Search

CN-115690460-B - Remote sensing image fusion method considering image distortion and computer readable medium

CN115690460BCN 115690460 BCN115690460 BCN 115690460BCN-115690460-B

Abstract

The invention provides a remote sensing image fusion method and a computer readable medium for considering image distortion. The method solves the intersecting polygons participating in fusion of the images, and constructs the minimum circumscribed rectangle of the intersecting polygons according to the geographic range of the intersecting polygons. And performing grid division on the minimum circumscribed rectangle to obtain a plurality of blocks, performing initial matching, and performing affine transformation processing on the corrected high-spectral-resolution image based on a matching result. Updating the intersected polygons and the corresponding minimum circumscribed rectangles of the fusion images, performing grid division to obtain a plurality of blocks, performing fine matching processing, and further performing polynomial correction on the hyperspectral resolution remote sensing images subjected to affine transformation processing based on the matching result. And respectively carrying out fusion processing on each block participating in the fusion image, thereby obtaining a fused remote sensing image. The invention can effectively solve the problem of poor fusion effect caused by distortion in the images and can realize fusion between large-breadth remote sensing images.

Inventors

  • WANG TAOYANG
  • CHENG QIAN
  • LI XIN

Assignees

  • 武汉大学

Dates

Publication Date
20260508
Application Date
20221116

Claims (9)

  1. 1. The remote sensing image fusion method taking image distortion into consideration is characterized by comprising the following steps of: step 1, obtaining an original high-spatial-resolution remote sensing image and a corresponding RPC positioning model thereof, and performing geometric correction to obtain a corrected high-spatial-resolution remote sensing image; Step 2, calculating the intersected polygons of the corrected high-spatial resolution image and the corrected high-spectral resolution image, obtaining the geographic range of the intersected polygons, and constructing the minimum circumscribed rectangle of the intersected polygons according to the geographic range of the intersected polygons; 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 obtaining a plurality of groups of homonymous point pairs of each block through initial matching processing of each corrected high-spatial resolution remote sensing block image and each corresponding corrected high-spectral resolution remote sensing block image; Step 4, constructing an affine transformation model according to a plurality of groups of point pairs with the same name of a plurality of blocks, and carrying out affine transformation processing on the corrected hyperspectral resolution image according to the affine transformation model to obtain an affine transformed hyperspectral resolution image; Step 5, calculating the corrected high-spatial resolution image and the affine transformed high-spectral resolution image, updating the intersected polygons, obtaining the geographical ranges of the updated intersected polygons, and constructing the minimum circumscribed rectangles of the updated intersected polygons according to the geographical ranges of the updated intersected polygons; 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 affine transformed high spectral resolution remote sensing image to obtain each corrected high spectral resolution remote sensing block image, and carrying out 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 a plurality of groups of homonymous point pairs of each block; step 7, constructing a polynomial transformation model of each block according to a plurality of groups of homonymous point pairs of each block, and solving by a least square method by combining the plurality of groups of homonymous points of each sub-block to obtain parameters of the polynomial transformation model of each sub-block to obtain the polynomial transformation model of each block; Step 8, performing secondary polynomial correction on the affine transformed hyperspectral resolution remote sensing image according to the secondary polynomial transformation model of each block to obtain a hyperspectral resolution remote sensing block image corrected by each secondary polynomial; Step 9, carrying out fusion processing on each corrected high-spatial-resolution remote sensing segmented image and the corresponding secondary polynomial corrected high-spectral-resolution remote sensing segmented image to obtain a fused high-spatial-resolution and high-spectral-resolution remote sensing image; and 7, constructing a polynomial transformation model of each block according to a plurality of groups of homonymous point pairs of each block, wherein the polynomial transformation model is specifically as follows: Obtaining corrected high spatial resolution remote sensing segmented image of the qth row and p column Attributes of , If it is Then the corrected high-spatial resolution remote sensing segmented image of the q-th row and the p-th column is directly adopted Affine transformed hyperspectral resolution remote sensing segmented image with corresponding q rows and p columns The matched homonymy point pairs are used for solving a polynomial model; If it is Sequentially judging corrected high-spatial-resolution remote sensing segmented images of the (q) th row and the (p-1) th column Attributes of Corrected high spatial resolution remote sensing segmented image of row q and column p+1 Attributes of Corrected high spatial resolution remote sensing segmented image of q-1 row and p column Attributes and properties Corrected high spatial resolution remote sensing segmented image of row q+1st column p Attributes of If the number is 1, matching a plurality of groups of homonymous point pairs by the corresponding blocks as a plurality of groups of homonymous point pairs of the current q-th row and p-th column sub-blocks if the number is 1; the construction of the q-th row and p-th column block polynomial transformation model is specifically as follows: Obtaining corrected high spatial resolution remote sensing segmented image of the qth row and p column Matched h-th set of pixel coordinates Wherein, the High spatial resolution remote sensing segmented image representing correction of row and column q Affine transformed hyperspectral resolution remote sensing segmented image with q rows and p columns The number of matched homonymy points; The number nSBK _lon is the number of divided grids in the longitudinal direction, and nSBK _lat is the number of divided grids in the latitudinal direction; Obtaining affine transformed hyperspectral resolution remote sensing segmented image of the q-th row and p-th column Matched h-th set of pixel coordinates Wherein, the High spatial resolution remote sensing segmented image representing correction of row and column q Affine transformed hyperspectral resolution remote sensing segmented image with q rows and p columns The number of matched homonymy points; The number nSBK _lon is the number of divided grids in the longitudinal direction, and nSBK _lat is the number of divided grids in the latitudinal direction; high spatial resolution remote sensing segmented image corrected sequentially according to row q and column p Matched h-th set of pixel coordinates Affine transformed hyperspectral resolution remote sensing segmented image with q-th row and p-th column Matched h-th set of pixel coordinates And constructing an h group of quadratic polynomials of the q row and the p column, thereby constructing a quadratic polynomial model of the whole q row and the p column block, and solving corresponding quadratic polynomial model parameters by least square.
  2. 2. The remote sensing image fusion method considering image distortion as claimed in claim 1, wherein: step 2, calculating the intersected polygons 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 high-spatial-resolution image four-corner geographic coordinates; Constructing a corrected hyperspectral resolution image vector polygon according to the corrected hyperspectral resolution image four-corner geographic coordinates; calculating to obtain an intersecting polygon between the corrected high-spatial-resolution image vector polygon and the corrected high-spectral-resolution image vector polygon by a spatial intersection calculation method; Step 2, obtaining the geographical range of the intersecting polygon, and constructing a minimum bounding rectangle of the intersecting area according to the geographical range of the intersecting polygon, wherein the specific process is as follows: Acquiring longitude of each point and latitude of each point in the intersecting polygon in sequence; The longitude minimum value of each point in the intersecting polygon and the longitude maximum value of each point in the intersecting polygon are screened out from the longitudes of a plurality of points in the intersecting polygon, and the longitudes range for constructing the intersecting polygon is defined as follows: [FMIN_LON,FMAX_LON] wherein fmin_lon represents a minimum longitude value for each point in the intersecting polygon, and fmax_lon represents a maximum longitude value for each point in the intersecting polygon; Screening out the minimum latitude value of each point in the intersecting polygon and the maximum dimension value of each point in the intersecting polygon from the latitudes of a plurality of points in the intersecting polygon, constructing the latitude range of the intersecting polygon, and defining the following steps: [FMIN_LAT,FMAX_LAT] wherein fmin_lat represents a minimum latitude value for each point in the intersecting polygon, and fmax_lat represents a maximum latitude value for each point in the intersecting polygon; taking (FMAX_LON, FMIN_LAT), (FMAX_LON, FMAX_LAT), (FMAX_LON, FMIN_LAT) as four vertices of the minimum bounding rectangle of the intersecting polygon to construct the minimum bounding rectangle of the intersecting polygon; And step 2, performing grid division processing on the minimum circumscribed rectangle of the intersected polygon to obtain a plurality of blocks of the minimum circumscribed rectangle, wherein the blocks are specifically as follows: Setting the number of dividing grids in the longitudinal direction as nFBK _lon and the number of dividing grids in the latitudinal direction as nFBK _lat; The longitude width of each minimum circumscribed rectangle block and the latitude height of each minimum circumscribed rectangle block are calculated, and the method specifically comprises the following steps: Wherein, the For the longitude width of each minimum bounding rectangle partition, For the latitude height of each minimum circumscribed rectangular block, nFBK _lon is the number of longitudinal grid divisions, nFBK _lat is the set number of latitudinal grid divisions, FMIN_LON represents the minimum value of longitude of each point in the intersecting polygon, FMAX_LON represents the maximum value of longitude of each point in the intersecting polygon, FMIN_LAT represents the minimum value of latitude of each point in the intersecting polygon, and FMAX_LAT represents the maximum value of latitude of each point in the intersecting polygon; The smallest circumscribed rectangular block of the j-th row and i-th column is defined as: , Wherein i represents a block number in the longitudinal direction, j represents a block number in the latitudinal direction, nFBK _lon represents a division number of a directional grid, nFBK _lat represents a division number of a latitudinal grid Sequentially calculating a longitude minimum value, a longitude maximum value, a latitude minimum value and a latitude maximum value of the minimum circumscribed rectangle of the j-th row and i-th column, wherein the longitude minimum value, the longitude maximum value, the latitude minimum value and the latitude maximum value of the minimum circumscribed rectangle are specifically as follows: Wherein, the A minimum value of longitude representing the minimum circumscribed rectangular block of the j-th row and i-th column, Representing the longitude maximum of the minimum bounding rectangular tile of the j-th row and i-th column, Representing the minimum value of the latitude of the smallest circumscribed rectangular block of the j-th row and i-th column, Representing the maximum latitude value of the minimum circumscribed rectangular block of the j-th row and the i-th column; 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, FMIN_LAT represents the minimum latitude value of each point in the intersecting polygon, and FMAX_LAT represents the maximum latitude value of each point in the intersecting polygon; Will be% 、 )、( 、 )、( 、 )、 、 ) And the four vertexes are used for constructing a geographic rectangular block corresponding to the smallest circumscribed rectangular block of the jth row and the ith column.
  3. 3. The remote sensing image fusion method considering image distortion as claimed in claim 2, wherein: and step 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: Dividing four vertexes according to the minimum circumscribed rectangle of the j-th row and the i-th column 、 )、( 、 )、( 、 )、 、 ) The composed geographic range, and obtaining image content corresponding to the geographic range from the corrected high-spatial resolution remote sensing image; The corrected high spatial resolution remote sensing segmented image of the j-th row and i-th column is defined as: , wherein i represents a block number in a longitude direction, j represents a block number in a latitude direction, nFBK _lon is a direction division number of a grid, and nFBK _lat is a latitude direction division number of a grid; and step 3, mapping to the corrected hyperspectral resolution remote sensing images to obtain each hyperspectral resolution remote sensing block image, wherein the specific process is as follows: Dividing four vertexes according to the minimum circumscribed rectangle of the j-th row and the i-th column 、 )、( 、 )、( 、 )、 、 ) The composed geographical range, and obtaining the image content of the corresponding geographical range from the corrected hyperspectral resolution remote sensing image; The corrected hyperspectral resolution remote sensing segmented image of the j-th row and i-th column is defined as: , wherein i represents a block number in a longitude direction, j represents a block number in a latitude direction, nFBK _lon is a direction division number of a grid, and nFBK _lat is a latitude direction division number of a grid; And step 3, obtaining a plurality of groups of homonymous point pairs of each block through initial matching processing, wherein the specific process is as follows: Corrected high spatial resolution remote sensing segmented image for jth row and ith column And corresponding corrected hyperspectral resolution remote sensing segmented image of jth row and ith column Obtaining a j-th row and i-th column corrected high-spatial-resolution remote sensing segmented image by adopting SIFT+KNN matching algorithm High spectral resolution remote sensing segmented image corrected by pixel coordinates and jth row and jth column Pixel coordinates; Acquiring minimum value of longitude of ith row and ith column sub-block Minimum value of latitude High-spatial resolution remote sensing segmented image corrected by jth row and ith column The pixel coordinates are mapped into high-spatial resolution remote sensing block image geographic coordinates according to the sum of the pixel coordinates, the resolution product and the minimum value; Acquiring minimum value of longitude of ith row and ith column sub-block Minimum value of latitude High spectral resolution remote sensing segmented image corrected by jth row and ith column The pixel coordinates map the hyperspectral resolution remote sensing block images into geographic coordinates according to the sum of the pixel coordinates, the resolution product and the minimum value; And combining the high-spatial resolution remote sensing segmented image geographic coordinates corresponding to the jth row and ith column sub-blocks with the high-spectral resolution remote sensing segmented image geographic coordinates.
  4. 4. The remote sensing image fusion method considering image distortion as claimed in claim 3, wherein: and 4, constructing an affine transformation model according to a plurality of groups of point pairs with the same name of a plurality of blocks, wherein the specific process is as follows: Constructing a global affine transformation model according to the geographic coordinates of the high-spatial resolution remote sensing segmented images and the geographic coordinates of the high-spectral resolution remote sensing segmented images corresponding to the same-name point pairs of all the segmented blocks, and solving affine transformation model parameters by adopting a least square method; And step 4, carrying out affine transformation processing on the corrected hyperspectral resolution image according to an affine transformation model, wherein the specific process is as follows: combining affine transformation model parameters And carrying out affine transformation correction on the corrected hyperspectral resolution image by adopting an indirect correction method to obtain an affine transformed hyperspectral resolution image.
  5. 5. The remote sensing image fusion method considering image distortion as claimed in claim 4, wherein: and 5, calculating the intersected polygons after updating the corrected high-spatial-resolution image and the affine transformed high-spectral-resolution image, wherein the specific process is as follows: Constructing corrected high-spatial resolution image vector polygons according to the corrected high-spatial resolution image four-corner geographic coordinates, constructing affine transformed high-spectral resolution image vector polygons according to the affine transformed high-spectral resolution image four-corner geographic coordinates, and calculating to obtain intersecting polygons between the corrected high-spatial resolution image vector polygons and the affine transformed high-spectral resolution image vector polygons by a spatial intersection calculation method, namely updating the intersecting polygons; And 5, acquiring the geographical range of the updated intersecting polygon, and constructing the minimum circumscribed rectangle of the intersecting area according to the geographical range of the updated intersecting polygon, wherein the specific process is as follows: acquiring longitude of each point and latitude of each point in the updated intersecting polygon in sequence; the longitude minimum value of each point in the updated intersecting polygon and the longitude maximum value of each point in the updated intersecting polygon are screened out from the longitudes of a plurality of points in the updated intersecting polygon, and the longitude range for constructing the updated intersecting polygon is defined as follows: [SMIN_LON,SMAX_LON] wherein SMIN_LON represents the minimum longitude value of each point in the updated intersecting polygon, and SMAX_LON represents the maximum longitude value of each point in the updated intersecting polygon; The method comprises the steps of screening out the minimum latitude value of each point in the updated intersecting polygon and the maximum latitude value of each point in the updated intersecting polygon from the latitudes of a plurality of points in the updated intersecting polygon, constructing the latitude range of the updated intersecting polygon, and defining the following steps: [SMIN_LAT,SMAX_LAT] wherein SMIN_LAT represents the minimum latitude value of each point in the updated intersecting polygon, and SMAX_LAT represents the maximum latitude value of each point in the updated intersecting polygon; taking (smax_lon, smin_lat), (smax_lon, smax_lat), (smax_lon, smin_lat) as four vertices of the minimum bounding rectangle of the updated intersecting polygon to construct the minimum bounding rectangle of the updated intersecting polygon; and 5, performing grid division processing on the minimum circumscribed rectangle of the updated intersecting polygon to obtain a plurality of blocks of the minimum circumscribed rectangle, wherein the blocks are specifically as follows: setting the number of dividing grids in the longitudinal direction as nSBK _lon and the number of dividing grids in the latitudinal direction as nSBK _lat; The longitude width of each minimum circumscribed rectangle block and the latitude height of each minimum circumscribed rectangle block are calculated, and the method specifically comprises the following steps: Wherein, the For the longitude width of each minimum bounding rectangle partition, For the latitude height of each minimum circumscribed rectangular block, nSBK _lon is the number of longitudinal grid divisions, nSBK _lat is the number of latitudinal grid divisions, SMIN_LON represents the minimum value of the longitude of each point in the updated intersecting polygon, SMAX_LON represents the maximum value of the longitude of each point in the updated intersecting polygon, SMIN_LAT represents the minimum value of the latitude of each point in the updated intersecting polygon, and SMAX_LAT represents the maximum value of the latitude of each point in the updated intersecting polygon; The smallest circumscribed rectangular block of the q-th row and p-th column is defined as: , wherein p represents a block number in the longitudinal direction, q represents a block number in the latitudinal direction, nSBK _lon represents a division number of a directional grid, nSBK _lat represents a division number of a latitudinal grid Sequentially calculating a longitude minimum value, a longitude maximum value, a latitude minimum value and a latitude maximum value of the minimum circumscribed rectangle of the q-th row and p-th column, wherein the longitude minimum value, the longitude maximum value, the latitude minimum value and the latitude maximum value of the minimum circumscribed rectangle are specifically as follows: Wherein, the The minimum value of the longitude of the smallest circumscribed rectangular block representing the q-th row and p-th column, Represents the longitude maximum of the smallest circumscribed rectangular block of the q-th row and p-th column, Representing the minimum value of the latitude of the smallest circumscribed rectangular block of the q-th row and p-th column, Represents the latitude maximum of the smallest circumscribed rectangular block of the q-th row and p-th column, SMIN_LON represents the minimum longitude value of each point in the updated intersecting polygon, SMAX_LON represents the maximum longitude value of each point in the updated intersecting polygon, SMIN_LAT represents the minimum latitude value of each point in the updated intersecting polygon, SMAX_LAT represents the maximum latitude value of each point in the updated intersecting polygon, and the method is to be used 、 )、( 、 )、( 、 )、 、 ) As four vertexes to construct a geographic rectangular block corresponding to the smallest bounding rectangular block of the qth row and p column.
  6. 6. The remote sensing image fusion method considering image distortion as claimed in claim 5, wherein: And step 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: Dividing four vertexes according to the minimum circumscribed rectangle of the q-th row and the p-th column 、 )、( 、 )、( 、 )、 、 ) The composed geographic range, and obtaining image content corresponding to the geographic range from the corrected high-spatial resolution remote sensing image; the corrected high spatial resolution remote sensing segmented image of row q and column p is defined as: , wherein p represents a block number in the longitudinal direction, q represents a block number in the latitudinal direction, nSBK _lon represents the number of division of the latitudinal direction into grids, and nSBK _lat represents the number of division of the latitudinal direction into grids; and step 6, mapping to the affine transformed hyperspectral resolution remote sensing image to obtain each corrected high spatial resolution remote sensing segmented image, wherein the specific process is as follows: Dividing four vertexes according to the minimum circumscribed rectangle of the q-th row and the p-th column 、 )、( 、 )、( 、 )、 、 ) The composed geographic range, and obtaining image content corresponding to the geographic range from the corrected high-spatial resolution remote sensing image; The high-spatial resolution remote sensing segmented image of affine transformation of the q-th row and p-th column is defined as: , wherein p represents a block number in the longitudinal direction, q represents a block number in the latitudinal direction, nSBK _lon represents a direction division number of the grid, and nSBK _lat represents a latitudinal direction division number of the grid.
  7. 7. The remote sensing image fusion method considering image distortion as claimed in claim 6, wherein: and step 6, obtaining a plurality of groups of homonymous point pairs of each block through fine matching, wherein the specific process is as follows: Corrected high spatial resolution remote sensing segmented image for row q and column p The validity determination is carried out as follows: If it is Then ; If it is Then 0; Wherein, the High spatial resolution remote sensing segmented image representing correction of row and column q The number of pixels having a value greater than 0, Represents a partition validity determination threshold value, Representing the effectiveness attribute of the q-th row and p-th column corrected high-spatial resolution remote sensing segmented image; If it is Corrected high spatial resolution remote sensing segmented image of row q and column p Affine transformed hyperspectral resolution remote sensing segmented image with corresponding q rows and p columns Subsequent matching is not performed; If it is Corrected high spatial resolution remote sensing segmented image of row q and column p Affine transformed hyperspectral resolution remote sensing segmented image with corresponding q rows and p columns Matching is carried out, and the matching is concretely as follows: Corrected high spatial resolution remote sensing segmented image for q row and p column Affine transformed hyperspectral resolution remote sensing segmented image with corresponding q rows and p columns Obtaining a high-spatial resolution remote sensing segmented image corrected by the row (q) and the column (p) by adopting SIFT+RANSAC matching algorithm High spectral resolution remote sensing segmented image corrected by pixel coordinates and q-th row and p-th column Pixel coordinates; And combining the pixel coordinates of the high-spatial resolution remote sensing segmented image corresponding to the p-th row and p-th column segmented image and the pixel coordinates of the affine transformed high-spectral resolution remote sensing segmented image into a plurality of groups of homonymous point pairs of the q-th row and p-th column segmented image.
  8. 8. The remote sensing image fusion method considering image distortion as claimed in claim 7, wherein: and 9, performing fusion processing to obtain a remote sensing image with fused high spatial resolution and high spectral resolution, wherein the method comprises the following steps of: Setting blue band weight as The green band weight is The weight of the red wave band is The weight of the near infrared band is ; Obtaining pixel values of each corrected high-spatial-resolution remote sensing block image and the corresponding quadratic polynomial corrected high-spectral-resolution remote sensing block image pixel by pixel, and calculating a blue-green-red three-band adjustment value corresponding to each pixel value Near infrared adjustment value corresponding to each pixel value The calculation method is as follows: Wherein, the Representing the corresponding panchromatic pixel values on the corrected high spatial resolution remote sensing segmented image, For each pixel value corresponding to the blue-green-red three-band adjustment value, For each pixel value corresponding near infrared adjustment value, For the weighted average of red band, green band and blue band corresponding to each pixel, For the weighted average of the red band, green band, blue band and near infrared band corresponding to each pixel, the calculation is as follows, Wherein, the Representing blue band pixel values in the secondary polynomial corrected hyperspectral resolution remote sensing segmented image, Representing the green band pixel value in the secondary polynomial corrected hyperspectral resolution remote sensing segmented image, Representing red band pixel values in the secondary polynomial corrected hyperspectral resolution remote sensing segmented image, Representing the near infrared band pixel value in the secondary polynomial corrected hyperspectral resolution remote sensing segmented image; blue-green-red three-band adjustment value corresponding to each pixel value Near infrared adjustment value corresponding to each pixel value The fused image value is calculated pixel by pixel, and is specifically as follows: Wherein, the Representing the blue band pixel value of the fused image, Representing the pixel value of the green wave band of the fused image, Representing the pixel value of the red wave band of the fused image, And representing the near infrared band pixel value of the fused image.
  9. 9. 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 any one of claims 1-8.

Description

Remote sensing image fusion method considering image distortion and computer readable medium Technical Field The invention belongs to the technical field of image processing, and particularly relates to a remote sensing image fusion method and a computer readable medium for considering image distortion. Background The fusion between the high spatial resolution image and the high spectral resolution image is called remote sensing image spatial spectrum fusion, and the spatial spectrum fusion is a hot spot 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, the fusion performance is greatly affected due to the difference of spatial spectrum image data sources and registration errors between images. In addition, the characteristics of wide remote sensing image coverage and large internal topography fluctuation lead to the local geometric deformation of the image to be generally different, namely the image has internal distortion, and the whole image is difficult to obtain higher registration precision by adopting the same registration model at the moment, so that the fusion effect is further restricted. How to effectively improve the registration accuracy is a difficult problem to be solved by spatial spectrum fusion. Disclosure of Invention The invention provides a remote sensing image fusion method and a computer readable medium for 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 the fusion effect, a remote sensing image fusion method considering image distortion is provided by combining a coarse registration method and a blocking fine registration method when fusion is carried out, and the method is suitable for fusion between high-spatial resolution and high-spectral resolution images and can provide technical guarantee for further application of remote sensing technology. In order to achieve the above purpose, the technical scheme provided by the invention is a remote sensing image fusion method considering image distortion, comprising the following steps: step 1, obtaining an original high-spatial-resolution remote sensing image and a corresponding RPC positioning model thereof, and performing geometric correction to obtain a corrected high-spatial-resolution remote sensing image; And 2, calculating the intersected polygons of the corrected high-spatial resolution image and the corrected high-spectral resolution image, acquiring the geographic range of the intersected polygons, and constructing the minimum circumscribed rectangle of the intersected polygons according to the geographic range of the intersected polygons. Performing grid division processing on the minimum circumscribed rectangle of the intersected polygon to obtain a plurality of blocks of the minimum circumscribed rectangle; 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 obtaining a plurality of groups of homonymous point pairs of each block through initial matching processing of each corrected high-spatial resolution remote sensing block image and each corresponding corrected high-spectral resolution remote sensing block image; Step 4, constructing an affine transformation model according to a plurality of groups of point pairs with the same name of a plurality of blocks, and carrying out affine transformation processing on the corrected hyperspectral resolution image according to the affine transformation model to obtain an affine transformed hyperspectral resolution image; And 5, calculating the updated intersecting polygon of the corrected high-spatial resolution image and the affine transformed high-spectral resolution image, acquiring the geographical range of the updated intersecting polygon, and constructing the minimum circumscribed rectangle of the updated intersecting polygon according to the geographical range of the updated intersecting polygon. Performing grid division processing on the minimum circumscribed rectangle of the updated intersecting polygon to obtain a plurality of blocks of the minimum circumscribed rectangle; 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 affine transformed high spectral resolution remote sensing image to obtain each corrected high spectral resolution rem