Local Orthorectification Method Based on Video Satellite Imagery and Range Vector

Abstract

This invention relates to the field of satellite image processing technology and proposes a local orthorectification method based on video satellite imagery and range vectors. It addresses the problems of existing orthorectification processes, such as errors in direct positioning of single images, poor positioning accuracy, and feature occlusion in the generated orthorectified images due to pitch and swivel angles. This invention utilizes video satellite imagery data from multiple shooting perspectives, selecting images with small pitch and swivel angles for orthorectification. An intermediate orthorectified image is generated from a single frame of the video image, and this intermediate orthorectified image is matched with a reference base map to obtain tie points. Control point information is constructed using an auxiliary DEM, and an image-side compensation parameter model is calculated. Fourier-Melin transform matching is used during the matching process to obtain high-precision, uniformly distributed tie points. This method avoids manual editing, saving time and labor costs.

Description

Technical Field

[0001] This invention relates to the field of satellite image processing technology, and specifically to a local orthorectification method based on video satellite imagery and range vectors. Background Technology

[0002] During orthorectification, orthorectified images of the entire image coverage area are usually not needed. Instead, orthorectified images of a local area are required. Therefore, after generating the orthorectified image, manual cropping is required based on the range vector to obtain the local area. This requires a lot of manpower and time. Furthermore, direct positioning of a single image has errors and low positioning accuracy. In addition, the captured images will have pitch and sway angles, and the generated orthorectified image will have feature occlusion.

[0003] The "Jilin-1" series of video satellites, independently developed by Changguang Satellite Technology Co., Ltd., has achieved meter-level resolution and an image size of 12k×5k. Its imaging mode is staring imaging, which can capture images of the same area on the ground from different angles over a long period of time to obtain satellite video. This video is composed of single frames of video images from different angles. Images with smaller pitch and sway angles are selected for orthorectification to obtain high-quality orthorectified images. In the production process, it is not necessary to cover all areas of the image. Therefore, after generating the orthorectified image, manual editing is required to crop out the required vector coverage area, which consumes a lot of time and manpower.

[0004] Currently, no one in the industry has combined range vectors with video satellite imagery for orthorectification, and has implemented fully automated local production processing during the correction process. Existing technologies first generate orthorectified images, and then use commercial software to manually edit and crop the orthorectified images and range vectors together to generate orthorectified images within the range vector coverage area. There is currently no method to automatically process range vectors, reference images, and single-frame video images together to generate locally orthorectified images. Therefore, there is a need to research a fully automated local orthorectification method based on video satellite imagery and range vectors. Summary of the Invention

[0005] To address the problems in existing orthorectification processes, such as errors in direct positioning of individual images, poor positioning accuracy, and the presence of pitch and sway angles in captured images, as well as feature occlusion in the generated orthorectified images, this invention provides a local orthorectification method based on video satellite imagery and range vectors.

[0006] A local orthorectification method based on video satellite imagery and range vectors is implemented through the following steps:

[0007] Step 1: Select a single frame of video image and process the single frame of video image to obtain the image to be processed;

[0008] Step 2: Determine the bounding rectangle of the image to be processed as the coverage area, establish an intermediate orthophoto, and process the intermediate orthophoto to obtain a consistent intermediate orthophoto:

[0009] Step 3: Obtain the coordinates of the matching point. The specific steps are as follows:

[0010] Step 3: Select N uniformly distributed pixels on the consistent intermediate orthophoto as reference points, and calculate the latitude and longitude coordinates corresponding to the N pixels based on the geographic reference information of the consistent intermediate orthophoto.

[0011] Step 32: Based on the geographic reference information of the reference base map, convert the latitude and longitude coordinates into N corresponding matching points on the reference base map; take each reference point in Step 31 as the center, and take an image sub-block with m0×n0 pixels on the consistent intermediate orthophoto; take each matching point as the center, and take an image sub-block with m0×n0 pixels on the reference base map.

[0012] Step 33: Perform Fourier Merlin transform matching on the image patch of each reference point in Step 32 and the image patch of the corresponding matching point;

[0013] Step 4: Solve the image-space compensation parameter model. The specific process is as follows:

[0014] Step 41: Set a set of image-space compensation parameters for the single-frame video image to be processed to eliminate systematic errors in the single-frame video image;

[0015] Step 42: Convert the coordinates of the matching points extracted from the consistent intermediate orthophoto into the corresponding coordinates on the single frame of the video to be processed, and then perform joint adjustment with the control point information and image compensation parameters to solve and save the parameter values ​​of the image compensation parameter model.

[0016] Step 5: Generate a local orthorectified image. The specific process is as follows:

[0017] Step 51: Using the bounding rectangle Rangle range described in Step 21 as the image coverage area, establish a local orthorectified image. The resolution GSD_orth, number of channels Channel_orth, and bit depth Deep_orth of the local orthorectified image are the same as the resolution GSD_m, number of channels Channel_m, and bit depth Deep_m of a single frame of the video image. Add georeferenced information Geo_orth to the local orthorectified image according to the resolution and image coverage area.

[0018] Step 52: Based on the georeferenced information Geo_orth from Step 51, calculate the geographic coordinates corresponding to the pixels in the local orthorectified image. Based on the video image RPC parameters and the reference DEM, calculate the pixel position of the geographic coordinates in a single frame of the video image.

[0019] Step 53: Perform bilinear interpolation on the pixel coordinates of the obtained single-frame video image to obtain the pixel value, and assign the pixel value to the pixel on the local orthorectified image.

[0020] Step 54: Traverse all pixels on the local orthorectified image, obtain their corresponding pixel values, and complete the generation of the local orthorectified image.

[0021] The beneficial effects of this invention are as follows: The orthorectification method described in this invention uses satellite video imagery, not traditional pushbroom imagery. Furthermore, by applying range vectors to the orthorectification process, it can automatically generate orthorectified images for local areas. The size of the intermediate orthorectified image is controlled by the local range vectors, improving data processing efficiency. The Fourier-Melin transform matching algorithm can obtain high-precision connection points between two images with phase differences, constructing control information and improving the positioning accuracy of single-frame video images. Orthorectification is performed on single-frame video images with the smallest pitch and yaw angles, ultimately generating high-quality locally orthorectified images without feature occlusion. Specific advantages are as follows:

[0022] First, video images have multiple shooting angles, allowing for the selection of vertically shot images with minimal feature occlusion for orthorectification, resulting in high-quality orthorectified images with minimal feature occlusion.

[0023] Second, the range vector is jointly processed during the orthorectification process. The size of the intermediate orthorectified image is controlled by the local range vector, which reduces the time consumption of image generation and downsizing calculation. The size of the final generated orthorectified image is controlled by the range vector. The local orthorectified image covered by the range vector is generated automatically without subsequent manual editing, saving a lot of time and manpower costs. Attached Figure Description

[0024] Figure 1 This is a flowchart of the local orthorectification method based on video satellite imagery and range vectors described in this invention;

[0025] Figure 2 The image shown is a video satellite image taken by Jilin-1 Video 03 satellite, serving as a rendering of the image to be processed.

[0026] Figure 3 The orthorectification effect diagram of the selected single video frame with the pitch angle closest to 0;

[0027] Figure 4 To assist in the rendering of DEM images;

[0028] Figure 5 (5a) is a schematic diagram of the reference image, and (5b) is a distribution diagram of the bounding rectangle of the range vector;

[0029] Figure 6 (6a) is a consistent intermediate orthophoto image, and (6b) is a roll-up comparison image of the consistent intermediate orthophoto image and the reference map.

[0030] Figure 7 A strategy diagram for the Fourier Merlin transform matching method;

[0031] Figure 8 (8a) shows the distribution of the tie points on the consistent intermediate orthophoto, and (8b) shows the distribution of the tie points on the reference map.

[0032] Figure 9 This is a fully automatically generated local orthorectified image.

[0033] Figure 10 This is a schematic diagram comparing a locally orthorectified image with a reference base map. Detailed Implementation

[0034] Combination Figures 1 to 8 This embodiment describes a fully automated local orthorectification method based on video satellite imagery and range vectors. In this embodiment, the range vector is imported into the production process during orthorectification. The coverage area extracted by the range vector controls the range size of the intermediate and final orthorectified images, effectively reducing time consumption and avoiding manual cropping after image generation, thus saving time and labor costs. In summary, to utilize video satellite imagery and range vectors acquired in staring mode by the Jilin-1 video satellite for automated local orthorectification and obtain high-quality locally orthorectified images without feature occlusion, this method is specifically implemented through the following steps:

[0035] 1. Select a single frame of video image, wherein the single frame of video image is a staring video satellite image;

[0036] In staring shooting mode, video satellite image data is acquired and then stabilized. The stabilized image data consists of a series of single video frames, such as... Figure 1 As shown; after acquiring a series of single-frame video images, by comparing the metadata of each single-frame video image, the image with the smallest absolute value of the pitch angle is selected as the image to be processed. The single-frame video image to be processed is set as Image_m, the number of image channels is Channel_m, the resolution of the single-frame video image is set as GSD_m, and the bit depth is Deep_m. Figure 2 As shown.

[0037] 2. Generation of intermediate orthophotos, including the following steps:

[0038] 2-1. Open the range vector Shp_m, extract the coordinates (Long_i, Lat_i), i = 1, 2, 3... of all vector points in the range vector. By comparing the magnitudes of the x and y coordinates, obtain the maximum and minimum values ​​of the x-coordinate Long_i (Long_max, Long_min) and the y-coordinate Lat_i (Lat_max, Lat_min). Construct the bounding rectangle Rangle of the range vector based on the maximum and minimum values, such as... Figure 5 As shown, the minimum value of the x-coordinate and the maximum value of the y-coordinate are combined to form the coordinates of the top left corner of the circumscribed rectangle (Long_min, Lat_max), and the maximum value of the x-coordinate and the minimum value of the y-coordinate are combined to form the coordinates of the bottom right corner of the circumscribed rectangle (Long_max, Lat_min).

[0039] 2-2. Using the bounding rectangle Rangle from step 2-1 as the coverage area, establish an intermediate orthophoto image Image_mid_orth. Set the number of channels of the intermediate orthophoto image as Channel_mid_orth, the resolution as GSD_mid_orth, and the bit depth as Deep_mid_orth. The resolution, number of channels, and bit depth of the intermediate orthophoto image are the same as those of a single frame of the video image, i.e., GSD_mid_orth = GSD_m, Channel_mid_orth = Channel_m, Deep_mid_orth = Deep_m. Add georeferenced information Geo_mid to the intermediate orthophoto image according to the resolution and the coverage area of ​​the image.

[0040] 2-3. Based on the georeferenced information Geo_mid, calculate the geographic coordinates (X, Y) corresponding to the pixel (r_mid, c_mid) in the intermediate orthophoto. i Y i Based on the RPC (Rational Polynomial Coefficients) parameters of the video imagery and the reference DEM (Digital Elevation Model), the geographic coordinates (X) are calculated. i Y i In a single frame of video imagery, the radiometric information of pixel (r_mid, c_mid) is obtained through sampling. This process is repeated for each pixel in the intermediate orthophoto imagery to obtain the radiometric information for each pixel. The calculation model between geographic coordinates and pixel positions in a single frame of video imagery is as follows:

[0041]

[0042] In the formula, l,s are the row and column values ​​of a single frame of video image obtained through the solution; Z i The elevation values ​​corresponding to the ground points provided by the DEM, R s R0, C s C0 is the normalized parameter in the RPC parameters of a single video frame, and p1 to p4 represent the forward form of the RPC polynomial.

[0043] p i (X,Y,Z)=a1+a2X+a3Y+a4Z+a5XY+a6XZ+a7YZ+a8X 2 +a9Y 2 +a 10 Z 2 +a 11 YXZ+a 12 X 3 +a 13 XY 2 +a 14 XZ 2 +a 15 X 2 Y+a 16 Y 3 +a 17 YZ 2 +a 18 X 2 Z+a 19 Y 2 Z+a 20 Z 3

[0044] Where a1-a 20 The corresponding parameter values ​​in the RPC parameter file.

[0045] 2-4. Extract the resolution GSD_map and bit depth Deep_map of the reference base image Image_map, such as... Figure 5 As shown; it checks whether the resolution of the reference base map (GSD_map) is the same as the resolution of the intermediate orthophoto (GSD_mid_orth). If they are different, the intermediate orthophoto is resampled. It then checks whether the bit depth of the reference base map (Deep_map) is the same as the bit depth of the intermediate orthophoto (Deep_mid_orth). If they are different, the intermediate orthophoto is upsampled or downsampled. Finally, after resampling and upsampling / downsampling, a consistent intermediate orthophoto (Image_s_mid_orth) with the same resolution and bit depth as the reference base map is generated, and its corresponding georeferenced information is updated to Geo_s_mid. At this point, the consistent intermediate orthophoto and the reference base map have errors; the same ground feature is misaligned in the two images, such as... Figure 6 As shown.

[0046] 3. Obtain the coordinates of the matching point, including the following steps:

[0047] 3-1. Select N uniformly distributed pixels as reference points on the consistent intermediate orthophoto image Image_s_mid_orth (r k ,c k ), k = 0, 1, ..., N, r k c is the row coordinate. k For column coordinates; based on the georeferenced information Geo_s_mid from the consistent intermediate orthophoto Image_s_mid_orth, calculate the latitude and longitude coordinates (M_Long) corresponding to N pixels. k ,M_Lat k ), k = 0, 1...N, N>6.

[0048] 3-2 Based on the georeference information Geo_map from the base map Image_map, the latitude and longitude coordinates (M_Long) will be... k ,M_Lat k Convert ) to N corresponding matching points (R) on the reference base map k C k Taking each reference point in step 3-1 as the center, take an image sub-block with m0×n0 pixels on the consistent intermediate orthophoto Image_s_mid_orth, and take an image sub-block with m0×n0 pixels on the reference base map Image_map, taking each matching point as the center.

[0049] 3-3. Perform Fourier Merlin transform matching on the image patch of each reference point in 3-2 and the image patch of the corresponding matching point, such as... Figure 7 As shown, a similarity threshold of 0.5 is set. The coordinates of matching points with similarity exceeding the threshold are retained, and erroneous matches are removed using the RANSAC algorithm. It is then determined whether the number of remaining matching point pairs after removing erroneous matches is greater than or equal to 10. If so, the location information of evenly distributed, high-precision matching points is saved, such as... Figure 8 As shown, and based on the latitude and longitude information corresponding to the matching points from the auxiliary DEM, such as... Figure 4 As shown, extract the elevation values ​​of the corresponding matching points on the reference image. The latitude and longitude coordinates and elevation values ​​of the matching points form the control point information; if not, re-enter step 3-1.

[0050] 4. Solve the image-space compensation parameter model, including the following steps:

[0051] 4-1. Set a set of image-side compensation parameters a0, a2, ... for a single frame of video image Image_m. l a s b0, b l b sIt is used to eliminate systematic errors in single-frame video images, and its specific form is as follows:

[0052]

[0053] Where (l′,s′) is the correct row and column value of a certain point T(X,Y,Z) on the ground in a single frame of video image, and (l,s) is the row and column value of point T in a single frame of video image calculated based on the RPC parameters of the single frame of video image;

[0054] 4-2. The coordinates of the matching points extracted from the consistent intermediate orthophoto Image_s_mid_orth are converted into the corresponding coordinates on the single-frame video image Image_m. Then, they are jointly adjusted with the control information and image-side compensation parameters. The adjustment model is as follows:

[0055]

[0056] Where X i Y i Z i Let X be the ground coordinates corresponding to the connection point. i Y i The coordinates of the matching points obtained from the reference base map in section 3.3 are converted based on the georeference information of the reference base map, Z. i The value is based on X i Y i The coordinates are obtained through an auxiliary DEM, R s R0, C s C0 is the normalized parameter in the video image RPC parameters, p i The positive solution form of the RPC polynomial is expressed as follows:

[0057] p i (X,Y,Z)=a1+a2X+a3Y+a4Z+a5XY+a6XZ+a7YZ+a8X 2 +a9Y 2 +a 10 Z 2 +a 11 YXZ+a 12 X 3 +a 13 XY 2 +a 14 XZ 2 +a 15 X 2 Y+a 16 Y 3 +a 17 YZ 2 +a 18 X 2 Z+a 19 Y2 Z+a 20 Z 3

[0058] Where a1-a 20 The corresponding parameter values ​​in the RPC parameter file.

[0059] 4-3 Solve and save the parameter values ​​of the image compensation parameter model.

[0060] 5. Generation of local orthophotos, including the following steps:

[0061] 5-1. Using the bounding rectangle Rangle in 2-1 as the image coverage area, establish a local orthorectified image Image_orth. The resolution GSD_orth, number of channels Channel_orth, and bit depth Deep_orth of the local orthorectified image are consistent with the relevant indicators of a single frame of the video image, i.e., GSD_orth = GSD_m, Channel_orth = Channel_m, Deep_orth = Deep_m. Add georeferenced information Geo_orth to the local orthorectified image according to the resolution and image coverage area.

[0062] 5-2. Based on the georeferenced information Geo_orth, calculate the geographic coordinates (X, Y, F) of the pixels (r_orth, c_orth) in the locally orthorectified image. orth Y orth Based on the video image RPC parameters and the reference DEM, the geographic coordinates (X) are calculated. orth Y orth The pixel position in a single frame of video image. The solution model is as follows:

[0063]

[0064] The specific parameters are the same as in 4-2, and the image compensation parameters are known values ​​at this time.

[0065] 5-3. Perform bilinear interpolation on the pixel coordinates of the solved single-frame video image to obtain the pixel value, and assign the pixel value to the pixel (r_orth, c_orth) on the local orthorectified image.

[0066] 5-4. Traverse all pixels on the local orthorectified image, obtain their corresponding pixel values, and complete the generation of the local orthorectified image.

[0067] 5-5. Output the locally orthorectified image in GeoTIFF format, such as... Figure 9 As shown, at this point, the local orthorectified image closely matches the reference base map. Figure 10 As shown.

[0068] Specific Implementation Method Two: Combination Figures 1 to 10 This embodiment is a verification example of the local orthorectification method based on video satellite imagery and range vectors described in Specific Embodiment 1: First, orthorectified images are generated using satellite video imagery.

[0069] The Fourier Merlin transform matching algorithm is used to extract high-precision control point information from the reference base map and auxiliary DEM, and generate the image-space compensation parameter model of a single frame of video image to improve positioning accuracy.

[0070] By combining range vector and orthorectified data processing, local orthorectified images can be generated automatically without subsequent manual processing, thereby improving data processing efficiency and saving manpower and time costs.

[0071] The technical method flowchart, satellite video image data, reference image and auxiliary DEM image, intermediate orthorectified image, Fourier Merlin matching strategy and results are presented in sequence, and the generated local orthorectified image results are presented at the end.

[0072] like Figure 1 The diagram shows the technical flowchart of a fully automatic local orthorectification method based on video satellite imagery and range vectors. Figure 2 The image shown is video satellite imagery data captured by Jilin-1 Video 03 satellite, which has multiple shooting perspectives; Figure 3 The image shown is a single video frame with the pitch angle closest to 0, which is then subjected to orthorectification. Figure 4 To assist in the processing of DEM image data; Figure 5 (5a) is a schematic diagram of the reference image, and (5b) is a distribution diagram of the bounding rectangle of the range vector; Figure 6 (6a) is a consistent intermediate orthophoto, and (6b) is a comparison between the consistent intermediate orthophoto and the reference map. It can be seen that the same feature has a large positional error in the two images, and the error is obvious. Furthermore, the reference map and the single frame video image have large differences in feature characteristics. This is because the shooting time of the two images is significantly different. Figure 7 A strategy diagram for the Fourier Merlin transform matching method; Figure 8 (8a) shows the distribution of the connection points on the consistent intermediate orthophoto, and (8b) shows the distribution of the connection points on the reference map. It can be seen that the connection points are all within the bounded rectangle of the range vector and have a uniform distribution. Figure 9 This is a fully automated, locally orthorectified image. Figure 10 This is a schematic diagram comparing a locally orthorectified image with a reference base map. It can be seen that the error between the two is extremely small, and the same ground feature is located in the same position in both images, with almost no misalignment.

[0073] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0074] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A local orthorectification method based on video satellite imagery and range vectors, characterized by: This method is implemented by the following steps: Step 1: Select a single frame of video image and process the single frame of video image to obtain the image to be processed; Step 2: Determine the bounding rectangle of the image to be processed as the coverage area, establish an intermediate orthophoto, and process the intermediate orthophoto to obtain a consistent intermediate orthophoto: Step 3: Obtain the coordinates of the matching point. The specific steps are as follows: Step 3: Select N uniformly distributed pixels on the consistent intermediate orthophoto as reference points, and calculate the latitude and longitude coordinates corresponding to the N pixels based on the geographic reference information of the consistent intermediate orthophoto. Step 32: Based on the geographic reference information of the reference base map, convert the latitude and longitude coordinates into N corresponding matching points on the reference base map; take each reference point in Step 31 as the center, and take an image sub-block with m0×n0 pixels on the consistent intermediate orthophoto; take each matching point as the center, and take an image sub-block with m0×n0 pixels on the reference base map. Step 33: Perform Fourier Merlin transform matching on the image patch of each reference point in Step 32 and the image patch of the corresponding matching point; Step 4: Solve the image-space compensation parameter model. The specific process is as follows: Step 41: Set a set of image-space compensation parameters for the single-frame video image to be processed to eliminate systematic errors in the single-frame video image; Step 42: Convert the coordinates of the matching points extracted from the consistent intermediate orthophoto into the corresponding coordinates on the single frame of the video to be processed, and then perform joint adjustment with the control point information and image compensation parameters to solve and save the parameter values ​​of the image compensation parameter model. Step 5: Generate a local orthorectified image. The specific process is as follows: Step 51: Using the bounding rectangle Rangle range described in Step 2 as the image coverage area, establish a local orthorectified image. The resolution GSD_orth, number of channels Channel_orth, and bit depth Deep_orth of the local orthorectified image are the same as the resolution GSD_m, number of channels Channel_m, and bit depth Deep_m of a single frame of the video image. Add georeferenced information Geo_orth to the local orthorectified image according to the resolution and image coverage area. Step 52: Based on the georeferenced information Geo_orth from Step 51 and the parameter values ​​of the image compensation parameter model from Step 42, calculate the geographic coordinates corresponding to the pixels in the local orthorectified image. Based on the video image RPC parameters and the reference DEM, calculate the pixel position of the geographic coordinates in a single frame of the video image. Step 53: Perform bilinear interpolation on the pixel coordinates of the obtained single-frame video image to obtain the pixel value, and assign the pixel value to the pixel on the local orthorectified image. Step 54: Traverse all pixels on the local orthorectified image, obtain their corresponding pixel values, and complete the generation of the local orthorectified image.

2. The local orthorectification method based on video satellite imagery and range vectors according to claim 1, characterized in that: In step one, the single-frame video image is a staring video satellite image; In staring shooting mode, video satellite image data is acquired and image stabilization is performed. The image data after image stabilization consists of a series of single-frame video images. After acquiring a series of single-frame video images, the image with the smallest absolute value of the pitch angle is selected as the image to be processed by comparing the metadata of each single-frame video image. Let the selected single-frame video image to be processed be Image_m, the number of image channels be Channel_m, the resolution of the single-frame video image be GSD_m, and the bit depth be Deep_m.

3. The local orthorectification method based on video satellite imagery and range vectors according to claim 1, characterized in that: The specific process of step two is as follows: Step 2: First, determine the range vector Shp_m, extract the coordinates (Long_i, Lat_i) of all vector points of the range vector, obtain the maximum value (Long_max) and minimum value (Long_min) of the horizontal coordinate Long_i and the maximum value (Lat_max) and minimum value (Lat_min) of the vertical coordinate Lat_i by comparing the horizontal and vertical coordinates, and construct the outer rectangle Rangle of the range vector based on the obtained maximum and minimum values ​​of the horizontal and vertical coordinates. The minimum value of the x-coordinate and the maximum value of the y-coordinate are combined to form the coordinates of the top left corner of the circumscribed rectangle Rangle (Long_min, Lat_max), and the maximum value of the x-coordinate and the minimum value of the y-coordinate are combined to form the coordinates of the bottom right corner of the circumscribed rectangle Rangle (Long_max, Lat_min). Step 22: Using the bounding rectangle Rangle determined in Step 21 as the coverage area, establish the intermediate orthophoto image Image_mid_orth; The resolution of the intermediate orthophoto Image_mid_orth is set to GSD_mid_orth, the number of channels is Chanel_mid_orth, and the bit depth is Deep_mid_orth. The resolution GSD_mid_orth, the number of channels Chanel_mid_orth, and the bit depth Deep_mid_orth of the intermediate orthophoto are all the same as the resolution GSD_m, the number of channels Chanel_m, and the bit depth Deep_m of a single frame of the video image. Georeference information Geo_mid is added to the intermediate orthophoto Image_mid_orth according to the resolution and the image coverage area. Steps 2 and 3: Based on the added georeferenced information Geo_mid, calculate the geographic coordinates (X, Y) corresponding to the pixel (r_mid, c_mid) in the intermediate orthophoto. i Y i ); Based on the RPC parameters of the single-frame video image to be processed and the auxiliary DEM image, the corresponding position of the pixel (r_mid, c_mid) in the single-frame video image is calculated, and the radiometric information of the corresponding position in the single-frame video image to be processed is assigned to the intermediate orthophoto pixel (r_mid, c_mid). Step 24: Extract the resolution GSD_map and bit depth Deep_map of the reference base map Image_map, and determine whether the resolution GSD_map of the reference base map is the same as the resolution GSD_mid_orth of the intermediate orthophoto. If they are different, resample the intermediate orthophoto. It is determined whether the depth of the reference base map (Deep_map) and the depth of the intermediate orthophoto (Deep_mid_orth) are the same. If they are different, the intermediate orthophoto is resampled or resampled. Finally, after resampling and resampling, a consistent intermediate orthophoto (Image_s_mid_orth) with the same resolution and depth as the reference base map is generated. The georeferenced information corresponding to the consistent intermediate orthophoto is updated to Geo_s_mid.

4. The local orthorectification method based on video satellite imagery and range vectors according to claim 3, characterized in that: In steps two and three, the solution model between geographic coordinates and pixel positions in a single frame of video imagery is as follows: In the formula, l,s are the row and column values ​​of a single frame of video image obtained through the solution; Z i The elevation values ​​corresponding to the ground points provided by the DEM, R s R0, C s C0 is the normalized parameter in the RPC parameters of a single video frame, and p1 to p4 represent the forward form of the RPC polynomial. p i (X,Y,Z)=a1+a2X+a3Y+a4Z+a5XY+a6XZ+a7YZ+a8X 2 +a9Y 2 +a 10 Z 2 +a 11 YXZ+a 12 X 3 +a 13 XY 2 +a 14 XZ 2 +a 15 X 2 Y+a 16 Y 3 +a 17 YZ 2 +a 18 X 2 Z+a 19 Y 2 Z+a 20 Z 3 Where a1-a 20 The corresponding parameter values ​​in the RPC parameter file.

5. The local orthorectification method based on video satellite imagery and range vectors according to claim 1, characterized in that: In step 3, the matching similarity threshold is set to 0.

5. The coordinates of matching points with similarity greater than the threshold are retained, and erroneous matching points are removed based on the RANSAC algorithm. It is then determined whether the number of remaining matching point pairs after removing erroneous matching points is greater than or equal to 10. If yes, save the location information of the evenly distributed high-precision matching points, and extract the elevation values ​​of the corresponding matching points from the reference image based on the latitude and longitude information of the matching points. The latitude and longitude coordinates and elevation values ​​of the matching points form the control point information. If no, return to step 31.

6. The local orthorectification method based on video satellite imagery and range vectors according to claim 1, characterized in that: In step four, the adjustment model is as follows: Where X i Y i Z i Let X be the ground coordinates corresponding to the connection point. i Y i Z is obtained by converting the coordinates of the matching points on the reference base map based on the georeferenced information of the reference base map. i The value is based on X i ,Y i The coordinates are obtained through an auxiliary DEM, R s ,R0,C s C0 is the normalized parameter in the video image RPC parameters, p i The positive solution form of the RPC polynomial is expressed as follows: p i (X,Y,Z)=a1+a2X+a3Y+a4Z+a5XY+a6XZ+a7YZ+a8X 2 +a9Y 2 +a 10 Z 2 +a 11 YXZ+a 12 X 3 +a 13 XY 2 +a 14 XZ 2 +a 15 X 2 Y+a 16 Y 3 +a 17 YZ 2 +a 18 X 2 Z+a 19 Y 2 Z+a 20 Z 3 Where a1-a 20 The corresponding parameter values ​​in the RPC parameter file.

Images