Wide area network collaborative mode large-scale aerial triangulation method

Abstract

The application discloses a kind of wide area network cooperation mode large-scale image aerial triangulation method.It includes the following steps: first, according to certain rules, the survey area is decomposed into multiple sub-regions with certain overlap, and they are allocated to off-site computing clusters;Then, according to the sub-region, independent aerial triangulation processing is carried out, and the simplified aerial triangulation data is uploaded to the cloud or the central server for merging;Finally, the overall area network adjustment solution is carried out for the merged survey area, and the adjustment results are decomposed to the sub-region.The application has the advantages of maximizing the use of off-site computing clusters for distributed aerial triangulation, reducing the data transmission volume of wide area network, improving the operation efficiency and the overall accuracy consistency of the survey area.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of photogrammetry and remote sensing, and particularly relates to a large-scale image aerial triangulation method based on a wide area network collaborative mode. BACKGROUND

[0002] In the field of aerospace photogrammetry technology, high-precision geometric positioning of optical sensor image data is usually achieved through aerial triangulation. Aerial triangulation refers to obtaining image orientation parameters and three-dimensional coordinates of points to be measured (also known as encrypted points) by using a certain number of ground control points and image observation data and auxiliary data processing according to the geometric relationship between the captured image and the ground object. Aerial triangulation can maintain consistent relative accuracy and absolute positioning accuracy for all image data in the survey area, and is the core link of aerospace image geometric processing.

[0003] In view of the increasing amount of image data, the aerial triangulation operation mode has gradually developed from single-machine independent processing to multi-machine cluster processing, that is, a computing cluster composed of several computing nodes in a local area network is used for parallel processing to improve processing efficiency. However, with the increasing demand for large-scale geographic information product production and updating, such as covering a country or even a continent, the task of large-scale satellite image aerial triangulation is increasing. Such tasks have huge data volume and require rapid completion within a certain time. The existing cluster aerial triangulation operation mode is also difficult to meet the efficiency requirements (the existing cluster aerial triangulation operation mode relies on the manual connection of control points by the operator at the edge of the connection, and also requires the export of RPC to check the control points in the stereoscopic condition, which has a huge workload and is extremely low in efficiency in large-scale satellite image aerial triangulation operation). In order to accelerate the cluster aerial triangulation operation mode, the common method is to divide the complete survey area into multiple sub-areas, then perform aerial triangulation operation in each sub-area, and finally realize the connection of the regional network through offline survey area edge connection. However, the entire process of sub-area division and edge connection is realized through manual assistance, and by means of external software, manual data division is used, which is suitable for small survey areas and has low efficiency, and is difficult to automate. SUMMARY

[0004] The purpose of this invention is to provide a large-scale aerial triangulation method for satellite imagery using a wide area network (WAN) collaborative model. This method integrates the resources of multiple distributed computing clusters across different locations to perform aerial triangulation on large-scale satellite imagery. The method employs a workflow of "survey area decomposition—off-site sub-area processing—sub-area merging and overall adjustment—distribution of adjustment results" for aerial triangulation. It fully utilizes off-site computing clusters to quickly complete aerial triangulation of the task area, automating both partitioning and edge matching. It is applicable to a wide range of survey areas and is highly efficient. This invention addresses the problem that existing cluster-based aerial triangulation methods cannot effectively utilize multi-cluster computing resources across different locations. It also solves the problem that existing technologies rely almost entirely on manual assistance for partitioning and edge matching, requiring external software for data division, resulting in limited applicability to specific survey areas, low efficiency, and difficulty in automation.

[0005] To achieve the above objectives, the technical solution of the present invention is: a wide area network collaborative mode large-scale image aerial triangulation method, characterized by comprising the following steps:

[0006] Step 1, Decompose the survey area;

[0007] The complete survey area is automatically divided into multiple sub-areas based on a certain grid or task sub-area range;

[0008] Step 2, aerial triangulation of the sub-region;

[0009] For each sub-region, after obtaining the original imagery and geometric control data, conduct aerial triangulation independently in accordance with the aerial triangulation operation specifications to ensure that the results of the sub-region meet the requirements;

[0010] Step 3: Upload and summarize simplified aerial triangulation data for sub-regions;

[0011] Upload and aggregate the simplified aerial triangulation data for each sub-region that can be used for adjustment to the central server or cloud.

[0012] Step 4: Merge survey areas;

[0013] The aggregated simplified aerial triangulation data of all sub-regions are merged into a complete survey area and the survey area is automatically edge-connected.

[0014] Step 5, Joint Adjustment;

[0015] The combined complete survey area is subjected to joint adjustment calculation to obtain the precise positioning parameters of all images, and then uploaded to the central server or cloud.

[0016] Step 6: Update the sub-region image positioning parameters;

[0017] Extract the positioning parameters of the sub-region image from all the image precise positioning parameters uploaded in step 5 and save them.

[0018] In the above technical solution, in step 1, the automatic decomposition of the survey area includes the following steps:

[0019] First, the geographic range of each image in the survey area is calculated according to the initial image positioning parameters or directly extracted from the image metadata information;

[0020] Second, the union of all image geographic ranges is calculated to obtain the geographic range of the entire survey area;

[0021] Then, the entire survey area is automatically decomposed into multiple sub-areas according to certain rules, and the automatic division of the survey area includes:

[0022] (1) Automatic division according to a regular grid; the survey area is automatically divided into multiple sub-areas according to a certain grid size;

[0023] (2) Automatic division according to administrative regions; automatic division according to the administrative regions corresponding to the task areas of multiple off-site units participating in the work;

[0024] (3) Automatic division according to clustering; clustering processing is performed on the point set formed by the center points of the stereo image pairs composed of multiple images, and the number of clusters is set according to actual conditions; after clustering, the convex hull of each cluster is calculated;

[0025] (4) Automatic division according to geographic attributes;

[0026] The above automatic division methods of the survey area can be selected according to needs, or combined with several methods.

[0027] In the above technical solution, in order to ensure that the subsequent sub-areas can be merged, adjacent sub-areas should ensure that they overlap by 1-2 image pairs along their boundary lines, or more image pairs along the boundary lines according to actual conditions.

[0028] In the above technical solution, in step 2, each sub-area work unit applies for the original images and control data within the divided sub-area range, and independently performs aerial triangulation work according to relevant work specifications. Each sub-area work unit only needs to apply for all original image data and control data within the divided sub-area range, and does not need to apply for original image data and control data of other sub-areas.

[0029] In the above technical solution, in step 3, after completing the aerial triangulation task in the sub-area, the aerial triangulation results need to be uploaded to the central server or cloud through the wide area network; during transmission, each sub-area only needs to upload simplified aerial triangulation data for adjustment and calculation, which has small data volume and fast transmission; wherein, the simplified aerial triangulation data includes control point files, sub-area image point files, aerial triangulation engineering files containing image lists and stereo model list information, and coordinate system parameter files, but does not include original image files.

[0030] In the above technical solution, in step 4, each sub-area needs to be merged and summarized, and the aerial triangulation results of each sub-area are summarized as a complete survey area by merging the uploaded aerial triangulation simplified data of all sub-areas; the sub-area merging mainly includes project merging and connection point merging:

[0031] (1) Project merging, including image list merging, stereo model merging, merging the project files of multiple sub-areas into a complete survey area project file, and deleting duplicate images and stereo models during merging;

[0032] (2) Connection point merging, since the connection point numbers of different sub-areas may have the problem of duplication, in order to ensure the uniqueness of the connection point numbers after merging the survey area, it is necessary to remove and uniformly number the connection points. The specific method is to add a large number before the sub-area connection point number, the large number is the sub-area number, and the number of digits of the large number is one less than the maximum number of digits of all connection point numbers, so as to retain all connection points of duplicate images in different sub-areas, and establish the connection relationship between different survey areas through the connection points of different survey areas existing on the duplicate images, so as to realize automatic edge connection of the survey area without the need for matching or manual connection point conversion.

[0033] In the above technical solution, in step 5, the regional network adjustment of the complete survey area obtained by merging is solved, the positioning parameters of the images meeting the accuracy requirements are obtained according to the operation specification, and are uploaded to the center server or the cloud for downloading by each sub-area.

[0034] It should be noted that the present application has strong practicability and is suitable for aerial triangulation of multiple types of images (such as satellite images, aerial images, and unmanned aerial vehicle images), and the positioning parameter forms of different types of images are different. Satellite images are generally RPC parameters, and aerial and unmanned aerial vehicle images are inner and outer orientation elements. However, the present method can be applied to different types of positioning parameters, and is uniformly uploaded to the center server as a result, and each sub-area can be downloaded separately, which can realize compatibility for various positioning parameter forms; and the present method overcomes the problem of small application range of the prior art, which is only suitable for a certain type of image (aerial image or unmanned aerial vehicle image) and cannot be applied to satellite image aerial triangulation.

[0035] In the above technical solution, in step 6, each sub-area obtains the precise positioning parameters after the overall solution from the cloud or the center server, updates the sub-area image positioning parameters, and obtains the sub-area aerial triangulation result. Each sub-area takes the image name contained in its own survey area as a unique identifier, takes the positioning parameters of the sub-area image from all the image precise positioning parameters uploaded in step 5, and saves and updates them. When downloading, each sub-area only needs to download the positioning parameters corresponding to the images contained in its own survey area.

[0036] The large-scale image in the wide area network cooperative mode large-scale image aerial triangulation method refers to continent-level satellite images. The method of automatically dividing a survey area and edge connection, sub-area aerial triangulation, and returning simplified aerial triangulation parameters through a wide area network is adopted to realize the reuse of computing resources in each place, realize large-range (such as covering a country or even a continent) satellite image aerial triangulation, and is suitable for large data and high measurement efficiency. The method overcomes the problems that the existing technology adopts a manual point insertion method or a partition method for aerial triangulation operation, the partition and edge connection are basically realized by manual assistance, the method is only suitable for small data survey areas (such as a province or several cities) for aerial image aerial triangulation, cannot be used for decomposition and merging of large survey areas, cannot effectively use multi-cluster computing resources in different places, cannot realize large-scale satellite image aerial triangulation with large data, and has low measurement efficiency.

[0037] The present application has the following advantages:

[0038] (1) The computing resources of multi-clusters in different places are fully utilized, the problem that only the computing resources in one place are difficult to complete the task on time in the case of needing to quickly process the task is solved, the computing efficiency and the ability of emergency production are improved, the problems that the existing technology can process small data (mainly in the partition aerial triangulation in one area), cannot use the computing resources of multi-clusters in different places (the existing technology depends on original image data to use the computing resources in different places, needs to transfer image data, the data volume of image data is large, the transmission speed is slow, and the cost is high), the computing efficiency is low, and the emergency situation cannot be completed are overcome;

[0039] (2) The partition production method is adopted, the requirement for the operation performance of a single cluster is reduced, the problem that the computing resources are insufficient to cause difficulty in processing a large amount of data is solved, the ability and efficiency of processing a large amount of data are improved, the automatic partition method of the present application is not only suitable for the case that the survey area is small (such as covering a certain area in a city), but also suitable for the case that the survey area is large (such as covering a province / country or even a continent), the application range is large and the efficiency is high, the problems that the existing technology is only suitable for the case that the survey area is small (such as covering a certain area in a city), manually divides data by means of external software, and the application range of the survey area is small and the efficiency is low are overcome;

[0040] (3) The characteristics that adjustment calculation does not depend on original image data are fully utilized, the data volume of cloud uploading and downloading of data is small, the transmission is fast, and the time waste of a large amount of image data transmission is avoided;

[0041] (4) The application adopts the parallel aerial triangulation method of partition data processing for the processing of satellite image data at the continent level, can fully utilize the computing resources of multi-cluster in different places, and simultaneously utilizes the characteristics that adjustment calculation is not dependent on original image data, in the transmission process of uploading the aerial triangulation result to the central server or cloud through the wide area network, each sub-measurement area only needs to upload the simplified aerial triangulation data (including the control point file, the sub-area image point file, the aerial triangulation engineering file containing the image list and the stereoscopic model list information, and the coordinate system parameter file) for adjustment calculation, ensures that the data amount of cloud uploading and downloading is small, and the transmission is fast; simultaneously, the realization of large-scale data processing can make the integrity of the adjustment result better, and reduce the system error; overcomes the problems that the image aerial triangulation of the prior art needs to depend on the original image data and control data of the measurement area, and the data amount is large, simultaneously overcomes the problems that the prior art cannot realize large-scale satellite image aerial triangulation, and can only separately perform aerial triangulation measurement on different regional areas, cannot realize the overall processing of large-scale data, and the system error is large (for example, the aerial triangulation measurement processing of each province is separately performed, which may lead to the corresponding system error of different provinces, thereby increasing the system error of large-scale data processing);

[0042] (5) The application is suitable for the overall process of satellite aerial triangulation, has high measurement efficiency, and processes large data amount; overcomes the problems that the prior art is only suitable for the measurement of control data under the condition of partition aerial triangulation, has low measurement efficiency, and processes small data amount;

[0043] (6) The clustering method of setting parameters (such as the number of clusters) can realize uniform automatic partition, or automatic division according to the scope of administrative region, or automatic division according to the regular grid of longitude and latitude, or automatic division according to geographical properties under the condition of adding auxiliary data; these partition methods do not need direct participation of manual work, only need auxiliary work such as manual adjustment of options and setting of some parameters, and do not need mobilization of other software; the application automatically completes edge connection, partition and edge connection efficiency and precision are high, and the application is fully automatic during the adjustment of the combined area; overcomes the problems that the partition and edge connection of the prior art basically depend on manual realization, are difficult to realize automation, have low partition and edge connection efficiency, and have low precision. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 It is the overall technical flowchart of the application;

[0045] Figure 2 It is the overall image distribution schematic diagram of the measurement area of the embodiment of the application;

[0046] Figure 3 It is the result schematic diagram of the K-means clustering division measurement area method of the application;

[0047] Figure 4Fig. 1 is a schematic diagram of overlapping images along the boundary line for dividing the survey area of the present application;

[0048] Figure 5 Fig. 2 is a distribution map of images for each sub-area of the present application;

[0049] Figure 6 Fig. 3 is a distribution map of control points and checkpoints for each sub-area of the present application;

[0050] Figure 7 Fig. 4 is a distribution map of connection points for each sub-area of the present application;

[0051] Figure 8 Fig. 5 is a diagram of simplified aerial triangulation data composition of the present application;

[0052] Figure 9 Fig. 6 is a schematic diagram of the unified numbering method for connection points of the present application. DETAILED DESCRIPTION

[0053] The embodiments of the present application will be described in detail below with reference to the accompanying drawings, but they do not constitute a limitation on the present application, but only serve as an example. At the same time, the advantages of the present application are made clearer and easier to understand through the description.

[0054] The present application discloses a large-scale image aerial triangulation method based on a wide area network collaborative mode, which core idea is to decompose the large-scale image aerial triangulation task into sub-tasks, and distribute them to multiple computing clusters located in different places through a wide area network, thereby improving the processing efficiency. First, according to certain rules, the survey area is automatically decomposed into multiple sub-areas with certain overlap, which are distributed to computing clusters in different places; then, independent aerial triangulation processing is performed according to the sub-areas, and the simplified aerial triangulation data is uploaded to the cloud or the central server for merging; finally, the overall block adjustment solution is performed on the merged survey area, and the adjustment results are distributed to the sub-areas. This method is no longer limited to a single computer or a single computing cluster, and can maximize the use of distributed computing clusters for distributed aerial triangulation, solving the problem that only relying on the computing resources in one place is difficult to complete the task on time in the case of needing to process the task quickly, improving the computing efficiency and the ability of emergency production. In addition, the present application also solves the problems of inconsistency between the aerial triangulation results of each data processing node and low transmission efficiency, proposes a collaborative aerial triangulation method of task distribution, task merging and adjustment solution, and distribution of adjustment results, and the concept of simplified aerial triangulation data, reduces the data transmission amount of the wide area network, and improves the operation efficiency and the overall accuracy consistency of the survey area.

[0055] The present application is mainly realized by the following technical solutions: based on the idea of wide area network cooperation mode, the aerial triangulation process is decomposed from the original data stage, multiple clusters located in different places are used to give full play to their respective computing resources to complete the aerial triangulation task of each sub-area, to ensure that the results of each sub-area meet the requirements, then the simplified aerial triangulation data of each sub-area is uploaded to the central server or cloud, and the sub-area merging and joint adjustment calculation is completed to ensure the consistency of the overall aerial triangulation results. The overall technical process is shown in Figure 1

[0056] The core process includes the following steps:

[0057] Step 1, area decomposition. The complete survey area is automatically divided into multiple sub-areas according to a certain grid or task sub-area range;

[0058] Step 2, sub-area aerial triangulation. For each sub-area, after obtaining the original image and control data, the aerial triangulation operation is independently carried out according to the aerial triangulation operation specification to ensure that the results of the sub-area meet the requirements;

[0059] Step 3, sub-area aerial triangulation simplified data uploading and summarizing. The simplified aerial triangulation data of each sub-area which can be used for adjustment calculation is uploaded and summarized to the central server or cloud;

[0060] Step 4, area merging. The summarized aerial triangulation simplified data of all sub-areas is merged into the complete survey area and the survey area automatic edge matching is carried out;

[0061] Step 5, overall adjustment. The complete survey area obtained by merging is subjected to joint adjustment calculation to obtain the precise positioning parameters of all images, and is uploaded to the central server or cloud.

[0062] ​Step 6, update sub-area image positioning parameters. Obtain the positioning parameters of the sub-area image from all the image positioning parameters uploaded in step 5 and save them. The present application first automatically partitions the image and integrates the aerial triangulation without relying on the original data. Through automatic partitioning of the original image and corresponding algorithms, the extraction of sub-area aerial triangulation data and the operation of integrated aerial triangulation are realized, while the resources are dispersed to different places for collaborative processing. It is suitable for aerial triangulation of satellite images with large data volume and coverage range (capable of obtaining a large amount of data at one time), real-time monitoring and analysis of global or large-scale scenes (such as climate change research, global forest coverage monitoring, and marine protection), and improves processing efficiency; solves the problem that the prior art is only suitable for aerial triangulation of aerial images with small data volume and coverage range, which is suitable for scenes requiring high precision and local detailed information (such as urban three-dimensional modeling, building deformation monitoring, and archaeological site identification), and cannot realize large-scale satellite image data division, calculation, and transmission for aerial triangulation, and cannot realize large-scale satellite image aerial triangulation with too much data (for example, satellite image data with a data level of TB).

[0063] In step 1, the steps of measuring area decomposition are as follows:

[0064] First, calculate the geographic range of each image in the measuring area according to the initial image positioning parameters or directly extract it from the image metadata information;

[0065] Second, count the union of the geographic ranges of all images to obtain the geographic range of the entire measuring area as shown in Figure 2

[0066] Then, the entire measuring area is divided into multiple sub-areas according to certain rules, including:

[0067] (1) Division according to a regular grid. The measuring area is divided into multiple sub-areas according to a certain grid size;

[0068] (2) Division according to administrative regions. Division according to the administrative regions corresponding to the task areas of multiple off-site units participating in the work;

[0069] (3) Division according to clustering. Cluster the point set formed by the center points of the stereo image pairs composed of multiple images (such as K-means clustering), and the number of clusters can be set according to actual conditions. After clustering, calculate the convex hull (i.e., the minimum circumscribed convex polygon) of each cluster. The division effect is shown in Figure 3

[0070] (4) Division according to geographic attributes. Such as division according to high and low altitude regions, urban and forest regions, etc. ​​

[0071] The above division manners can be selected as needed, or several of them can be combined.

[0072] Finally, in order to ensure that the subsequent sub-areas can complete the merging, the adjacent sub-areas should ensure that they overlap by 1-2 image pairs or images along their boundary lines. As shown in Figure 4 , which is a schematic diagram of the division result in a rectangular grid and the sub-area overlap, wherein the small quadrilateral vectors are the geographic ranges of the image pairs, the blue color is the overlapped image pairs, the black color is the non-overlapped images, and the red vectors are the ranges of the divided sub-areas.

[0073] In step 2, after assigning the corresponding work units to each sub-area, the original images and geometric control data within the range of the sub-area are respectively applied for, the aerial triangulation work is independently carried out according to the relevant work specifications, and it is ensured that the aerial triangulation result meets the accuracy requirements. In addition, each sub-area should measure the control points within the range of the sub-area according to the work specifications. The image distribution of each sub-area, the distribution of the control points and the check points, and the distribution of the connection points are respectively shown in Figure 5 , Figure 6 and Figure 7 The above Figure 1 shows the entire process; Figure 2 shows the image distribution of the entire survey area; the above Figure 3 and Figure 4 respectively embody the two automatic division manners of the survey area according to the clustering manner and the latitude and longitude manner; Figure 5 shows the image distribution of the sub-survey area; Figure 6 , Figure 7 respectively show that the distribution of the connection points and the control points of each sub-survey area in the present application is uniform, which indicates that the aerial triangulation effect of the sub-survey area in the present application is good and the accuracy is high; Figure 2 , Figure 5 , Figure 6 , Figure 7 The blue boundary lines in Figure 2 , Figure 4 , Figure 5 , Figure 6 all represent the boundary lines of a province in the embodiment of the present application and the surrounding provinces; Figure 6 The small quadrilaterals in Figure 7 all represent the distribution of the connection points, wherein the red color represents one-degree overlap, the magenta color represents two-degree overlap, and the green color represents three-degree or more overlap.

[0074] In step 3, after completing the sub-area aerial triangulation task, the sub-area aerial triangulation result needs to be uploaded to the central server or cloud through the wide area network for subsequent aggregation and merging. In order to improve transmission efficiency, the data volume is reduced as much as possible. Therefore, in the actual transmission process, each sub-area only needs to upload the simplified aerial triangulation data used for adjustment and solution, avoiding the transmission of large-volume data such as original image data. As shown in FIG. 8, the simplified aerial triangulation data includes the control point file, the sub-area image point file, the aerial triangulation engineering file containing image list and stereo model list information, and the coordinate system parameter file. Figure 8 The wide area network of the present application refers to a computer network covering a wide geographical area, which can connect local area networks, computers, terminals and other devices distributed in different places to realize resource sharing and communication. The present application uploads the sub-area aerial triangulation result to the central server or cloud through the wide area network to realize the extraction of sub-area aerial triangulation data and the operation of merged area aerial triangulation, realize large-scale satellite image aerial triangulation, and at the same time, distribute to different places to utilize the resources of different places for collaborative processing, fully utilize the computing cluster of different places, and improve the processing efficiency. The present application overcomes the problem of the prior art that the connection of the regional network is realized by offline area connection and edge connection (the regional network of photogrammetry refers to a specific area divided from multiple image areas with overlapping images in the photogrammetry process for overall adjustment and improvement of measurement accuracy), and the applicable range of the measurement area is small and the efficiency is low.

[0075] This step only transmits the necessary data for area merging and joint adjustment, i.e. simplified aerial triangulation data, including image list and positioning parameter information, stereo model list, image point observation data, and control point coordinates, but does not include original image files, thereby greatly reducing the data and improving the transmission speed, making the partitioned aerial triangulation result suitable for transmission in the wide area. The transmission of partitioned aerial triangulation result in the wide area can realize the sharing of aerial triangulation result and facilitate subsequent area merging operation (first transmit the simplified aerial triangulation data to one place and then perform area merging operation), realize large-scale satellite image aerial triangulation, reduce the cost of data transmission in processing large-scale satellite image data, improve data transmission efficiency and save cost. The present application solves the problem of the prior art that the original image data and control data are relied on, the aerial triangulation result data is too large to be transmitted in the wide area. At the same time, the present application overcomes the problem of the prior art that the aerial triangulation processing flow is only applicable to small data volume and coverage range, and is suitable for aerial images requiring high precision and local detailed information (such as urban three-dimensional modeling, building deformation monitoring, and archaeological site identification), and cannot be applied to satellite images with large data volume and coverage range (capable of acquiring a large amount of data at one time) and suitable for global or large-scale real-time monitoring and analysis (such as climate change research, global forest coverage monitoring, and marine protection).

[0076] In step 4, the sub-regions are merged and summarized. By merging the simplified aerial triangulation data from all uploaded sub-regions, the aerial triangulation results from each sub-region are summarized into a complete survey area. Sub-region merging includes:

[0077] (1) Project merging. This includes merging image lists and stereo pairs, combining project files from multiple sub-regions into a single project file for a complete survey area. Since adjacent sub-regions have a certain degree of overlap, duplicate images and stereo pairs should be removed during the merging process.

[0078] (2) Connector point merging and unified numbering: Connectors on overlapping images within sub-regions need to be merged. Since connector numbers in different sub-regions may have duplicate names, to ensure the uniqueness of connector numbers after merging, deduplication and unified numbering are required. Specifically, a large number is added before the sub-region connector number; this large number represents the sub-region number. The added digit is the first digit of the largest digit in all connector numbers, ensuring that connector numbers in different sub-regions are located in different, non-overlapping intervals. For example... Figure 9 As shown, the new connection point number consists of the sub-region number and the sub-region connection point number. For example, connection point 1000345 in sub-region 1 is renamed 11000345 after merging, and connection point 1000345 in sub-region 2 is renamed 21000345 after merging, ensuring the uniqueness of the point name and identifying which sub-region the connection point comes from. Deduplication and unified numbering of connection points are performed to retain all connection points of duplicate images in different sub-regions. Connection relationships between different survey areas are established through connection points existing in different survey areas on duplicate images, thus achieving automatic edge joining of survey areas. Since the automatic division of survey areas in step 1 repeatedly divides some images into different sub-regions (several images overlapped during the automatic division of survey areas in step 1), connection point merging and unified numbering during survey area merging avoid point name conflicts, thus retaining all connection points of duplicate images in different sub-regions, thereby achieving automatic edge joining of survey areas without the need for re-matching or manual remapping of connection points. For example, image a is repeatedly assigned to survey area 1 and survey area 2. When aerial triangulation is performed in the survey area, there will be connection points. Image a may have connection points such as 3003, 3004, 3005 in survey area 1, and 6111, 6112, 6113 in survey area 2. When the sub-areas are merged, they are renumbered, and all connection points of image a in survey areas 1 and 2 are retained. Therefore, after merging the areas, there will be connection points such as 13003, 13004, 13005, 26111, 26112, 26113. Since one image a has points in two survey areas (survey area 1 and survey area 2), this image a can establish the connection relationship between the two survey areas (survey area 1 and survey area 2), thereby realizing automatic edge stitching.

[0079] In step 5, the integrated area is subjected to overall area network adjustment calculation, so that accurate positioning parameters of all images are obtained, and are uploaded to the central server or cloud for downloading by each sub-area. When the integrated area is subjected to joint adjustment calculation, the operation specification also needs to be met, so that the accurate positioning parameters of the images meeting the accuracy requirement are obtained.

[0080] It should be noted that the positioning parameter forms of different types of images are different. The satellite images are generally RPC parameters, and the aerial and unmanned aerial vehicle images are inner and outer orientation elements. However, the method can be applied to different types of positioning parameters, and is uniformly uploaded to the central server as the result, and each sub-area can be downloaded, and compatibility can be realized for various positioning parameter forms.

[0081] In step 6, each sub-area obtains the accurate positioning parameters after the overall calculation from the cloud or the central server, and updates the sub-area image positioning parameters. Each sub-area takes the image name contained as the unique identifier, and obtains the positioning parameters of the sub-area image from all the image accurate positioning parameters uploaded in step 5, and saves and updates.

[0082] Embodiment

[0083] Now, the application is described in detail by taking the satellite image aerial triangulation of a province as an embodiment, and the application has a guiding role in other large-scale image aerial triangulation.

[0084] In this example, 138 stereo scene resource three satellite image data (each stereo scene contains three images of front view, normal view and lower view) of a province are selected, the area of the survey area is about 160,000 square kilometers, and a total of 404 images are included.

[0085] The technical scheme provided by the application is: firstly, the sub-area is decomposed, then the sub-area aerial triangulation is performed, the aerial triangulation data is uploaded after the completion of the aerial triangulation of each sub-area, the merging of the sub-area projects and the adjustment calculation of the overall project are performed in the cloud, the adjustment results are generated, then the corresponding updated adjustment results are imported by each sub-area, and the update of the aerial triangulation results is completed.

[0086] The process of collaborative aerial triangulation is as shown in Figure 1 , and includes the following steps:

[0087] Step 1: firstly, the geographical range of each image in the survey area is calculated according to the initial positioning parameters, and the survey area range is obtained by merging, as shown in Figure 2 . The original data is divided into four survey areas, and the data amount of each sub-area is ensured to be similar, and there is an overlap of 1-2 stereo scenes at the sub-area boundary.

[0088] Since the stereoscopic scene data of the target survey area 138 is distributed in the east-west direction, according to the actual situation, due to the irregular distribution of the survey area, the K-means clustering method is used for division, so as to ensure that the image data is as evenly distributed as possible and that there is 1-2 scene images overlapping at the indirect edge of the sub-area. The actual division is divided into 4 sub-areas, which respectively contain 44, 45, 34 and 41 stereoscopic scenes (in this embodiment, the division and edge connection of the survey area are automatically realized by the set program without manual intervention; the program for automatic division of the survey area in this embodiment is: clustering method automatic division program; the automatic edge connection program of the survey area is: when merging the survey area, the merging and unified numbering are realized through the connection points to avoid point name conflicts, so as to retain all connection points of the repeated images in different sub-areas).

[0089] Step 2: According to the sub-area divided in step 1, independent space triangulation is carried out in the sub-area to ensure that the results of the sub-area meet the requirements (such as Figure 3 as shown, Figure 3 which is the effect of automatic division of the clustering method in this embodiment).

[0090] For each sub-area, first match the connection points to ensure uniform distribution of the connection points, then measure the control points, the control point distribution is as shown in Figure 6 , and finally perform adjustment calculation. The limit error of the image points in the adjustment process is 2 pixels, and the limit error of the control points in the plane and elevation is 12 meters and 10 meters respectively. The control point distribution of the four sub-areas is as shown in Figure 6 , and the number of control points is 283, 225, 168 and 199 respectively; the connection point distribution is as shown in Figure 7 , and the number is 374918, 369589, 263662 and 245881 respectively.

[0091] Step 3: Upload the simplified space triangulation data of each sub-area, including image point files, engineering files, control point files and other data to the cloud. The size of the space triangulation simplified data file of the four sub-areas is 30.3 MB, 34.0 MB, 21.4 MB and 23.7 MB respectively (the data amount of the space triangulation simplified data file in this embodiment is small, and the transmission is fast; compared with other existing methods, the image data transmitted by the method of this embodiment is much smaller in magnitude, and the image data transmitted by other existing methods is generally in the order of GB or more).

[0092] Step 4: Merge the space triangulation results of each sub-area in the cloud, eliminate the repeated images and image points, and combine the space triangulation results of the four sub-areas into a complete survey area and perform automatic edge connection of the survey area. After merging, the number of connection points and control points of the survey area is 1254050 and 875 respectively. The engineering merging includes merging of image list and merging of stereoscopic model, merging the engineering files of the four sub-areas into a complete engineering file of the survey area, and deleting the repeated images and stereoscopic models during merging. When merging the connection points, the connection points of different sub-areas are uniformly renumbered.

[0093] Step 5, the combined complete survey area is subjected to area network adjustment calculation. The point error of adjustment is 0.34 pixels, the plane and height errors of the control points are 2.84 meters and 1.59 meters, the plane and height errors of the check points are 2.74 meters and 1.90 meters, the size of the generated exterior orientation element file is 76 KB, and the file is uploaded to the central server or cloud for downloading by each sub-area.

[0094] Step 6, the updated geometric positioning parameters are distributed according to the sub-areas. Each sub-area processing unit opens the project of the sub-area, imports the adjustment results, automatically downloads and updates the newly generated geometric positioning parameter file according to the image name of the sub-area from the central server or cloud (when the application is applied to other large-scale image (such as national or continental) satellite image aerial triangulation, the application can be used to automatically divide the large-scale image into multiple provinces or multiple countries, and each province or country (i.e. each sub-area) processes the data of its own province or country (i.e. performs sub-area aerial triangulation) to improve the calculation efficiency and the ability of emergency production by using the computing resources of other places (i.e. the computing resources of each province), and then performs merging and overall adjustment of each province or country (i.e. sub-area), and finally distributes the adjustment results; when the application is applied to other types of images (such as satellite images, aerial images, unmanned aerial vehicle images) aerial triangulation, the measurement method is the same as above).

[0095] Conclusion: In this embodiment, the application is used for satellite image aerial triangulation in a province, and the computing clusters in other places are fully utilized to quickly complete the aerial triangulation of the task area, and the partitioning and edge connecting can be automatically performed, the partitioning and edge connecting efficiency and accuracy are high, the data amount uploaded and downloaded in the cloud is small, and the transmission speed is fast.

[0096] The large-scale image aerial triangulation method based on the wide area network collaborative mode provided by the application aims to integrate multiple computing clusters distributed in other places through a wide area network, divide the aerial triangulation of the overall survey area into multiple sub-areas through partitioning, and perform independent aerial triangulation in other places, and after the operation is completed, submit the simplified aerial triangulation data (not containing image data, only containing point observation, control points, image list and positioning parameter information) of the sub-area through the wide area network, and then perform overall area network adjustment calculation, so that the task requirements of large-scale satellite image rapid processing are maximized; and the partitioning and edge connecting in the prior art are basically realized by manual assistance, which is difficult to automate, and the overall accuracy of the aerial triangulation is also affected to a certain extent.

[0097] The present application mainly solves the problem that the traditional cluster parallel processing method cannot effectively utilize the resources of off-site computing cluster, and proposes a large-scale aerial triangulation method of a task flow of "measurement area decomposition-sub-area off-site processing-sub-area merging and overall adjustment-adjustment result distribution". The original data is divided from the task level, each sub-area off-site only needs to process the data of the measurement area of the local sub-area, and finally uploaded to the cloud is only the aerial triangulation simplified data without the original image file, the data amount is small, can realize fast uploading, and meets the demand of emergency production.

[0098] The specific embodiments described herein are merely illustrative of the present application. Various modifications or changes can be made to the described embodiments without departing from the spirit of the application. For example, the specific configurations described are illustrative only and other configurations can be substituted to implement the principles of the present application. Accordingly, the disclosure is not intended to be limited to the described embodiments, but is to be accorded the widest scope consistent with the claims.

[0099] Other parts not described belong to the prior art.

Claims

1. A method for large-scale aerial triangulation of imagery in a WAN collaborative mode, characterized by: Based on the wide area network cooperation mode, the aerial triangulation process is decomposed from the original data stage, the images are automatically partitioned, the large-scale image aerial triangulation task is decomposed into subtasks, and they are distributed to multiple computing clusters located in different places through the wide area network; the aerial triangulation of satellite images, aerial images and unmanned aerial vehicle images suitable for global or large-scale real-time monitoring and analysis; The method comprises the following steps: Step 1, area decomposition; According to a certain grid or task sub-area range, the complete survey area is automatically divided into multiple sub-areas; The steps of area decomposition are as follows: Firstly, the geographical range of each image in the survey area is calculated according to the initial positioning parameters of the image or directly extracted from the image metadata information; Secondly, the union of the geographical ranges of all images is calculated to obtain the geographical range of the entire survey area; Then, the entire survey area is divided into multiple sub-areas according to certain rules, and the division methods include: (1) division according to a regular grid; the survey area is divided into multiple sub-areas according to a certain grid size; (2) division according to administrative regions; division is performed according to the administrative regions corresponding to the task areas of multiple units participating in the work; (3) division according to clustering; clustering is performed on the point set formed by the center points of the stereo image pairs composed of multiple images, and the number of clusters can be set according to actual conditions; (4) division according to geographical attributes; The above division methods can be selected according to needs, or several methods can be combined; Finally, in order to ensure that the subsequent sub-areas can be combined, adjacent sub-areas should ensure that there are overlapping image pairs or images along their boundary lines; Step 2, sub-area aerial triangulation; After obtaining the original images and geometric control data for each sub-area, independent aerial triangulation is performed according to the aerial triangulation operation specification to ensure that the sub-area results meet the requirements; Step 3, uploading and summarizing of sub-area simplified aerial triangulation data; The simplified aerial triangulation data of each sub-area that can be used for adjustment and calculation is uploaded and summarized to a central server or cloud; Step 4, area combination; All the summarized sub-area aerial triangulation simplified data is combined into a complete survey area and the survey area is automatically connected; The sub-area combination includes: (1) project combination, including image list combination, stereo model combination, combining the project files of multiple sub-areas into a complete survey area project file, and deleting duplicate images and stereo models during combination; (2) connection point combination, a large number is added before the connection point number of the sub-area, the large number is the sub-area number, and the number of digits of the large number is one less than the maximum number of digits of all connection point numbers, so as to retain all connection points of the duplicate images in different sub-areas, and establish the connection relationship between different survey areas through the different survey area connection points existing on the duplicate images, so as to realize the automatic connection of the survey area without the need for matching or manual connection point conversion; Step 5, joint adjustment; The complete survey area obtained by combination is subjected to joint adjustment calculation to obtain the precise positioning parameters of all images, and the parameters are uploaded to a central server or cloud; Step 6, updating of sub-area image positioning parameters; The positioning parameters of the sub-area images are taken from the precise positioning parameters of all images uploaded in step 5 and saved.

2. The WAN-cooperative mode large-scale image aerial triangulation method according to claim 1, characterized in that: In step 2, each sub-area work unit respectively applies for original images and control data within the sub-area range, and independently performs aerial triangulation work according to relevant work specifications.

3. The WAN-cooperative mode large-scale image aerial triangulation method according to claim 2, characterized in that: In step 3, after completing the aerial triangulation task in the sub-area, the aerial triangulation result is uploaded to the central server or cloud through the wide area network; in the transmission process, each sub-area uploads simplified aerial triangulation data for adjustment and solution; wherein the simplified aerial triangulation data includes a control point file, a sub-area image point file, an aerial triangulation project file containing image list and stereoscopic model list information, and a coordinate system parameter file.

Images