Aerial photogrammetry production method and system without field control points

Abstract

The application provides a method for aerial photogrammetry production without external control points, which comprises the following steps: obtaining new aerial photos according to aerial photos of a target area, querying oriented aerial photos of nearby positions in a database of oriented image information by using GPS information recorded in aerial photos, carrying out image matching between the new aerial photos and the oriented aerial photos to obtain homonymous points, obtaining corresponding three-dimensional object coordinates by using multi-photo space forward intersection, carrying out template matching based on gray scale to obtain connecting points in overlapping ranges between the new aerial photos, obtaining image point coordinates of the connecting points on different new aerial photos, taking the matching points with the obtained object coordinates as control points, taking corresponding image point coordinates on the new aerial photos and the image point coordinates of the connecting points as observation values, taking exterior orientation elements of the new aerial photos as unknowns, and carrying out adjustment to obtain the exterior orientation elements of the new aerial photos, so as to realize aerial photogrammetry production.

Description

Technical Field

[0001] This invention belongs to the field of remote sensing and mapping technology, specifically relating to a method for aerial photogrammetry processing without field control points based on oriented aerial photographs. Background Technology

[0002] Aerial photogrammetry, utilizing aerial cameras on aircraft and other flying platforms, continuously captures aerial images of the ground and processes them to obtain accurate orthophotos, serving as a crucial means of acquiring geographic information. The most critical technology in aerial photogrammetry production is image orientation, which involves calculating the position and attitude of the aerial image during imaging. The professional term for the position and attitude of the aerial image is exterior orientation elements, including linear elements (X, Y, Z coordinates) describing spatial position and angular elements (rotation angles ψ, ω, κ with the three coordinate axes) describing attitude. With exterior orientation elements, the ground coordinates of any image point can be calculated according to the imaging equation. In traditional aerial photogrammetry processing methods, image orientation is achieved through aerial triangulation using rationally distributed ground control points. The specific operation involves two steps: first, reasonable control points are manually selected within the measurement area, and their geographic coordinates are measured on-site using a total station, theodolite, or GPS device; then, the image point positions of the control points are specified in the aerial image, and the exterior orientation elements of the image are solved mathematically (professionally termed bundle adjustment). This operational method is already very mature and has been used in actual production. However, this method requires on-site measurement of control points for each production run, which is extremely labor-intensive and resource-intensive. This is especially true in complex and challenging geographical environments (such as Mount Everest surveying), posing significant risks and difficulties to surveying and mapping production. On the other hand, during the research process of this invention, it was noted that in today's era of digital photogrammetry, with the continuous deepening of traditional aerial photogrammetry work and the constant accumulation and updating of measurement data, surveying and mapping production units have collected massive amounts of comprehensive image data. How to rationally utilize this geographically encoded data to conduct aerial photogrammetry production with minimal or even no control is of extremely important research value and application prospects in the current era of big data, but currently, no relevant practical technical solutions have emerged. Summary of the Invention

[0003] This invention addresses the problem that existing photogrammetric production methods require measured control points, and proposes an aerial photogrammetric production method without field control points based on oriented aerial photographs.

[0004] The technical solution of this invention provides an aerial photogrammetry production method without field control points, comprising the following steps:

[0005] Step 1: Based on the newly captured aerial photographs of the target area, use the GPS information recorded during the aerial photography to search for oriented aerial photographs of nearby locations in the oriented image information database.

[0006] Step 2: Based on the adjacency relationship between the newly captured aerial photographs and the oriented aerial photographs obtained in Step 1, perform image matching between the newly captured aerial photographs and the oriented aerial photographs to obtain corresponding points;

[0007] Step 3: Based on the corresponding image points of the three oriented aerial photographs obtained in Step 2, the corresponding object-side three-dimensional coordinates are obtained by using the forward intersection of multiple images based on their image point coordinates and exterior orientation elements.

[0008] Step 4: Extract Harris feature points from the affine transformation image corresponding to each new aerial photograph, and perform gray-scale template matching again on the affine transformation images of other new aerial photographs to obtain the connection points within the overlapping range between the new aerial photographs, and obtain the image point coordinates of the connection points on different new aerial photographs.

[0009] Step 5: Using the matching points of the obtained object coordinates as control points, the corresponding image point coordinates and the image point coordinates of the connecting points on the newly captured aerial photograph as observation values, and the exterior orientation elements of the newly captured aerial photograph as unknowns, perform bundle adjustment on the survey area and obtain the residuals of each control point. Remove points with large residuals, perform adjustment again, and finally obtain the exterior orientation elements of the newly captured aerial photograph, thus realizing aerial photogrammetry production.

[0010] Furthermore, step 1 is implemented by using the GPS coordinates (X) of the newly captured aerial photograph. gps ,Y gps Z gps Iterate through the oriented aerial photograph data, calculate the Euclidean distance between the GPS coordinates and the line element of the exterior orientation element of the oriented aerial photograph, search for oriented aerial photographs near the newly captured aerial photograph, and retain at least three of the closest oriented aerial photographs for each newly captured aerial photograph.

[0011] Furthermore, step 2, the matching of the newly captured aerial photographs with the already oriented aerial photographs, includes the following steps:

[0012] Step 2.1: For each newly captured aerial photograph, take the three oriented aerial photographs corresponding to it as the reference plane, and use the collinearity condition equation to project the oriented aerial photographs onto the reference plane to obtain the newly generated image.

[0013] Step 2.2: Extract SIFT feature points from the newly captured aerial photograph and the oriented aerial photograph with the closest Euclidean distance and perform matching. When matching, the Euclidean distance of the key point feature vectors is used as the similarity metric for key points in the two images. Several corresponding points with the highest matching degree are retained. The image point coordinates in the new aerial photograph and the oriented aerial photograph are obtained respectively. The two images are approximated as affine transformations, and the affine transformation parameters are obtained using the affine transformation formula.

[0014] The newly captured aerial photographs are subjected to an overall affine transformation based on the affine transformation parameters to obtain a corresponding approximate image. After the transformation, the approximate overlapping part with the image generated corresponding to the oriented aerial photograph is found, and a maximum rectangular region is extracted.

[0015] Step 2.3: Extract Harris feature points from the affine transformed image corresponding to the new aerial photograph. Then, perform grayscale-based template matching in the overlapping area of ​​the image generated from the three oriented aerial photographs. If a matching point can be found in all three oriented aerial photographs, the Harris feature point is retained; otherwise, it is discarded. After obtaining the corresponding points of the affine image of the new aerial photograph and the oriented aerial photographs, the image point coordinates of the corresponding points in the new aerial photograph are calculated in reverse according to the affine transformation parameters.

[0016] Furthermore, in step 3, when performing spatial forward intersection of the matched corresponding points on the oriented aerial photographs, the strict solution method of the collinearity equation is used to establish the collinearity condition equation for each oriented aerial photograph, and the least squares method is used to solve for the object space three-dimensional spatial coordinates corresponding to the matched corresponding image points.

[0017] Furthermore, the method for obtaining the exterior orientation elements of the new aerial photograph in step 5 is as follows:

[0018] Using the image point coordinates of corresponding points as observations, a bundle adjustment is performed. Assuming the interior orientation elements are known, the error equation is expressed as:

[0019]

[0020] The constant term l in the formula x =x-(x),l y = y - (y), where x and y are the observed values, and the corresponding correction is v. x ,v y ;(x) and (y) are the results obtained by substituting the initial values ​​of the unknowns into the collinear equations; X, Y, Z are the coordinates of the ground points; X S ,Y S Z S , ω and κ are parameters to be determined, and the corresponding correction is ΔX. S ,ΔY S ,ΔZ S , Δω,Δκ;a ij Let i = 1, 2, j = 1, 2, 3, 4, 5, 6.

[0021] The exterior orientation elements of the newly captured aerial photographs are obtained by solving the error equations using the bundle adjustment method. After each bundle adjustment, the coordinate corrections of the undetermined points are treated as random variables. In the least squares adjustment calculation, the variance-covariance matrix of the coordinate corrections is calculated. If the coordinate error is greater than 3 times the error of the unit weight observation, it is deleted accordingly. The remaining matching points are then adjusted again using the bundle adjustment method. This process is repeated until the accuracy of all points meets the preset requirements and converges, resulting in exterior orientation elements of the new aerial photographs with higher accuracy. These elements are then used for subsequent aerial photogrammetry data processing.

[0022] On the other hand, the present invention provides an aerial photogrammetry production system without field control points, for implementing the aerial photogrammetry production method without field control points as described above.

[0023] Moreover, it includes the following modules,

[0024] The first module is used to search for oriented aerial photos of nearby locations in the oriented image information database based on newly captured aerial photos obtained from aerial photography of the target area, using GPS information recorded during aerial photography.

[0025] The second module is used to perform image matching between the newly captured aerial images and the oriented aerial images obtained from the first module, based on the adjacency relationship between the newly captured aerial images and the oriented aerial images, to obtain corresponding points;

[0026] The third module is used to obtain the corresponding three-dimensional object coordinates by using the forward intersection of multiple images based on the image points of the three oriented aerial photographs obtained in the second module, according to their image point coordinates and exterior orientation elements.

[0027] The fourth module is used to extract Harris feature points from the affine transformation image corresponding to each new aerial photograph, and then perform gray-scale template matching again on the affine transformation images of other new aerial photographs to obtain the connection points within the overlapping range between the new aerial photographs, and to obtain the image point coordinates of the connection points on different new aerial photographs.

[0028] The fifth module is used to take the matching points of the obtained object coordinates as control points, the corresponding image point coordinates and the image point coordinates of the connecting points on the newly taken aerial photograph as observation values, and the exterior orientation elements of the newly taken aerial photograph as unknowns. The module performs bundle adjustment on the survey area and obtains the residuals of each control point. Points with large residuals are removed, and the adjustment is performed again. Finally, the exterior orientation elements of the newly taken aerial photograph are obtained, thus realizing aerial photogrammetry production.

[0029] Alternatively, it may include a processor and a memory, the memory being used to store program instructions, and the processor being used to call the stored instructions in the memory to execute an aerial photogrammetry production method without field control points as described above.

[0030] Alternatively, it may include a readable storage medium storing a computer program that, when executed, implements an aerial photogrammetry production method without field control points as described above.

[0031] Compared with existing technologies, this invention eliminates the need for manual selection of control points in the field, greatly reducing the amount of manual work, improving the speed and efficiency of aerial photogrammetry processing, effectively utilizing existing data, improving data information utilization, avoiding repetitive work and data redundancy, and meeting the requirements of information utilization in the era of big data. Attached Figure Description

[0032] Figure 1 This is a flowchart of an embodiment of the present invention.

[0033] Figure 2 This is a schematic diagram of the horizontal image of the reference plane corresponding to the oriented aerial photograph in an embodiment of the present invention.

[0034] Figure 3 This is a schematic diagram of a forward rendezvous of oriented aerial photographs according to an embodiment of the present invention. Detailed Implementation

[0035] This invention provides a method for aerial photogrammetry production without field control points based on oriented aerial photographs. The method first uses newly captured aerial photographs of the target area and GPS information recorded during aerial photography to search for oriented aerial photographs of nearby locations in an existing image database. Then, image matching is performed between the newly captured and oriented aerial photographs. Based on the matching results, the geographic coordinates of the matching points are calculated using parameters from the oriented aerial photographs, and these coordinates are used as control points for the newly captured aerial photographs. Finally, aerial triangulation is performed using the acquired control points and the newly captured aerial photographs. Through a specific adjustment algorithm, the exterior orientation elements of the newly captured aerial photographs can be calculated.

[0036] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0037] like Figure 1 As shown, this embodiment of the invention provides an aerial photogrammetry production method without field control points, the process of which includes the following steps:

[0038] Step 1: Based on the newly captured aerial photographs of the target area, use the GPS information recorded during the aerial photography to search for oriented aerial photographs of nearby locations in the oriented image information database.

[0039] In this embodiment, this step is specifically implemented as follows: according to the GPS coordinates (X) of the newly captured aerial photograph. gps ,Y gps Z gps Step 1: Traverse the oriented aerial image data, calculate the Euclidean distance between the GPS coordinates and the line elements of the exterior orientation elements of the oriented aerial image, search for oriented aerial images near the newly captured aerial image, and retain at least three nearest oriented aerial images for each newly captured aerial image. This ensures that there are enough corresponding points to match with the newly captured aerial image to achieve a certain level of accuracy, while minimizing computational load and shortening the processing time. Step 2: Based on the adjacency relationship between the newly captured aerial image and the oriented aerial images obtained in Step 1, perform image matching between the newly captured aerial image and the oriented aerial images to obtain corresponding points.

[0040] In this embodiment, the matching of the newly captured aerial photographs with the already oriented aerial photographs in step 2 is preferably achieved using the following steps:

[0041] Step 2.1: For each newly captured aerial photograph, the relative flight altitude and exterior orientation elements of the corresponding three oriented aerial photographs are known. However, since the photographs were not perfectly horizontal during capture, direct matching of corresponding points between the oriented aerial photographs yields poor results. Therefore, based on their relative flight altitude and exterior orientation elements, it is assumed that the Z-axis of the aerial camera was at the time of image capture. gps The difference between the coordinates and the relative flight altitude is the average ground elevation. For three oriented aerial photographs, using the average ground elevation as the reference plane, the oriented photographs are projected onto the reference plane using the collinearity condition equation to obtain a new image. The formula is:

[0042]

[0043] Where x, y are the image plane coordinates of the image point; x0, y0, f are the interior orientation elements of the image; X S ,Y S Z S X, Y, Z are the object space coordinates of the camera station; a is the object space coordinates of the object point. i ,b i ,c i (i = 1, 2, 3) are the three exterior orientation elements of the image. The nine direction cosines composed of κ and ω are as follows:

[0044]

[0045] Setting Z equal to the average elevation yields the X and Y coordinates of each pixel in the oriented aerial photograph, thus creating a new image. Then, the ground regions corresponding to the three aerial photographs are represented, and their overlapping areas are identified, such as... Figure 2 As shown, S1, S2, and S3 are the photographic centers of the three oriented aerial photographs closest to each newly taken aerial photograph;

[0046] Step 2.2: Extract SIFT feature points from the newly captured aerial photograph and the oriented aerial photograph with the closest Euclidean distance, and perform matching. During matching, the Euclidean distance between the keypoint feature vectors is used as the similarity metric for keypoints in the two images. Only a small number of points with the highest matching degree are retained. The image coordinates of these points in the new and oriented aerial photographs are obtained respectively. The two images are approximated by an affine transformation, and the affine transformation parameters are obtained using the affine transformation formula, as follows:

[0047]

[0048] Where (x',y′) are the image point coordinates before transformation, (x,y) are the corresponding image point coordinates after transformation, and m0,m1,m2,n0,n1,n2 are affine transformation parameters, which can be solved using the least squares method. The new aerial photograph is subjected to an overall affine transformation according to the affine transformation parameters to obtain its corresponding approximate image, so that it has a similar size and angle to the image generated from the already oriented aerial photograph. After transformation, the approximate overlapping part with the corresponding image of the already oriented aerial photograph is found, and a maximum rectangular region is extracted.

[0049] Step 2.3: Extract Harris feature points from the affine transformed image corresponding to the newly captured aerial photograph. To avoid too many extracted points and uneven distribution, it is preferable to set the minimum interval of extracted feature points to 81 pixels, thus obtaining a series of Harris feature points. Then, perform grayscale-based template matching between these points and the images generated from the three oriented aerial photographs in the overlapping area. Since the images have similar scales and orientations, matching is very easy. If a Harris feature point can find a matching point in all three oriented aerial photographs, it is retained; otherwise, it is discarded. Thus, the corresponding points of the affine image of the new aerial photograph and the oriented aerial photographs can be obtained, and the image point coordinates of the corresponding points in the new aerial photograph can be calculated in reverse according to the affine transformation parameters.

[0050] Step 3: Based on the corresponding image points of the three oriented aerial photographs obtained in Step 2, the corresponding three-dimensional object coordinates are obtained using the multi-image spatial forward intersection method according to their image point coordinates and exterior orientation elements.

[0051] like Figure 3 As shown, S1, S2, and S3 are the photographic centers of three oriented aerial photographs, p1(x1,y1), p2(x2,y2), and p3(x3,y3) are the coordinates of the object point P(X,Y,Z) in the three images, and S1-X1Y1Z1, S2-X2Y2Z2, and S3-X3Y3Z3 are local spatial coordinate systems. Furthermore, the embodiment utilizes a rigorous solution to the collinearity equation to perform spatial forward intersection of matched points on the oriented aerial photographs. For each oriented aerial photograph, the collinearity condition equation can be obtained as follows:

[0052]

[0053] Where x, y are the image plane coordinates of the image point; x0, y0, f are the interior orientation elements of the image; X S ,Y S Z S Let X, Y, and Z be the object space coordinates of the camera station; X, Y, and Z are the object space coordinates of the object point, which are unknowns to be determined; a i ,b i ,c i (i=1,2,3) has the same meaning and values ​​as in equation (1).

[0054] The equations are linear with respect to the unknowns X, Y, and Z. For a pair of corresponding points on three oriented aerial photographs, six of the above linear equations can be written. Since there are three unknowns, the least squares method can be used to solve this system of linear equations to obtain the object-space coordinates of the matched corresponding image points.

[0055] Step 4: The Harris feature points extracted from the affine transformed image corresponding to each new aerial photograph are then subjected to grayscale-based template matching with the affine transformed images of other new aerial photographs to obtain the connection points within the overlapping range between the new aerial photographs. The image point coordinates of the connection points on different new aerial photographs are then obtained for subsequent calculations.

[0056] Step 5: Using the matching points of the obtained object coordinates as control points, the corresponding image point coordinates and the image point coordinates of the connecting points on the newly captured aerial photograph as observation values, and the exterior orientation elements of the newly captured aerial photograph as unknowns, perform bundle adjustment on the survey area and obtain the residuals of each control point. Remove points with large residuals, perform adjustment again, and finally obtain the exterior orientation elements of the newly captured aerial photograph, thus realizing aerial photogrammetry production.

[0057] The present invention further proposes that, in step 5, the image point coordinates of corresponding points are used as observations for bundle adjustment of the local area network. When the interior orientation elements are considered known, the error equation can be expressed as:

[0058]

[0059] The constant term l in the formula x =x-(x),l y = y - (y), where x and y are the observed values, and the corresponding correction is v. x ,v y ;(x) and (y) are the results obtained by substituting the initial values ​​of the unknowns into the collinear equations; X, Y, Z are the coordinates of the ground points; X S ,Y S Z S , ω and κ are parameters to be determined, and the corresponding correction is ΔX. S ,ΔY S ,ΔZ S , Δω,Δκ;a ij Let be a series of partial derivatives, i = 1, 2, j = 1, 2, 3, 4, 5, 6. Specifically:

[0060]

[0061]

[0062]

[0063]

[0064]

[0065]

[0066]

[0067]

[0068]

[0069]

[0070]

[0071]

[0072] The exterior orientation elements of the newly captured aerial photograph can be obtained by solving the error equation. Since Harris automatically matches a large number of image points and there is a certain probability of matching non-corresponding points, an iterative solution method is adopted. After each bundle adjustment, the coordinate correction of the undetermined point is treated as a random variable. In the least squares adjustment calculation, the variance-covariance matrix of the coordinate correction is calculated. If the coordinate error is greater than 3 times the error of the unit weight observation, it is deleted accordingly. The remaining matching points are used to perform bundle adjustment again in this way. This process is repeated several times (usually 3-5 times) until the accuracy of all points meets the preset requirements and converges. This yields the exterior orientation elements of the new aerial photograph with higher accuracy, which can then be used for subsequent aerial photogrammetry data processing.

[0073] In specific implementation, the method proposed in the technical solution of this invention can be automatically executed by those skilled in the art using computer software technology. System devices for implementing the method, such as computer-readable storage media storing the corresponding computer program of the technical solution of this invention and computer equipment including the computer program running the corresponding computer program, should also be within the protection scope of this invention.

[0074] In some possible embodiments, an aerial photogrammetry production system without field control points is provided, comprising the following modules:

[0075] The first module is used to search for oriented aerial photos of nearby locations in the oriented image information database based on newly captured aerial photos obtained from aerial photography of the target area, using GPS information recorded during aerial photography.

[0076] The second module is used to perform image matching between the newly captured aerial images and the oriented aerial images obtained from the first module, based on the adjacency relationship between the newly captured aerial images and the oriented aerial images, to obtain corresponding points;

[0077] The third module is used to obtain the corresponding three-dimensional object coordinates by using the forward intersection of multiple images based on the image points of the three oriented aerial photographs obtained in the second module, according to their image point coordinates and exterior orientation elements.

[0078] The fourth module is used to extract Harris feature points from the affine transformation image corresponding to each new aerial photograph, and then perform gray-scale template matching again on the affine transformation images of other new aerial photographs to obtain the connection points within the overlapping range between the new aerial photographs, and to obtain the image point coordinates of the connection points on different new aerial photographs.

[0079] The fifth module is used to take the matching points of the obtained object coordinates as control points, the corresponding image point coordinates and the image point coordinates of the connecting points on the newly taken aerial photograph as observation values, and the exterior orientation elements of the newly taken aerial photograph as unknowns. The module performs bundle adjustment on the survey area and obtains the residuals of each control point. Points with large residuals are removed, and the adjustment is performed again. Finally, the exterior orientation elements of the newly taken aerial photograph are obtained, thus realizing aerial photogrammetry production.

[0080] In some possible embodiments, an aerial photogrammetry production system without field control points is provided, including a processor and a memory. The memory is used to store program instructions, and the processor is used to call the stored instructions in the memory to execute an aerial photogrammetry production method without field control points as described above.

[0081] In some possible embodiments, an aerial photogrammetry production system without field control points is provided, including a readable storage medium on which a computer program is stored. When the computer program is executed, it implements an aerial photogrammetry production method without field control points as described above.

[0082] The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to substitute them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

Claims

1. A method for aerial photogrammetry production without field control points, characterized in that, Includes the following steps: Step 1: Based on the newly captured aerial photographs of the target area, use the GPS information recorded during the aerial photography to search for oriented aerial photographs of nearby locations in the oriented image information database. Step 2: Based on the adjacency relationship between the newly captured aerial photographs and the oriented aerial photographs obtained in Step 1, perform image matching between the newly captured aerial photographs and the oriented aerial photographs to obtain corresponding points; Step 3: Based on the corresponding image points of the three oriented aerial photographs obtained in Step 2, the corresponding object-side three-dimensional coordinates are obtained by using the forward intersection of multiple images based on their image point coordinates and exterior orientation elements. Step 4: Extract Harris feature points from the affine transformation image corresponding to each new aerial photograph, and perform gray-scale template matching again on the affine transformation images of other new aerial photographs to obtain the connection points within the overlapping range between the new aerial photographs, and obtain the image point coordinates of the connection points on different new aerial photographs. Step 5: Using the matching points of the obtained object coordinates as control points, the corresponding image point coordinates and the image point coordinates of the connecting points on the newly captured aerial photograph as observation values, and the exterior orientation elements of the newly captured aerial photograph as unknowns, perform bundle adjustment on the survey area and obtain the residuals of each control point. Remove points with large residuals, perform adjustment again, and finally obtain the exterior orientation elements of the newly captured aerial photograph, thus realizing aerial photogrammetry production.

2. The aerial photogrammetry production method without field control points as described in claim 1, characterized in that: Step 1 is implemented by using the GPS coordinates of the newly captured aerial photograph. Iterate through the oriented aerial photograph data, calculate the Euclidean distance between the GPS coordinates and the line element of the exterior orientation element of the oriented aerial photograph, search for oriented aerial photographs near the newly captured aerial photograph, and retain at least three of the closest oriented aerial photographs for each newly captured aerial photograph.

3. The aerial photogrammetry production method without field control points as described in claim 1, characterized in that: Step 2, the matching of the newly captured aerial photographs with the already oriented aerial photographs, includes the following steps: Step 2.1: For each newly captured aerial photograph, take the three oriented aerial photographs corresponding to it as the reference plane, and use the collinearity condition equation to project the oriented aerial photographs onto the reference plane to obtain the newly generated image. Step 2.2: Extract SIFT feature points from the newly captured aerial photograph and the oriented aerial photograph with the closest Euclidean distance and perform matching. When matching, the Euclidean distance of the key point feature vectors is used as the similarity metric for key points in the two images. Several corresponding points with the highest matching degree are retained. The image point coordinates in the new aerial photograph and the oriented aerial photograph are obtained respectively. The two images are approximated as affine transformations, and the affine transformation parameters are obtained using the affine transformation formula. The newly captured aerial photographs are subjected to an overall affine transformation based on the affine transformation parameters to obtain a corresponding approximate image. After the transformation, the approximate overlapping part with the image generated corresponding to the oriented aerial photograph is found, and a maximum rectangular region is extracted. Step 2.3: Extract Harris feature points from the affine transformed image corresponding to the new aerial photograph. Then, perform grayscale-based template matching in the overlapping area of ​​the image generated from the three oriented aerial photographs. If a matching point can be found in all three oriented aerial photographs, the Harris feature point is retained; otherwise, it is discarded. After obtaining the corresponding points of the affine image of the new aerial photograph and the oriented aerial photographs, the image point coordinates of the corresponding points in the new aerial photograph are calculated in reverse according to the affine transformation parameters.

4. The aerial photogrammetry production method without field control points as described in claim 1, characterized in that: In step 3, when performing spatial forward intersection of the matched corresponding points on the oriented aerial photographs, the rigorous solution method of the collinearity equation is used to establish the collinearity condition equation for each oriented aerial photograph, and the object space three-dimensional spatial coordinates corresponding to the matched corresponding image points are obtained by solving the least squares method.

5. A method for aerial photogrammetry production without field control points as described in claim 1, 2, 3, or 4, characterized in that: In step 5, the outer bearing elements of the new aerial photograph are obtained as follows: Using the image point coordinates of corresponding points as observations, a bundle adjustment is performed. Assuming the interior orientation elements are known, the error equation is expressed as: constant term in the formula , For the observed value, the corresponding correction is: ; and It is the result obtained by substituting the initial value of the unknown into the collinear equation; These are the coordinates of a point on the ground. For parameters to be determined, the corresponding correction values ​​are: ; Let be a series of partial derivatives, i=1,2, j=1,2,3,4,5,6. , , , , , , , , , , , ; The exterior orientation elements of the newly captured aerial photographs are obtained by solving the error equations using the bundle adjustment method. After each bundle adjustment, the coordinate corrections of the undetermined points are treated as random variables. In the least squares adjustment calculation, the variance-covariance matrix of the coordinate corrections is calculated. If the coordinate error is greater than 3 times the error of the unit weight observation, it is deleted accordingly. The remaining matching points are then adjusted again using the bundle adjustment method. This process is repeated until the accuracy of all points meets the preset requirements and converges, resulting in exterior orientation elements of the new aerial photographs with higher accuracy. These elements are then used for subsequent aerial photogrammetry data processing.

6. An aerial photogrammetry production system without field control points, characterized in that: This method is used to implement an aerial photogrammetry production method without field control points as described in any one of claims 1-5.

7. The aerial photogrammetry production system without field control points according to claim 6, characterized in that: Includes the following modules, The first module is used to search for oriented aerial photos of nearby locations in the oriented image information database based on newly captured aerial photos obtained from aerial photography of the target area, using GPS information recorded during aerial photography. The second module is used to perform image matching between the newly captured aerial images and the oriented aerial images obtained from the first module, based on the adjacency relationship between the newly captured aerial images and the oriented aerial images, to obtain corresponding points; The third module is used to obtain the corresponding three-dimensional object coordinates by using the forward intersection of multiple images based on the image points of the three oriented aerial photographs obtained in the second module, according to their image point coordinates and exterior orientation elements. The fourth module is used to extract Harris feature points from the affine transformation image corresponding to each new aerial photograph, and then perform gray-scale template matching again on the affine transformation images of other new aerial photographs to obtain the connection points within the overlapping range between the new aerial photographs, and to obtain the image point coordinates of the connection points on different new aerial photographs. The fifth module is used to take the matching points of the obtained object coordinates as control points, the corresponding image point coordinates and the image point coordinates of the connecting points on the newly taken aerial photograph as observation values, and the exterior orientation elements of the newly taken aerial photograph as unknowns. The module performs bundle adjustment on the survey area and obtains the residuals of each control point. Points with large residuals are removed, and the adjustment is performed again. Finally, the exterior orientation elements of the newly taken aerial photograph are obtained, thus realizing aerial photogrammetry production.

8. A device, characterized in that: It includes a processor and a memory, the memory being used to store program instructions, and the processor being used to call the stored instructions in the memory to execute the aerial photogrammetry production method without field control points as described in any one of claims 1-5.

9. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which, when executed, implements an aerial photogrammetry production method without field control points as described in any one of claims 1-5.

Images