correcting the jth row and ith column of the high-spatial-resolution remote sensing blocked image HSPA _ FBK i,j The pixel coordinates are mapped into high spatial resolution remote sensing block image geographical coordinates in a mode of the sum of the product of the pixel coordinates and the resolution and the minimum value;
obtaining the minimum longitude value of the jth row and ith column sub-block
Minimum latitude
High-spectral-resolution remote sensing blocked image HSPE _ FBK corrected in jth row and ith column i,j The pixel coordinates are mapped into geographical coordinates according to the mode of the sum of the pixel coordinates, the resolution product and the minimum value;
and combining the geographic coordinates of the high-spatial-resolution remote sensing block images and the geographic coordinates of the high-spectral-resolution remote sensing block images corresponding to the jth row and ith column sub-blocks into a plurality of groups of dotted pairs in the jth row and ith column.
4. The remote sensing image fusion method considering image distortion according to claim 3, characterized in that:
step 4, establishing an affine transformation model according to the multiple groups of homonymous point pairs of the multiple blocks, which specifically comprises the following steps:
constructing a global affine transformation model according to the geographic coordinates of the high-spatial-resolution remote sensing block images and the geographic coordinates of the high-spectral-resolution remote sensing block images corresponding to all block homonymy point pairs, and solving parameters of the affine transformation model by adopting a least square method;
and 4, carrying out affine transformation processing on the corrected high-spectral-resolution image according to an affine transformation model, wherein the specific process is as follows:
incorporating affine transformation model parameters (e) 0 ,e 1 ,e 2 ,f 0 ,f 1 ,f 2 ) And performing affine transformation correction on the corrected high-spectral-resolution image by adopting an indirect correction method to obtain the high-spectral-resolution image with affine transformation.
5. The remote sensing image fusion method considering image distortion according to claim 4, characterized in that:
step 5, calculating the intersection polygon after the corrected high spatial resolution image and the affine transformation high spectral resolution image are updated, and the specific process is as follows:
constructing a corrected high-spatial-resolution image vector polygon according to the corrected four-corner geographical coordinates of the high-spatial-resolution image; constructing an affine-transformed high-spectral-resolution image vector polygon according to the geographical coordinates of the four corner points of the affine-transformed high-spectral-resolution image; calculating an intersection polygon between the corrected high-spatial-resolution image vector polygon and the affine-transformed high-spectral-resolution image vector polygon by using a spatial intersection calculation method, wherein the intersection polygon is an updated intersection polygon;
step 5, obtaining the updated geographic range of the intersected polygon, and constructing the minimum circumscribed rectangle of the intersected area according to the updated geographic range of the intersected polygon, wherein the specific process is as follows:
sequentially acquiring the longitude of each point and the latitude of each point in the updated intersected polygon;
screening out the longitude minimum value of each point in the updated intersected polygon and the longitude maximum value of each point in the updated intersected polygon from the longitudes of the points in the updated intersected polygon, and constructing the longitude range of the updated intersected polygon to define as follows:
[SMIN_LON，SMAX_LON]
wherein SMIN _ LON represents the longitude minimum value of each point in the updated intersected polygon, and SMAX _ LON represents the longitude maximum value of each point in the updated intersected polygon;
screening out the latitude minimum value of each point in the updated intersected polygon and the latitude maximum value of each point in the updated intersected polygon from the latitudes of the points in the updated intersected polygon, and constructing the latitude range of the updated intersected polygon, wherein the latitude range is defined as follows:
[SMIN_LAT，SMAX_LAT]
wherein SMIN _ LAT represents the latitude minimum value of each point in the updated intersected polygon, and SMAX _ LAT represents the latitude maximum value of each point in the updated intersected polygon;
taking (SMAX _ LON, SMIN _ LAT), (SMAX _ LON, SMAX _ LAT), (SMAX _ LON, SMIN _ LAT) as four vertexes of a minimum bounding rectangle of the updated intersected polygon to construct a minimum bounding rectangle of the updated intersected polygon;
step 5, performing grid division processing on the updated minimum circumscribed rectangle of the intersected polygon to obtain a plurality of blocks of the minimum circumscribed rectangle, specifically as follows:
setting the division number of the longitudinal direction grid as nSBK _ lon and setting the division number of the latitudinal direction grid as nSBK _ lat;
calculating the longitude width of each minimum circumscribed rectangle block and the latitude height of each minimum circumscribed rectangle block, specifically as follows:
wherein, Δ S _ LON is the longitude width of each minimum circumscribed rectangle block, Δ S _ lat is the latitude height of each minimum circumscribed rectangle block, nSBK _ LON is the division number of the grid in the longitude direction, nSBK _ lat is the division number of the grid in the set latitude direction, SMIN _ LON represents the longitude minimum value of each point in the intersection polygon after updating, SMAX _ LON represents the longitude maximum value of each point in the intersection polygon after updating; wherein SMIN _ LAT represents the latitude minimum value of each point in the updated intersected polygon, and SMAX _ LAT represents the latitude maximum value of each point in the updated intersected polygon;
the minimum circumscribed rectangle block of the qth row and the pth column is defined as:
SBK p,q
p∈[1,nSBK_lon]，q∈[1,nSBK_lat]
wherein p represents the block serial number in the longitude direction, q represents the block serial number in the latitude direction, nSBK _ lon is the division number of the directional grid, nSBK _ lat is the division number of the grid in the latitude direction
Sequentially calculating the longitude minimum value, the longitude maximum value, the latitude minimum value and the latitude maximum value of the minimum circumscribed rectangle block of the qth row and the pth column, which are specifically as follows:
wherein the content of the first and second substances,
represents the minimum longitude of the smallest circumscribed rectangular block of the qth row and the pth column,
represents the maximum value of the longitude of the minimum bounding rectangle block of the qth row and the pth column,
represents the minimum value of latitude of the minimum circumscribed rectangle block of the qth row and the pth column,
represents the latitude maximum of the minimum bounding rectangle block of the qth row and the pth column,
SMIN _ LON represents the longitude minimum value of each point in the intersection polygon after updating, and SMAX _ LON represents the longitude maximum value of each point in the intersection polygon after updating; wherein SMIN _ LAT represents the latitude minimum value of each point in the updated intersected polygon, and SMAX _ LAT represents the latitude maximum value of each point in the updated intersected polygon; will be provided with
And the geographic rectangular blocks are used as four vertexes to construct the geographic rectangular blocks corresponding to the minimum circumscribed rectangular blocks of the qth row and the pth column.
6. The method for fusing remote sensing images considering image distortion according to claim 5, wherein:
and 6, mapping to the corrected high-spatial-resolution remote sensing images to obtain each corrected high-spatial-resolution remote sensing block image, wherein the specific process is as follows:
partitioning four vertexes according to the minimum circumscribed rectangle of the q-th row and the p-th column
The geographical range is formed, and the image content of the corresponding geographical range is obtained from the corrected remote sensing image with high spatial resolution;
the corrected high spatial resolution remote sensing block image of the q-th row and the p-th column is defined as follows:
HSPA_SBK p,q
p∈[1,nSBK_lon]，q∈[1,nSBK_lat]
wherein p represents the serial number of the blocks in the longitude direction, q represents the serial number of the blocks in the latitude direction, nSBK _ lon is the division number of the latticed network in the longitude direction, and nSBK _ lat is the division number of the latticed network in the latitude direction;
and 6, mapping the image to the affine transformed high-spectral-resolution remote sensing image to obtain each corrected high-spatial-resolution remote sensing block image, wherein the specific process is as follows:
partitioning four vertexes according to the minimum circumscribed rectangle of the q-th row and the p-th column
The geographical range is formed, and the image content of the corresponding geographical range is obtained from the corrected remote sensing image with high spatial resolution;
the affine transformation high spatial resolution remote sensing block image of the q-th row and the p-th column is defined as follows:
HSPE_SBK p,q
p∈[1,nSBK_lon]，q∈[1,nSBK_lat]
wherein p represents the serial number of the blocks in the longitude direction, q represents the serial number of the blocks in the latitude direction, nSBK _ lon is the division number of the directional grid, and nSBK _ lat is the division number of the grid in the latitude direction.
7. The method for fusing remote-sensing images considering image distortion according to claim 6, wherein:
and 6, obtaining a plurality of groups of same-name point pairs of each block through fine matching processing, wherein the specific process is as follows:
corrected high spatial resolution remote sensing blocked image HSPA _ SBK of q row and p column p，q The validity determination is specifically performed as follows:
if CNT _ HSPA _ SBK p,q Not less than alpha, then
If CNT _ HSPA _ SBK p,q <α, then
Wherein CNT _ HSPA _ SBK p,q Corrected high spatial resolution remote sensing blocked image HSPA _ SBK for expressing q row and p column p,q The number of pixel values greater than 0, alpha represents a block validity determination threshold,
representing the validity attribute of the high-spatial-resolution remote sensing blocked image corrected by the qth row and the pth column;
if it is
The corrected high spatial resolution remote sensing blocked image HSPA _ SBK of the q-th row and the p-th column p,q And a corresponding q-row p-column affine transformation high-spectral-resolution remote sensing blocked image HSPE _ SBK p,q Subsequent matching is not performed;
if it is
The corrected high spatial resolution remote sensing blocked image HSPA _ SBK of the q-th row and the p-th column p,q And a high-spectral-resolution remote sensing blocked image HSPE _ SBK of affine transformation of the p-th column of the corresponding q rows p,q Carrying out the matching, asThe following:
corrected high spatial resolution remote sensing blocked image HSPA _ SBK of p row and p column of q rows p,q And a corresponding q-row p-column affine transformation high-spectral-resolution remote sensing blocked image HSPE _ SBK p,q Acquiring the corrected high-spatial-resolution remote sensing segmented image HSPA _ SBK of the p-th row and the q-th column by adopting an SIFT + RANSAC matching algorithm p,q High-spectral-resolution remote sensing block image HSPE _ SBK corrected by pixel coordinates and p-th row and p-th column p,q Pixel coordinates;
and combining the pixel coordinates of the high-spatial-resolution remote sensing block images corresponding to the q-th row and p-th column blocks and the pixel coordinates of the affine-transformed high-spectral-resolution remote sensing block images into a plurality of groups of homonymous point pairs of the q-th row and p-th column blocks.
8. The remote sensing image fusion method considering image distortion according to claim 7, characterized in that:
step 7, constructing a polynomial transformation model of each block according to the multiple groups of homonymy point pairs of each block as follows:
obtaining corrected high spatial resolution remote sensing block image HSPA _ SBK of the q-th row and the p-th column p,q Properties
If it is
Directly adopting the corrected high spatial resolution remote sensing blocked image HSPA _ SBK of the q row and the p column p,q And a high-spectral-resolution remote sensing blocked image HSPE _ SBK of affine transformation of the p-th column of the corresponding q rows p,q The matched homonymous point pairs are used for solving a polynomial model;
if it is
Sequentially judging corrected high-spatial-resolution remote sensing blocked images HSPA _ SBK of the p-1 th row and the q-th column p-1,q Attribute
Corrected high-spatial-resolution remote sensing blocked image HSPA _ SBK of q row and p +1 column p+1,q Attribute
Corrected high-spatial-resolution remote sensing blocked image HSPA _ SBK of q-1 row and p column p,q-1 Property, property
Corrected high-spatial-resolution remote sensing blocked image HSPA _ SBK of q +1 row and p column p,q+1 Properties
If the number of the sub-blocks is 1, matching a plurality of groups of homonymous point pairs with the corresponding sub-blocks to serve as a plurality of groups of homonymous point pairs of the current qth row and the pth column sub-blocks;
the polynomial transformation model of the block of the qth row and the block of the pth column is specifically constructed as follows:
obtaining corrected high spatial resolution remote sensing block image HSPA _ SBK of the q-th row and the p-th column p,q Matched h-th group of pixel coordinates
h∈[1,CT p,q ]p h ∈[1,nSBK_lon]，q h ∈[1,nSBK_lat]
Wherein the content of the first and second substances,
corrected high spatial resolution remote sensing blocked image HSPA _ SBK representing q-th row and p-th column p,q And q rows and p columns of affine transformation high-spectral-resolution remote sensing blocked image HSPE _ SBK p,q The number of matched homonyms; h represents a serial number, nSBK _ lon is the division number of the grid network in the longitude direction, and nSBK _ lat is the division number of the grid network in the latitude direction;
obtaining a high-spectral-resolution remote sensing blocked image HSPE _ SBK of affine transformation of a q-th row and a p-th column p,q Matched h-th group of pixel coordinates
h∈[1,CT p,q ]p h ∈[1,nSBK_lon]，q h ∈[1,nSBK_lat]
Wherein the content of the first and second substances,
corrected high spatial resolution remote sensing blocked image HSPA _ SBK for expressing q row and p column p,q And q rows and p columns of affine transformation high-spectral-resolution remote sensing block image HSPE _ SBK p,q The number of matched homonyms; h represents a serial number, nSBK _ lon is the division number of the division network in the longitude direction, and nSBK _ lat is the division number of the division network in the latitude direction;
sequentially correcting the high-spatial-resolution remote sensing blocked image HSPA _ SBK according to the q-th row and the p-th column p,q Matched h-th group of pixel coordinates
And a q-th row and p-th column affine transformation high-spectral-resolution remote sensing blocked image HSPE _ SBK p,q Matched h-th group of pixel coordinates
And constructing an h group of quadratic polynomials of the p column of the q row so as to establish a quadratic polynomial model of the whole block of the p column of the q row, and solving corresponding quadratic polynomial model parameters by adopting least square.
9. The method for fusing remote-sensing images considering image distortion according to claim 8, wherein:
and 9, performing fusion processing to obtain a remote sensing image fused with high spatial resolution and high spectral resolution, which comprises the following specific steps:
setting the weight of blue band as B w Green band weight of G w The red band weight is R w The weight of the near infrared band is IR w ；
Acquiring pixel values of each corrected high-spatial-resolution remote sensing block image and the corresponding quadratic polynomial corrected high-spatial-resolution remote sensing block image pixel by pixel, and calculating blue-green-red three-band adjustment value ADJ corresponding to each pixel value 1 Near infrared adjustment value ADJ corresponding to each pixel value 2 The calculation method is as follows:
wherein, HSPA DN Representing corresponding panchromatic pixel values, ADJ, on a corrected high spatial resolution remote sensing tile image 1 Adjust values for each pixel value for three bands of cyan and red, ADJ 2 For each pixel value corresponding near-infrared adjustment value, AVGW 1 AVGW being weighted average of red, green and blue bands corresponding to each pixel 2 The weighted average of the red band, the green band, the blue band and the near-infrared band corresponding to each pixel is calculated as follows,
wherein, B DN Representing blue band pixel value G in quadratic polynomial corrected high spectral resolution remote sensing block image DN Representing green band pixel value, R, in a quadratic polynomial corrected high spectral resolution remote sensing block image DN High spectral resolution remote sensing block image red band pixel value, IR expressing quadratic polynomial correction DN Representing the pixel value of a near infrared band in a high-spectral-resolution remote sensing block image corrected by a quadratic polynomial;
a blue, green and red three-band adjustment value ADJ corresponding to each pixel value 1 Near infrared adjustment value A corresponding to each pixel valueDJ 2 Calculating the fused image value pixel by pixel, specifically as follows:
wherein, B FUS Representing the blue band pixel value, G, of the fused image FUS Representing the green band pixel value, R, of the fused image FUS Representing the red band pixel value, IR, of the fused image FUS And representing the pixel value of the fused image near infrared band.
10. A computer-readable medium, characterized in that it stores a computer program for execution by an electronic device, which computer program, when run on the electronic device, causes the electronic device to perform the steps of the method according to claims 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211434744.5A CN115690460A (en) | 2022-11-16 | 2022-11-16 | Remote sensing image fusion method considering image distortion and computer readable medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211434744.5A CN115690460A (en) | 2022-11-16 | 2022-11-16 | Remote sensing image fusion method considering image distortion and computer readable medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115690460A true CN115690460A (en) | 2023-02-03 |
Family
ID=85053137
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211434744.5A Pending CN115690460A (en) | 2022-11-16 | 2022-11-16 | Remote sensing image fusion method considering image distortion and computer readable medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115690460A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117197625A (en) * | 2023-08-29 | 2023-12-08 | 珠江水利委员会珠江水利科学研究院 | Remote sensing image space-spectrum fusion method, system, equipment and medium based on correlation analysis |
-
2022
- 2022-11-16 CN CN202211434744.5A patent/CN115690460A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117197625A (en) * | 2023-08-29 | 2023-12-08 | 珠江水利委员会珠江水利科学研究院 | Remote sensing image space-spectrum fusion method, system, equipment and medium based on correlation analysis |
| CN117197625B (en) * | 2023-08-29 | 2024-04-05 | 珠江水利委员会珠江水利科学研究院 | Remote sensing image space-spectrum fusion method, system, equipment and medium based on correlation analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110211043B (en) | Registration method based on grid optimization for panoramic image stitching | |
| CN103914808B (en) | Method for splicing ZY3 satellite three-line-scanner image and multispectral image | |
| US8577139B2 (en) | Method of orthoimage color correction using multiple aerial images | |
| CN111508029A (en) | Satellite-borne segmented linear array CCD optical camera overall geometric calibration method and system | |
| CN108759788B (en) | Unmanned aerial vehicle image positioning and attitude determining method and unmanned aerial vehicle | |
| AU2018206753A1 (en) | Method and device for filling invalid regions of terrain elevation model data | |
| CN115690460A (en) | Remote sensing image fusion method considering image distortion and computer readable medium | |
| CN111768337B (en) | Image processing method and device and electronic equipment | |
| CN106600551A (en) | High-precision geometry geometric correction method for large-scene spaceborne SAR image | |
| CN108629742B (en) | True ortho image shadow detection and compensation method, device and storage medium | |
| CN111709876B (en) | Image splicing method, device, equipment and storage medium | |
| Liu et al. | Robust radiometric normalization of multitemporal satellite images via block adjustment without master images | |
| CN114663298A (en) | Disparity map repairing method and system based on semi-supervised deep learning | |
| CN112734630A (en) | Ortho image processing method, device, equipment and storage medium | |
| CN107516291B (en) | Night scene image ortho-rectification processing method | |
| CN108985154B (en) | Small-size ground object sub-pixel positioning method and system based on image concentration | |
| CN116563460A (en) | Three-dimensional reconstruction optimization method and device, medium and equipment | |
| CN108052986B (en) | Least square matching method based on multiple channels | |
| CN113592744B (en) | Geometric fine correction method suitable for high-resolution remote sensing image | |
| CN112017108B (en) | Satellite image color relative correction method based on adjustment of independent model method | |
| CN113324527B (en) | Co-rail laser height measurement point and three-linear array three-dimensional image combined surveying and mapping processing method | |
| CN115423702A (en) | Method and system for manufacturing large-area space-borne optical and SAR (synthetic Aperture Radar) image DOM (document object model) | |
| CN115330619A (en) | Local geometric fine correction method suitable for high-resolution remote sensing image | |
| CN114581498A (en) | Iterative model registration system combining vector data and raster image | |
| CN114359389A (en) | Large image blocking epipolar line manufacturing method based on image surface epipolar line pair |
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 |