A glacier monitoring method and system based on binocular 3D reconstruction

Abstract

The present invention discloses a glacier monitoring method based on binocular 3D reconstruction. The method comprises: setting up an observation platform in the target glacier monitoring area, with two cameras on the observation platform, a GNSS device deployed at each camera position, and multiple measuring rods deployed within the camera's field of view. The two cameras photograph the measuring rods and the glacier to obtain a pair of glacier images; processing the glacier images; using a binocular stereo matching algorithm to find feature points in the corrected glacier images and construct a glacier digital elevation model; using an automatic matching algorithm to match the feature points on the measuring rods in the two glacier digital elevation models to obtain the glacier surface velocity; and using the GNSS device to measure the positional changes of the feature points on the glacier at different times to obtain the glacier surface velocity, which is used to verify the glacier ice flow velocity. Based on binocular 3D reconstruction, the present invention monitors glacier changes by building a 3D model of the survey area, effectively solving the problems of traditional measuring rod monitoring methods, such as small data volumes and limited coverage.

Description

Technical Field

[0001] The present invention relates to the field of photogrammetry, in particular to the field of glacier monitoring, and specifically to a glacier monitoring method based on binocular visual three-dimensional reconstruction. Background Art

[0002] Glacier flow velocity and its mass balance are important indicators of glacier change. They can be estimated through field observations, model simulations, satellite remote sensing, and drone imagery. Rod-based methods can obtain highly accurate measurements. However, due to glacier movement, the rods will also move. Measuring the rod's position at intervals provides a rough estimate of ice flow velocity at that point, but it is not possible to obtain a comprehensive picture of ice flow velocity. This is inefficient and labor-intensive. Simulations using the open-source Global Glacier Model (OGGM) can provide some predictions of future glacier change, but they require extensive data and are subject to uncertainty. Estimation using satellite remote sensing or drone imagery can provide detailed data over a larger survey area, but image registration and the computation of massive amounts of data remain major challenges.

[0003] With the development of photogrammetry and computer vision technology, binocular 3D reconstruction methods based on the principles of human vision have become very mature. By using two cameras to shoot the observed object from different angles and calculating the parallax of the observed object, the depth information of the observed object can be estimated, thereby restoring the coordinate information of the observed object in three-dimensional space to establish a three-dimensional model of the observed object. This technology has been widely used in target recognition and 3D reconstruction tasks in special scenes such as underwater, providing a possibility for target monitoring and 3D modeling in glacial scenes.

[0004] Therefore, this field urgently needs to propose new technical solutions to obtain DEM and ice flow velocity within a certain range to improve the existing single-point observation scheme of glaciers in the field, which has insufficient monitoring results accuracy. Summary of the Invention

[0005] To overcome the above-mentioned deficiencies of the prior art, the present invention provides a glacier monitoring method and system based on binocular three-dimensional reconstruction of vision, which constructs a glacier DEM by performing three-dimensional reconstruction of the measuring rod and its nearby glaciers to detect glacier changes.

[0006] According to one aspect of the present invention, a glacier monitoring method based on binocular 3D reconstruction is provided, comprising:

[0007] S1. Set up an observation platform in the target glacier monitoring area. Two cameras are spaced apart on the platform, and a GNSS device is deployed at each camera position to monitor the camera position. Multiple measuring rods are deployed within the camera's field of view, with markers placed on each rod at equal intervals. Two cameras periodically capture the measuring rods and the nearby glaciers to obtain a pair of glacier images.

[0008] S2, performing noise reduction processing, image distortion correction, stereo correction, and effective area extraction on the acquired glacier image in sequence to obtain a corrected glacier image;

[0009] S3. Based on the corrected glacier image, a binocular stereo matching algorithm is used to find feature points representing the same three-dimensional spatial point in a pair of glacier images. Disparity values ​​are calculated based on the relative displacement of the feature points to obtain a disparity map. The three-dimensional coordinates of the pixels are restored based on the disparity map and the camera projection matrix. A three-dimensional model of the measuring rod and its surroundings is obtained, and a digital elevation model of the glacier is constructed.

[0010] S4. Using an automatic matching algorithm, match the characteristic points on the measuring rods in the glacier digital elevation models of two adjacent periods to obtain the glacier surface velocity;

[0011] S5. By using the camera position monitored by the GNSS device as the measurement point and combining it with the image taken by the camera, a group of feature points on the glacier are selected in the glacier image, and the GNSS device is used to measure the position changes of the feature points on the glacier at different times to obtain the surface flow velocity of the glacier, which is used to verify the surface flow velocity of the glacier.

[0012] As a further technical solution, the method further includes: using the camera position monitored by the GNSS device as a reference point, combined with the constructed glacier digital elevation model, to obtain the absolute speed of change of the surrounding glaciers.

[0013] As a further technical solution, the method further includes: obtaining the glacier surface elevation difference at different periods by comparing the glacier digital elevation models of the monitoring area at different periods, and obtaining the change in the material balance of the glacier.

[0014] As a further technical solution, the method also includes: providing absolute glacier surface elevation change measurement results based on the scale of the ice surface corresponding to the measuring rod captured by the camera, which is used to verify the change in the glacier material balance.

[0015] As a further technical solution, the method further includes: using two identical cameras with known interior orientation parameters, fixing them at the observation site, and calibrating them on site to obtain the exterior orientation parameters.

[0016] As a further technical solution, the noise reduction process includes: using a non-local mean algorithm to perform smooth correction on the image through weighted calculation of pixel similarity.

[0017] As a further technical solution, the image distortion correction includes: based on the internal and external orientation parameters and distortion coefficients of the camera, using radial and tangential distortion formulas to correct errors in the image.

[0018] As a further technical solution, the stereo correction and effective area extraction include: calculating the correction and projection matrices of the left and right cameras based on the internal and external orientation parameters and distortion parameters of the two cameras, using the relative rotation and translation matrices, remapping the pixels of the original image to the corrected image to make the viewing angles of the left and right cameras parallel, and obtaining the matrix ROI of the effective area at the same time, and finally cropping to obtain the effective area.

[0019] As a further technical solution, the three-dimensional coordinates of the pixels are restored according to the disparity map and the camera projection matrix, including:

[0020] The pixel coordinates and disparity values ​​are input as a homogeneous coordinate, and the homogeneous coordinate conversion is performed using the disparity map and the camera's projection matrix;

[0021] The disparity map and the camera's projection matrix The calculation formula is as follows:

[0022] ;

[0023] Where, ( ) is the principal point coordinate of the camera, is the horizontal focal length of the camera, is the baseline distance between the two cameras, is the disparity value, (x, y) is the pixel coordinate, and (X, Y, Z) is the actual 3D point coordinate;

[0024] By dividing by W, the normalized coordinates of the actual 3D point are obtained. W represents the distance from the 3D point to the optical center of the camera, that is, the depth.

[0025] According to one aspect of the present invention, a glacier monitoring system based on binocular 3D reconstruction is provided, comprising a monitoring device and a data processing device, wherein:

[0026] The monitoring device includes an observation platform, on which two cameras are spaced apart, and a GNSS device is deployed at each camera position for monitoring the camera position. A plurality of measuring rods are deployed within the field of view of the cameras, and markers are evenly spaced on each measuring rod. The monitoring device is used to periodically photograph the measuring rods and the glaciers nearby using the two cameras to obtain a pair of glacier images;

[0027] The data processing device is used to perform noise reduction processing, image distortion correction, stereo correction and effective area extraction on the acquired glacier image in sequence to obtain a corrected glacier image; based on the corrected glacier image, a binocular stereo matching algorithm is used to find feature points representing the same three-dimensional space point in a pair of glacier images, and the disparity value is calculated according to the relative displacement of the above feature points to obtain a disparity map, and the three-dimensional coordinates of the pixels are restored according to the disparity map and the projection matrix of the camera, and a three-dimensional model of the measuring rod and its surrounding environment is calculated to construct a glacier digital elevation model; through an automatic matching algorithm, the feature points on the measuring rod in the glacier digital elevation models of two periods are matched to calculate the glacier surface flow velocity, and a group of feature points on the glacier are selected in the glacier image by using the camera position monitored by the GNSS device as the measurement point and combining it with the image taken by the camera, so that the GNSS device measures the position changes of the feature points on the glacier in different periods to obtain the glacier surface flow velocity, which is used to verify the glacier surface flow velocity.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] 1. This invention is based on binocular 3D reconstruction. By building a 3D model of the survey area to monitor glacier changes, it effectively solves the problems of small data volume and small coverage of traditional measuring rod monitoring methods.

[0030] 2. The solution provided by the present invention can realize automatic monitoring under the condition that the absolute position of the camera remains unchanged, without the need for manual re-measurement of the position information of the measuring rod to estimate the ice flow rate.

[0031] 3. The solution provided by the present invention can realize dynamic monitoring of the survey area. By comparing the three-dimensional models of different periods, the change information of glaciers in the survey area can be obtained more intuitively.

[0032] 4. The present invention uses a GNSS device to monitor whether the absolute position of the camera has changed to ensure the accuracy of the model analysis. By using the camera position monitored by the GNSS device as the measurement point and combining it with the image captured by the camera, a group of characteristic points on the glacier are selected from the glacier image. The GNSS device measures the position changes of the characteristic points on the glacier at different times to obtain the glacier surface flow velocity. This can verify the results of the glacier ice flow velocity and improve the accuracy of the monitoring results.

[0033] 5. This invention also uses the camera position monitored by the GNSS device as a reference point, combined with the constructed glacier digital elevation model, to obtain the absolute rate of change of the surrounding glaciers, thereby better understanding the mechanism of glacier change and providing a scientific basis for responding to climate change.

[0034] 6. Based on the camera capturing the scale of the ice surface on the measuring rod, the present invention provides absolute measurement results of glacier surface elevation changes, which is used to verify the changes in glacier mass balance and improve the accuracy of monitoring results. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, a brief introduction will be given below to the drawings used in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0036] Figure 1 A flowchart of a glacier monitoring method and system based on binocular 3D reconstruction is provided in an embodiment of the present invention.

[0037] Figure 2 A schematic diagram of the observation platform structure of a glacier monitoring method and system based on binocular visual 3D reconstruction provided in an embodiment of the present invention.

[0038] Figure 3 A schematic diagram of a stereo image pair of a glacier monitoring method and system based on binocular visual 3D reconstruction provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0039] The terms "including" and "having" and any variations thereof in the description and claims of the present invention and the above-mentioned drawings are intended to cover non-exclusive inclusions, for example, a process, method, system, product or apparatus that includes a series of steps or units is not necessarily limited to the steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products or apparatuses.

[0040] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In addition, the technical features in the various embodiments or single embodiments provided by the present invention are arbitrarily combined with each other to form a new technical solution. This combination is not restricted by the sequence of steps and / or structural composition mode, but must be based on the ability of ordinary technicians in this field to implement it. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that this combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0041] See also Figure 1-3 The embodiment of the present invention provides a glacier monitoring method based on binocular 3D reconstruction, comprising:

[0042] Step 1: Set up an observation platform in the glacier monitoring target area. Two cameras are installed on the observation platform at intervals to form a stereoscopic observation of the glacier. A GNSS device is deployed at each camera position to monitor the camera position. Multiple measuring rods are deployed within the camera's field of view. The number of measuring rods is no less than three, and each measuring rod is marked with equal intervals. The measuring rods are used to provide reference points for the glacier position. The two cameras regularly photograph the measuring rods and the glaciers nearby to obtain glacier images. Step 2: The obtained glacier images are subjected to noise reduction, image distortion correction, stereo correction, and effective area extraction in sequence to obtain the corrected glacier images.

[0043] Step 3: Based on the corrected glacier image, a binocular stereo matching algorithm (SGBM algorithm) is used to find feature points representing the same 3D spatial point in a pair of glacier images. The disparity values ​​are calculated based on the relative displacement of these feature points to obtain a disparity map. The 3D coordinates of the pixels are restored based on the disparity map and the camera projection matrix. A 3D model of the measuring rod and its surroundings is obtained to construct a glacier digital elevation model (DEM).

[0044] Under the premise of assuming that the absolute position of the camera remains unchanged, a grid is created based on the 3D point cloud data and a bilinear interpolation calculation is performed to obtain the DEM.

[0045] Step 4: Use an automatic matching algorithm, such as the SIFT algorithm, to match the characteristic points on the measuring rods in the glacier digital elevation models of the two periods to obtain the glacier surface velocity;

[0046] Step 5: By using the camera position monitored by the GNSS device as the measurement point and combining it with the image captured by the camera, a group of characteristic points on the glacier are selected from the glacier image. The GNSS device is then used to measure the position changes of the characteristic points on the glacier at different times to obtain the glacier surface velocity, which is used to verify the glacier surface velocity.

[0047] In addition, the embodiment of the present invention uses the camera position monitored by the GNSS device as a reference point, combined with the constructed glacier digital elevation model, to obtain the absolute speed of change of the surrounding glaciers, thereby better understanding the change mechanism of glaciers and providing a scientific basis for responding to climate change.

[0048] It is also possible to obtain the surface elevation difference of the glacier at different times and the change in the material balance of the glacier by comparing the digital elevation models of the glaciers in the monitoring areas at different times.

[0049] Specifically, based on the glacier DEM obtained at different times, the height difference between the time periods is obtained, the material balance changes of the glaciers in the survey area are estimated, and the material balance changes of the glaciers are obtained. The estimation of the glacier material balance changes is based on the DEMs obtained in different periods. By comparing and analyzing the data differences between the two DEMs, the material balance changes of the glaciers are obtained.

[0050] Since three measuring rods are arranged within the camera’s field of view as constraints and feature points, the scales on the measuring rods corresponding to the ice surface can be captured by the camera, providing absolute glacier surface elevation change measurements, which are used to verify the obtained glacier mass balance changes.

[0051] In step 1, a 1-meter-long horizontal arm is placed on the observation platform, and two cameras are fixed at both ends of the horizontal arm. By setting up the horizontal arm, an installation position is provided for the cameras.

[0052] The method further includes: using two identical cameras with known interior orientation parameters, completing fixation and on-site calibration at the observation site to obtain exterior orientation parameters.

[0053] The specific method for obtaining exterior orientation parameters is as follows:

[0054] Use a camera to take photos of the black and white chessboard on the spot;

[0055] Use the cv2.findChessboardCorners() function provided by the OPENCV library to detect chessboard corners and create the three-dimensional and two-dimensional coordinates of the corners;

[0056] The exterior orientation parameters of the camera are calculated based on the principle of rear intersection.

[0057] In addition, since the surface of the glacier is smooth and highly reflective, a polarizing filter must be installed on the camera to ensure that the pictures taken are clear and usable.

[0058] In step 2, the acquired glacier images are processed, including noise reduction, image distortion correction, stereo correction, and effective area extraction, to remove environmental noise and image distortion and ensure accurate image matching.

[0059] The image noise reduction process includes: using the non-local means algorithm (NLM) to perform smooth correction on the image through weighted calculation of pixel similarity.

[0060] Specifically, for noise reduction processing, the window size and smoothing parameter h are mainly selected, and the non-local means algorithm (NLM) is applied to update the value of the current pixel by taking a weighted average of the similarity weights of all pixels within the search range of the current pixel.

[0061] Image distortion correction includes: based on the camera's internal and external orientation parameters and distortion coefficients, using radial and tangential distortion formulas to correct errors in the image.

[0062] Specifically, for image distortion correction, the radial distortion parameters provided during the internal orientation parameter calibration process are used. and tangential distortion parameters Combined with calculation using theoretical distortion correction formula.

[0063] Stereo correction and effective area extraction include: according to the internal and external orientation parameters and distortion parameters of the two cameras, using the relative rotation and translation matrix to calculate the correction and projection matrices of the left and right cameras, remapping the pixels of the original image to the corrected image to make the perspectives of the left and right cameras parallel, and obtaining the matrix ROI of the effective area at the same time, and finally cropping to obtain the effective area.

[0064] The NLM algorithm uses each pixel The small window centered on is a pixel block, and similar pixel blocks are searched within the search range. The similarity between two pixel blocks is measured by the weighted Euclidean distance of their gray values;

[0065] (1);

[0066] in and In pixels and Two pixel blocks centered on is a parameter that controls the degree of smoothing, Is the similarity weight between the two pixel blocks. Then for the current pixel, use the similarity weights of all pixels in the search range to perform weighted average to update the value of the current pixel.

[0067] Combining the radial and tangential distortion parameters, each pixel is corrected using the following formula.

[0068] (2);

[0069] in is the distance from the point to the optical axis, , To describe the low-order distortion effects, high-order distortion effects (especially at the edge of the lens), and higher-order distortion effects, respectively, based on the radial distortion parameters provided during the internal orientation parameter calibration process. In order to control the components of the tangential distortion in the two main directions according to the tangential distortion parameters provided during the internal orientation parameter calibration process, the tangential distortion is described as the distortion caused by the imperfect alignment of the lens. 、 ) is the original pixel coordinate, ( 、 ) are the pixel coordinates after radial and tangential distortion.

[0070] The two cameras acquire images separately, and with the left camera as the reference, the rotation matrix of the stereo correction of the left and right cameras is calculated. , and the projection matrix , , the specific formula is as follows:

[0071] (3);

[0072] Where R is a common rotation matrix used to correct the perspectives of the two cameras so that their viewing planes are parallel and their optical axes are aligned. In the new coordinate system defined by it, the x-axis is aligned with the baseline, the y-axis is obtained by the cross product of the original optical axis direction and the baseline, and the z-axis is obtained by the cross product of the x-axis and the y-axis. is the relative rotation matrix between cameras, , R C1 With R C2 Represents the rotation matrix of the left and right cameras respectively; is the translation matrix of the left camera relative to the right camera, and They represent the translation matrices of the left and right cameras respectively, K1 and K2 are the intrinsic parameter matrices of the left camera and the right camera respectively.

[0073] Finally, the original image is remapped to the corrected image. The specific formula is as follows:

[0074] (4);

[0075] in,( ) is the pixel coordinate after distortion correction, ( 、 ) are the pixel coordinates of the original image.

[0076] After stereo rectification, some pixels are mapped to the edge of the image or exceed the original area of ​​the image, resulting in black edges. By detecting the area of ​​valid pixels in the image, a minimum rectangular bounding box is constructed to contain these valid areas.

[0077] The disparity map is generated by a stereo matching algorithm (such as the SGBM algorithm), and the three-dimensional point cloud data of the observation area is obtained based on the disparity information and the camera projection matrix.

[0078] In step 3, the principle diagram of using parallax information to restore the three-dimensional coordinates of the point is as follows: Figure 3 After stereo correction, the viewing angles of the two images are parallel, and the two camera planes can be regarded as two straight lines. arrive and The distance is the depth , and The distance is , the focal length is .

[0079] Based on the left camera projection center Draw parallel lines about the light in the right image. Based on the principle of similar triangles, we can get parallax and depth. Inversely proportional:

[0080] (5);

[0081] In the image space auxiliary coordinate system of the left camera, That is the point of Coordinates, based on the projection relationship between parallax and depth, are obtained by converting pixel coordinates of and Coordinates. Assuming that the camera's exterior orientation parameters remain unchanged, the modeling system no longer needs to convert the coordinate system into the world coordinate system.

[0082] In actual operation, after the same feature points are extracted by SGBM algorithm, the formula The disparity of each feature point is obtained to form a disparity map. Then the pixel coordinates and disparity values ​​are input as a homogeneous coordinate, and the disparity map and the camera's projection matrix are used to calculate the pixel coordinates. Perform homogeneous coordinate transformation, and finally, divide by , normalized to get the actual 3D point coordinates. Disparity map and camera projection matrix The calculation formula is as follows:

[0083] (6);

[0084] in,( ) is the principal point coordinate of the camera, is the horizontal focal length of the camera, is the baseline distance between the two cameras, (x, y) is the pixel coordinate, (X, Y, Z) is the actual 3D point coordinate, and W represents the distance from the 3D point to the optical center of the camera, that is, the depth.

[0085] The color information of the 3D point cloud data is derived from the average color information of the feature points in the two images. In addition, the point cloud can be filtered by setting a parallax threshold to filter out distant point clouds.

[0086] The implementation of each embodiment of the present invention is based on programmed processing by a device with processor functionality. Therefore, in practical engineering, the technical solutions and functionalities of each embodiment of the present invention are encapsulated into various modules. Based on this reality, and in addition to the aforementioned embodiments, an embodiment of the present invention provides a glacier monitoring system based on binocular 3D reconstruction. This system is used to implement the glacier monitoring method based on binocular 3D reconstruction described in the aforementioned method embodiments.

[0087] A glacier monitoring system based on binocular 3D reconstruction includes a monitoring device and a data processing device, wherein:

[0088] The monitoring device includes an observation platform with two cameras spaced apart. A GNSS device is deployed at each camera position to monitor the camera's position. Multiple measuring rods are deployed within the camera's field of view, each with markers spaced evenly apart. The monitoring device uses the two cameras to periodically photograph the measuring rods and the glaciers nearby, acquiring a pair of glacier images.

[0089] The data processing device is used to perform noise reduction processing, image distortion correction, stereo correction and effective area extraction on the acquired glacier image in sequence to obtain a corrected glacier image; based on the corrected glacier image, a binocular stereo matching algorithm is used to find feature points representing the same three-dimensional space point in a pair of glacier images, and the disparity value is calculated according to the relative displacement of the above feature points to obtain a disparity map, and the three-dimensional coordinates of the pixels are restored according to the disparity map and the projection matrix of the camera, and the three-dimensional model of the measuring rod and its surrounding environment is calculated to construct a glacier digital elevation model; through the automatic matching algorithm, the feature points on the measuring rod in the glacier digital elevation models of the two periods are matched, and the glacier surface flow velocity is calculated. By using the camera position monitored by the GNSS device as the measurement point and combining the image taken by the camera, a group of feature points on the glacier are selected in the glacier image, and the GNSS device measures the position changes of the feature points on the glacier in different periods to obtain the glacier surface flow velocity, which is used to verify the glacier surface flow velocity. An embodiment of the present invention provides a glacier monitoring system based on binocular visual three-dimensional reconstruction. By establishing a three-dimensional model of the survey area to monitor glacier changes, it effectively solves the problems of small data volume and small coverage of traditional measuring rod monitoring methods. Moreover, under the assumption that the absolute position of the camera remains unchanged, it can realize automatic monitoring without the need to manually re-measure the position information of the measuring rod to estimate the ice flow rate. At the same time, the solution provided by the present invention can realize dynamic monitoring of the survey area. By comparing the three-dimensional models of different periods, it can more intuitively obtain information on glacier changes in the survey area.

[0090] This invention applies binocular three-dimensional reconstruction technology to the field of glacier monitoring for the first time. The use of this invention helps to conduct fully automatic on-site glacier observations, obtain glacier monitoring data with wider coverage and more intuitiveness, and improve glacier monitoring efficiency.

[0091] It should be noted that the system embodiments provided by the present invention are not only used to implement the methods in the above-mentioned method embodiments, but also used to implement the methods in other method embodiments provided by the present invention. The only difference is the setting of corresponding functional modules. The principles thereof are basically the same as those of the above-mentioned system embodiments provided by the present invention. As long as those skilled in the art refer to the specific technical solutions in other method embodiments on the basis of the above-mentioned system embodiments, obtain corresponding technical means and technical solutions composed of these technical means by combining technical features, and on the premise of ensuring the practicality of the technical solutions, improve the modules in the above-mentioned system embodiments to obtain corresponding system-type embodiments, which are used to implement the methods in other method-type embodiments. For example:

[0092] Based on the contents of the above system embodiments, as a preferred embodiment, a glacier monitoring system based on visual binocular three-dimensional reconstruction is provided in the embodiments of the present invention, and the method also includes: using the camera position monitored by the GNSS device as a reference point, combined with the constructed glacier digital elevation model, to obtain the absolute speed of change of the surrounding glaciers.

[0093] Based on the contents of the above system embodiment, as a preferred embodiment, the method further includes: obtaining the glacier surface elevation difference in different periods by comparing the glacier digital elevation models of the monitoring area in different periods, and obtaining the change in the material balance of the glacier.

[0094] Based on the contents of the above system embodiments, as a preferred embodiment, a glacier monitoring system based on visual binocular three-dimensional reconstruction is provided in the embodiments of the present invention, and the method also includes: providing absolute glacier surface elevation change measurement results based on the scale of the ice surface corresponding to the measuring rod captured by the camera, which is used to verify the change in glacier material balance.

[0095] Based on the contents of the above system embodiments, as a preferred embodiment, a glacier monitoring system based on visual binocular three-dimensional reconstruction is provided in the embodiments of the present invention, and the method also includes: using two identical cameras with known internal orientation parameters, completing fixation and on-site calibration at the observation site to obtain external orientation parameters.

[0096] Based on the contents of the above system embodiments, as a preferred embodiment, the embodiments of the present invention provide a glacier monitoring system based on binocular visual three-dimensional reconstruction, noise reduction processing, and the use of non-local means algorithm (NLM) to smooth and correct the image through weighted calculation of pixel similarity.

[0097] Based on the contents of the above system embodiments, as a preferred embodiment, the embodiments of the present invention provide a glacier monitoring system based on visual binocular three-dimensional reconstruction, image distortion correction, based on the internal and external orientation parameters and distortion coefficients of the camera, using radial and tangential distortion formulas to correct errors in the image.

[0098] Based on the contents of the above system embodiments, as a preferred embodiment, the embodiments of the present invention provide a glacier monitoring system based on visual binocular three-dimensional reconstruction, stereo correction and effective area extraction. According to the internal and external orientation parameters and distortion parameters of the two cameras, the relative rotation and translation matrices are used to calculate the correction and projection matrices of the left and right cameras, and the pixels of the original image are remapped to the corrected image to make the viewing angles of the left and right cameras parallel. At the same time, the matrix ROI of the effective area is obtained, and finally the effective area is cropped.

[0099] Based on the contents of the above system embodiments, as a preferred embodiment, a glacier monitoring system based on binocular 3D reconstruction is provided in the embodiments of the present invention, which restores the 3D coordinates of pixels according to the disparity map and the camera projection matrix, including:

[0100] The pixel coordinates and disparity values ​​are input as a homogeneous coordinate, and the homogeneous coordinate conversion is performed using the disparity map and the camera's projection matrix;

[0101] Disparity map and camera projection matrix The calculation formula is as follows:

[0102] ;

[0103] Where, ( ) is the principal point coordinate of the camera, is the horizontal focal length of the camera, is the baseline distance between the two cameras, is the disparity value, (x, y) is the pixel coordinate, and (X, Y, Z) is the actual 3D point coordinate;

[0104] By dividing by W, the normalized coordinates of the actual 3D point are obtained. W represents the distance from the 3D point to the optical center of the camera, that is, the depth.

[0105] In summary, the present invention establishes a three-dimensional model of the survey area to monitor glacier changes, effectively solving the problems of small data volume and small coverage of traditional measuring rod monitoring methods; and under the assumption that the absolute position of the camera remains unchanged, it can realize automatic monitoring without the need to manually re-measure the position information of the measuring rod to estimate the ice flow rate. At the same time, the solution provided by the present invention can realize dynamic monitoring of the survey area. By comparing the three-dimensional models of different periods, the glacier change information in the survey area can be obtained more intuitively.

[0106] 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 it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.

Claims

1. A glacier monitoring method based on binocular 3D reconstruction, characterized in that: include: S1. Set up an observation platform in the target glacier monitoring area. Two cameras are spaced apart on the platform, and a GNSS device is deployed at each camera position to monitor the camera position. Multiple measuring rods are deployed within the camera's field of view, with markers placed on each rod at equal intervals. Two cameras periodically capture the measuring rods and the nearby glaciers to obtain a pair of glacier images. S2, performing noise reduction processing, image distortion correction, stereo correction, and effective area extraction on the acquired glacier image in sequence to obtain a corrected glacier image; S3. Based on the corrected glacier image, a binocular stereo matching algorithm is used to find feature points representing the same three-dimensional spatial point in a pair of glacier images. Disparity values ​​are calculated based on the relative displacement of the feature points to obtain a disparity map. The three-dimensional coordinates of the pixels are restored based on the disparity map and the camera projection matrix. A three-dimensional model of the measuring rod and its surroundings is obtained, and a digital elevation model of the glacier is constructed. S4. Using an automatic matching algorithm, match the characteristic points on the measuring rods in the glacier digital elevation models of two adjacent periods to obtain the glacier surface velocity; S5. By using the camera position monitored by the GNSS device as the measurement point and combining it with the image taken by the camera, a group of feature points on the glacier are selected in the glacier image, and the GNSS device is used to measure the position changes of the feature points on the glacier at different times to obtain the surface flow velocity of the glacier, which is used to verify the surface flow velocity of the glacier.

2. The glacier monitoring method based on binocular 3D reconstruction according to claim 1, characterized in that: The method further includes: obtaining the absolute speed of change of the surrounding glaciers by using the camera position monitored by the GNSS device as a reference point in combination with the constructed glacier digital elevation model.

3. The glacier monitoring method based on binocular 3D reconstruction according to claim 1, characterized in that: The method further includes: obtaining the glacier surface elevation difference at different periods by comparing the glacier digital elevation models of the monitoring area at different periods, and obtaining the material balance change of the glacier.

4. The glacier monitoring method based on binocular 3D reconstruction according to claim 3, characterized in that: The method further includes providing an absolute glacier surface elevation change measurement result based on the scale of the ice surface corresponding to the measuring rod captured by the camera, so as to verify the change in the glacier mass balance.

5. The glacier monitoring method based on binocular 3D reconstruction according to claim 1, characterized in that: The method further includes: using two cameras with the same known interior orientation parameters, fixing them at the observation site, and calibrating them on site to obtain the exterior orientation parameters.

6. The glacier monitoring method based on binocular 3D reconstruction according to claim 1, characterized in that: The noise reduction process includes: using a non-local mean algorithm to perform smooth correction on the image through weighted calculation of pixel similarity.

7. The glacier monitoring method based on binocular 3D reconstruction according to claim 6, characterized in that: The image distortion correction includes: based on the internal and external orientation parameters and distortion coefficients of the camera, using radial and tangential distortion formulas to correct errors in the image.

8. The glacier monitoring method based on binocular 3D reconstruction according to claim 7, characterized in that: The stereo correction and effective area extraction include: calculating the correction and projection matrices of the left and right cameras based on the internal and external orientation parameters and distortion parameters of the two cameras, using relative rotation and translation matrices, remapping the pixels of the original image to the corrected image to make the viewing angles of the left and right cameras parallel, obtaining the matrix ROI of the effective area at the same time, and finally cropping to obtain the effective area.

9. The glacier monitoring method based on binocular 3D reconstruction according to claim 1, characterized in that: Recover the 3D coordinates of the pixel based on the disparity map and the camera's projection matrix, including: The pixel coordinates and disparity values ​​are input as a homogeneous coordinate, and the homogeneous coordinate conversion is performed using the disparity map and the camera's projection matrix; The disparity map and the camera's projection matrix The calculation formula is as follows: ； Where, ( ) is the principal point coordinate of the camera, is the horizontal focal length of the camera, is the baseline distance between the two cameras, is the disparity value, (x, y) is the pixel coordinate, and (X, Y, Z) is the actual 3D point coordinate; By dividing by W, the normalized coordinates of the actual 3D point are obtained. W represents the distance from the 3D point to the optical center of the camera, that is, the depth.

10. A glacier monitoring system based on binocular 3D reconstruction, comprising a monitoring device and a data processing device, characterized in that: The monitoring device includes an observation platform, on which two cameras are spaced apart, and a GNSS device is deployed at each camera position to monitor the camera position. A plurality of measuring rods are deployed within the camera field of view, and markers are evenly spaced on each measuring rod. The monitoring device is used to periodically photograph the measuring rods and the glaciers nearby using the two cameras to obtain a pair of glacier images; The data processing device is used to perform noise reduction processing, image distortion correction, stereo correction and effective area extraction on the acquired glacier image in sequence to obtain a corrected glacier image; based on the corrected glacier image, a binocular stereo matching algorithm is used to find feature points representing the same three-dimensional space point in a pair of glacier images, and the disparity value is calculated according to the relative displacement of the above feature points to obtain a disparity map, and the three-dimensional coordinates of the pixels are restored according to the disparity map and the projection matrix of the camera, and a three-dimensional model of the measuring rod and its surrounding environment is calculated to construct a glacier digital elevation model; through an automatic matching algorithm, the feature points on the measuring rod in the glacier digital elevation models of two periods are matched to obtain the glacier surface flow velocity, and a group of feature points on the glacier are selected in the glacier image by using the camera position monitored by the GNSS device as the measurement point and combining it with the image taken by the camera, so that the GNSS device measures the position changes of the feature points on the glacier in different periods to obtain the glacier surface flow velocity, which is used to verify the glacier surface flow velocity.

Images