High-precision geological deformation monitoring method and device based on unmanned aerial vehicle image

Abstract

The invention provides a high-precision geological deformation monitoring method and device based on unmanned aerial vehicle images, and the method comprises the steps: merging two images from an unmanned aerial vehicle, and carrying out the bundle block adjustment, thereby obtaining a common connection point and a corresponding object space coordinate; splitting the two-stage image and respectively processing the two-stage image, re-adjusting the first-stage image, extracting a connection point with the minimum change with merging processing as a stable region connection point, and performing bundle adjustment on the stable region connection point as control information of the second-stage image so as to enable the stable region connection point to be registered with the first-stage image; and establishing a three-dimensional model and a digital surface model DSM by using the registered two-stage images, and respectively carrying out differential analysis to obtain geological deformation information. The method can be used for realizing high-precision ground surface deformation monitoring under the condition of no ground control point.

Description

Technical Field

[0001] The invention belongs to the technical field of remote sensing mapping, and in particular relates to a high-precision geological deformation monitoring method based on two-phase unmanned aerial vehicle images. Background Art

[0002] Geological disasters caused by geological deformation pose a serious threat to the safety of life and property of the country and the people. It is extremely important to carry out effective deformation monitoring. Traditional means generally adopt the method of manual field investigation. Leveling is a method with high deformation monitoring accuracy. It is widely used in engineering deformation monitoring such as surface deformation and building deformation, but it has high labor intensity and high economic cost. In addition, deformation monitoring can also be achieved by deploying GNSS monitoring points. Although this method has high monitoring accuracy and fast time sampling rate, it can only monitor the deformation at certain points. The spatial resolution is very low and cannot reflect the overall deformation. In recent years, InSAR technology has shown a blowout development in the field of deformation monitoring and has gradually become one of the commonly used means in the field of deformation monitoring. However, limited by the influence of SAR imaging geometry and microwave electromagnetic characteristics, even if the new generation of SAR satellites are used for monitoring, factors such as geometric distortion caused by terrain undulations, de-correlated noise caused by dense vegetation coverage, and atmospheric delay errors caused by water vapor changes still restrict the application of InSAR technology in the field of deformation monitoring. LiDAR technology can directly obtain three-dimensional point cloud data. It has the characteristics of high resolution, high precision and high efficiency in various complex terrain environments such as high steep slopes and dangerous rock masses. In areas covered with vegetation, the terrain under vegetation can be obtained, which has become the main means of producing high-precision digital elevation models (DEM). However, in complex terrain environments with high altitudes and gullies, airborne LiDAR often has GNSS signal interruption during flight and cannot continue to fly safely. On the other hand, due to the limited scanning angle and ranging accuracy of LiDAR on light and small drones, when collecting data in areas with large height differences, steep slopes and complex shapes, there will be some problems of missing point cloud data on steep slopes. In summary. Although the above methods have been developed relatively maturely through the efforts of countless scientific researchers, there are still various problems that cannot meet the current needs and requirements of geological disaster monitoring. With the development of drone technology, time-series high-resolution images can be flexibly and conveniently obtained, providing a solution for effective monitoring of geological deformation. The main process of using drones to monitor geological deformation is to first lay out a certain number of evenly distributed control points on the ground, use drones to take photos of the target area at certain time intervals to obtain multiple images, and perform aerial triangulation with control points, and then produce multiple fine three-dimensional models to find deformations by comparing the differences. However, in areas prone to geological disasters, the terrain is often complex. On the one hand, it is inconvenient to set up measurement equipment and there are great difficulties in laying out control points. On the other hand, due to the possible deformation of the surface, it is impossible to confirm whether the two control points are fixed. In addition, if the control points are re-measured every time a drone is used for photogrammetry, it will consume a lot of manpower and time, and the efficiency will be low. Therefore, the traditional drone photogrammetry method is not suitable for areas prone to geological deformation such as landslides, and new methods must be sought to solve this problem. Summary of the invention

[0003] In view of the above technical problems, the present invention introduces a deformation monitoring technical solution based on two-phase UAV images without the need for ground control points.

[0004] The present invention provides a high-precision geological deformation monitoring method based on drone images, including the following process: The two phases of images from the drone are combined to perform bundle block adjustment to obtain common tie points and corresponding object coordinates. The two phases of images are split and processed separately, the first phase of images are re-adjusted, and the connection points with the smallest change from the merged processing are extracted as stable area connection points, and used as control information for the second phase of images for bundle adjustment to align them with the first phase of images; The three-dimensional model and digital surface model (DSM) were established using the registered two-phase images, and differential analysis was performed to obtain geological deformation information.

[0005] Moreover, when the two phases of images are combined for bundle method regional block adjustment to obtain common connection points, the implementation method is to merge the two groups of images from different periods, perform joint aerial triangulation without ground control points, obtain the exterior orientation elements and image connection point information of the two phases of images, split the jointly processed images into two independent projects according to the shooting time, and retain their respective connection point information, and extract the connection points common to the first and second phases.

[0006] Furthermore, the extraction of the connection points common to the first phase and the second phase is implemented as follows: Make two copies of the merged aerial triangulation project. Delete the second-phase images and connection points from the first copy, and delete the first-phase images and connection points from the second copy. The connection points in the two projects are screened. If a connection point is included in both the first and second phase images, it is retained. If the connection point only exists in a single phase image, it is deleted. Finally, the common connection point information of the two phase images is obtained.

[0007] Moreover, the implementation of extracting the connection points of the stable region is as follows, Based on the first phase of images, independent aerial triangulation is performed without ground control points to update the image's exterior orientation parameters and tie point information; The new connection point Connecting points from the first of two phases of aerial triangulation Compare and extract the connection points with the same name; The three-dimensional coordinates of the corresponding connection points in the second-phase image are replaced with the three-dimensional coordinates in the first-phase image to obtain the stable area connection points with the first-phase image as the reference benchmark.

[0008] Moreover, the implementation of taking the connection point with the same name is as follows, The new connection point Connecting points from the first of two phases of aerial triangulation Compare and consider the connection points whose coordinate differences between images are lower than the corresponding threshold a as the connection points with the same name, and each connection point has projections in at least b images of the first and second phases; The selected connection points are marked accordingly. , and rank them according to the Euclidean distance, retaining the top c% of connection points with the smallest distance and discarding the rest, where a, b, and c are preset thresholds.

[0009] Moreover, the identified stable region connection points are further thinned out and used as control points to perform aerial triangulation on the second-phase images and update the exterior orientation elements of the images.

[0010] Moreover, it is used to achieve high-precision surface deformation monitoring without ground control points.

[0011] On the other hand, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the high-precision geological deformation monitoring method based on drone images as described above is implemented.

[0012] On the other hand, the present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored, characterized in that when the computer program is executed by a processor, the high-precision geological deformation monitoring method based on drone images as described above is implemented.

[0013] On the other hand, the present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the above-mentioned high-precision geological deformation monitoring method based on drone images.

[0014] In summary, the present invention proposes a deformation monitoring method based on two-phase UAV images without ground control points. The invention combines two phases of UAV images for aerial triangulation, extracts stable and unchanged areas from the first phase of images, and uses them as control information for the second phase of data. It can realize high-precision surface deformation monitoring without ground control points, and the monitoring accuracy of elevation changes can reach the centimeter level. Compared with the prior art, the present invention has the following advantages: in areas prone to natural geological disasters, it is difficult to deploy permanent ground control points. The present invention does not need to deploy artificial control points on the ground, which avoids a lot of field work, greatly improves safety and deformation monitoring efficiency, and can respond in a short time; the monitoring accuracy of the present invention is ten times higher than that of the traditional method without using ground control points, and is equivalent to the accuracy of using a large number of ground control points, which can reach the centimeter level; and the present invention can obtain three-dimensional surface change information of the monitoring area, and can comprehensively obtain high-precision deformation data of the monitoring area. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 Schematic diagram of the changing area of ​​an embodiment of the present invention.

[0016] Figure 2 is a flow chart of an embodiment of the present invention.

[0017] Figure 3 Schematic diagram of extracting connection points of stable regions according to an embodiment of the present invention.

[0018] Figure 4 Schematic diagram of aerial triangulation using stable connection points for the second phase of images according to an embodiment of the present invention.

[0019] Figure 5 This is a schematic diagram of gridding an experimental area according to an embodiment of the present invention.

[0020] Figure 6 A schematic diagram of relative accuracy and absolute accuracy of elevation using the method of the embodiment of the present invention and using the traditional method.

[0021] Figure 7 A schematic diagram of the physical structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The present invention introduces a deformation monitoring technology based on two-phase UAV images without ground control points. The technical solution first merges the two-phase images for bundle method regional block adjustment to obtain common connection points and corresponding object coordinates, then splits the two-phase images and processes them separately, re-adjusts the first-phase images, extracts the connection points with the smallest change from the merged processing as stable connection points, and uses them as control information for the second-phase images for bundle method adjustment to align them with the first-phase images; finally, uses the aligned two-phase images to establish a three-dimensional model and a digital surface model (DSM), and performs differential analysis on the two-phase three-dimensional models and DSM to obtain geological deformation information.

[0023] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0024] The research area used in the experiment of this embodiment is a mine. In 2020, 616 and 847 drone images of the experimental area were acquired in March 2020 and June 2020, respectively. The camera model is FC6310, the focal length is 9 mm, and the image resolution is 5472×3648 pixels. Figure 1 The figure shows the changes in the experimental area during the two impact acquisition periods. The left picture shows the situation in March, and the right picture shows the situation in June.

[0025] like Figure 2 As shown, an embodiment of the present invention provides a high-precision geological deformation monitoring method based on drone images, and the process includes the following steps: Step 1: Combine the first phase UAV images and the second phase UAV images, perform joint aerial triangulation without ground control points, and obtain the exterior orientation elements and image connection point information of the two phases of images.

[0026] The implementation method of step 1 further preferably adopted in the embodiment includes the following sub-steps: Step 1.1, use SIFT algorithm to extract feature points from two phases of drone images respectively, and calculate SIFT feature descriptors for each extracted feature point for subsequent matching. In specific implementation, the SIFT algorithm implementation method is prior art and will not be described in detail in the present invention.

[0027] Step 1.2, feature point matching is performed on the image, and the random sampling consistency algorithm is used to remove the wrong matching points, retain the correct matching point pairs, and improve the robustness of the matching result. In specific implementation, the implementation method of the random sampling consistency algorithm is the prior art, and the present invention will not be described in detail.

[0028] Step 1.3 Obtain the initial data of the shooting position and attitude of each image from the drone's track data (GPS data). This includes the coordinates of the center point of each image ( ) and attitude angle ( ) (roll angle around the X axis, pitch angle around the Y axis, yaw angle around the Z axis).

[0029] Step 1.4: All images are adjusted using the bundle method to establish an error equation that includes all images, matching points, and initial exterior orientation elements. The formula is: (1) The constant term , is the observed value of the image coordinates of the matching point, and the corresponding correction number is ; and It is the parameter calculated by substituting the initial value of the unknown into the collinear equation; is the coordinate of the ground point; is the parameter to be determined, and the corresponding correction number is ; is a series of partial derivatives, i =1,2, j =1,2,3,4,5,6. Specifically: (2) in, is the image principal point coordinate, is the camera focal length, parameter Specifically: (3) By solving the error equation, we can finally obtain the new exterior orientation elements and matching connection point information of all images. The connection points include the three-dimensional coordinates in space and the positions in the images.

[0030] Step 2: Split the jointly processed images into two independent projects according to the shooting time, while retaining their respective connection point information. Extract the connection points common to the first and second phases.

[0031] The implementation method of step 2 further preferably adopted in the embodiment includes the following sub-steps: Step 2.1, copy the merged aerial triangulation project twice, delete the second phase images and connection points from the first copy, and delete the first phase images and connection points from the second copy.

[0032] Step 2.2, the connection points in the two projects are screened. If a connection point is included in both the first and second phase images, it is retained. If the connection point only exists in a single phase image, it is deleted. Finally, the common connection point information of the two phase images is obtained.

[0033] Step 3: Use the tie points of the first phase image to perform aerial triangulation, obtain new exterior orientation elements and update tie point information. Calculate the difference in object coordinates of the common tie points in the first phase image processing alone and the joint processing of the two phases of images. The tie point with the smallest difference is considered to be the stable area tie point and used as a unified reference.

[0034] The stable region connection point extraction method further preferably adopted in the embodiment includes the following sub-steps: Step 3.1: Perform a separate aerial triangulation based on the first phase of images without ground control points to update the image's exterior orientation parameters and tie point information.

[0035] Step 3.2: Connect these new points Connecting points from the first of two phases of aerial triangulation If the image coordinates of the corresponding connection points are , the connection points whose coordinate differences between images are lower than the preset threshold a are considered to be the same-name connection points, that is, .

[0036] In this embodiment, it is preferred to set the difference a=0.01 pixel as the threshold, that is, (4) Each connection point has a projection in at least b images of the first phase and the second phase, and b=5 is preferably set in the embodiment.

[0037] The selected connection point is marked with , and rank them according to their Euclidean distance, retaining the first 50% of the connection points with the smallest distance and discarding the rest.

[0038] During specific implementation, the coordinate difference threshold a, the image number threshold b, and the connection point retention ratio threshold c can be set to other values ​​according to the situation.

[0039] Step 3.3, replace the three-dimensional coordinates of the corresponding connection points in the second phase image with the three-dimensional coordinates in the first phase image, and obtain the stable area connection points with the first phase image as the reference benchmark.

[0040] Figure 3 The process of extracting connection points in stable areas is demonstrated: first, the two phases of images are merged to perform aerial triangulation and obtain connection points; then, aerial triangulation is re-performed using the connection points of the first phase of images to obtain new exterior orientation elements and update the connection point information; finally, the connection points are screened to extract relatively stable connection points. Figure 3 The coordinate system in is the geodetic photogrammetric coordinate system. Points in the specified direction of the coordinate system.

[0041] Step 4: Thin out the identified stable area connection points and use them as control points to perform aerial triangulation on the second-phase image to update the exterior orientation elements of the image.

[0042] like Figure 4 As shown, the implementation method of step 4 further preferably adopted in the embodiment includes the following sub-steps: Step 4.1, such as Figure 5 As shown in the figure, the research area to be monitored for deformation is divided into regular rectangular grid blocks, and appropriate and evenly distributed stable area connection points are selected in each grid block as control points for the second phase of images. The division rule is: First, count the research area The maximum and minimum values ​​of the coordinates in two directions, with a specific length Divide the area into intervals, The formula for calculating the number of grids in a direction is as follows: (5) In the formula, maxX , maxY , x , Y are the maximum and minimum values ​​in the X and Y directions, is the grid side length, for The number of grids in the direction. , each grid takes a stable area connection point. In specific implementation, you can set For other values.

[0043] Step 4.2, the second phase image uses these control points for aerial triangulation. This process is similar to using ground control points, but the control points here already contain the image coordinate information, eliminating the step of manually placing control points. At this time, the three-dimensional coordinates of the control points are based on the exterior orientation elements of the first phase image, and the bundle method is adjusted to determine the exterior orientation elements of the second phase image, achieving relative registration between the two phases of images.

[0044] Step 5: Use the first and second phase aerial triangulation results to generate respective 3D models and digital surface models, and then perform differential processing on these models for deformation monitoring.

[0045] The implementation method of step 5 further preferably adopted in the embodiment includes the following sub-steps: Step 5.1, generate high-density three-dimensional point clouds for the two phases of images with updated exterior orientation elements. Each laser point in the point cloud contains its three-dimensional coordinates and color information. Interpolate and sample the dense point cloud model to obtain the two-phase digital surface model (DSM) at the specified resolution.

[0046] In step 5.2, a multi-scale model-to-model cloud comparison (M3C2) algorithm is preferably used to calculate the distance between the two phase point clouds.

[0047] For the sake of reference, the implementation process of the M3C2 algorithm is described as follows: 1. Select the core point cloud: Select a set of core points from the first phase of the point cloud. These points are usually evenly distributed and can represent the entire scene. 2. Calculate the three-dimensional surface normal: For each core point, fit a local plane in its neighborhood and calculate the normal vector of the plane. The surface normal is used to determine the direction of measuring distance in the point cloud. 3 Determine the point cloud distance: Calculate the distance between corresponding points in the two phases of the point cloud along the normal direction of each core point. 4. Establish spatial variable confidence intervals: Evaluate the impact of point cloud density and local surface roughness on distance measurement, and calculate the confidence interval of each distance measurement through statistical methods to determine its credibility. The M3C2 algorithm can directly detect ground deformation without meshing, and is particularly suitable for photogrammetric point cloud data. The differences generated by the M3C2 analysis reflect the three-dimensional deformation of the area between the two surveys.

[0048] Step 5.3, perform differential processing on the two phases of DSM models to generate a DSM differential map. Since the DSM is derived from the registered images, they share a unified geographic coordinate reference (i.e., the exterior orientation elements of the first phase of the image are used as the reference). The DSM differential map intuitively shows the settlement changes that occurred in the study area between the two surveys. In order to facilitate understanding of the technical effect of the present invention, as a comparison, the accuracy of this method is verified using control points with known spatial coordinates.

[0049] Figure 6 The elevation accuracy of this method and the traditional method is demonstrated. When ground control points are not used, the elevation accuracy is 0.9m, and the elevation accuracy of the method proposed by the present invention is 0.06m, which is close to the accuracy of the method using a large number of ground control points (more than 21). Compared with the traditional method without control points, this method greatly improves the relative accuracy of the two-phase UAV image registration; it is equivalent to the method using a large number of control points. The present invention greatly reduces the workload of field work while ensuring accuracy. Figure 7 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 7As shown, the electronic device may include: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus. The processor may call the logic instructions in the memory to execute a high-precision geological deformation monitoring method based on drone images. In addition, the logic instructions in the above-mentioned memory may be implemented in the form of a software functional unit and may be stored in a computer-readable storage medium when sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a disk, or an optical disk. On the other hand, the present invention also provides a computer program product, the computer program product includes a computer program, the computer program can be stored on a non-transitory computer-readable storage medium, and when the computer program is executed by a processor, the computer can execute the high-precision geological deformation monitoring method based on drone images provided by the above methods. On the other hand, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it is implemented to execute the high-precision geological deformation monitoring method based on drone images provided by the above methods. The device embodiments described above are only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the present embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor. Through the description of the above implementation methods, technicians in this field can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware.Based on this understanding, the above technical solutions are essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., and includes a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments. Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention is described in detail with reference to the above embodiments, ordinary technicians in this field should understand that they can still modify the technical solutions recorded in the above embodiments, or replace some of the technical features therein by equivalent; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A high-precision geological deformation monitoring method based on drone images, characterized by: Including the following process, The two phases of images from the drone are combined to perform bundle block adjustment to obtain common tie points and corresponding object coordinates. The two phases of images are split and processed separately, the first phase of images are re-adjusted, and the connection points with the smallest change from the merged processing are extracted as stable area connection points, and used as control information for the second phase of images for bundle adjustment to make them aligned with the first phase of images; The three-dimensional model and digital surface model (DSM) were established using the registered two-phase images, and differential analysis was performed to obtain geological deformation information.

2. A high-precision geological deformation monitoring method based on drone images as claimed in claim 1, characterized in that: When two phases of images are merged for bundle block adjustment to obtain common connection points, the implementation method is to merge the two groups of images from different periods, perform joint aerial triangulation without ground control points, obtain the exterior orientation elements and image connection point information of the two phases of images, split the jointly processed images into two independent projects according to the shooting time, retain their respective connection point information, and extract the connection points common to the first and second phases.

3. A high-precision geological deformation monitoring method based on drone images as claimed in claim 2, characterized in that: The extraction of the connection points common to the first phase and the second phase is implemented as follows: Make two copies of the merged aerial triangulation project. Delete the second-phase images and connection points from the first copy, and delete the first-phase images and connection points from the second copy. The connection points in the two projects are screened. If a connection point is included in both the first and second phase images, it is retained. If the connection point only exists in a single phase image, it is deleted. Finally, the common connection point information of the two phase images is obtained.

4. The high-precision geological deformation monitoring method based on drone images as claimed in claim 1, characterized in that: The implementation method of extracting stable region connection points is as follows: Based on the first phase of images, independent aerial triangulation is performed without ground control points to update the image's exterior orientation parameters and tie point information; The new connection point Connecting points from the first of two phases of aerial triangulation Compare and extract the connection points with the same name; The three-dimensional coordinates of the corresponding connection points in the second-phase image are replaced with the three-dimensional coordinates in the first-phase image to obtain the stable area connection points with the first-phase image as the reference benchmark.

5. The high-precision geological deformation monitoring method based on drone images as claimed in claim 4, characterized in that: The implementation of taking the connection point with the same name is as follows: The new connection point Connecting points from the first of two phases of aerial triangulation Compare and consider the connection points whose coordinate differences between images are lower than the corresponding threshold a as the connection points with the same name, and each connection point has projections in at least b images of the first and second phases; The selected connection points are marked accordingly. , and rank them according to the Euclidean distance, retaining the top c% of connection points with the smallest distance and discarding the rest, where a, b, and c are preset thresholds.

6. The high-precision geological deformation monitoring method based on drone images according to claim 1, characterized in that: The identified stable area connection points are further thinned and used as control points to perform aerial triangulation on the second phase of images and update the exterior orientation elements of the images.

7. A high-precision geological deformation monitoring method based on drone images as described in claim 1 or 2 or 3 or 4 or 5 or 6, characterized in that: Used to achieve high-precision surface deformation monitoring without ground control points.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the high-precision geological deformation monitoring method based on drone images as described in any one of claims 1 to 7 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the high-precision geological deformation monitoring method based on drone images as described in any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that: When the computer program is executed by a processor, the high-precision geological deformation monitoring method based on drone images as described in any one of claims 1 to 7 is implemented.

Images