Multi-view video acquisition and synchronous calibration method and system

Abstract

The invention discloses a multi-view video acquisition and synchronous calibration method and system, relates to the technical field of data calibration, and aims to solve the problem of high calibration difficulty caused by acquisition errors of each view acquisition device in an acquired article. Comprising the following steps: scanning collected articles to obtain the total number of planes of the collected articles, receiving the number of collected viewpoints transmitted by a user side, judging whether to enter a collection point calibration mechanism, obtaining the collection coverage range of viewpoint collection equipment when entering the collection point calibration mechanism, and screening the coordinates of the collection points in combination with the number of the collected viewpoints; setting a collection device to collect the video files according to preset collection points or screened collection points, receiving the video file analysis occupied space amount of each collection point, setting a feature screening mechanism to process the video files of each collection point, and obtaining the number of the same feature points of each video file; and integrating the occupied space amount of each video file and the number of the same feature points, and carrying out calibration and real-time application by using a swing balance algorithm.

Description

Technical Field

[0001] The present invention relates to the technical field of data calibration, and more specifically, to a multi-viewpoint video acquisition and synchronous calibration method and system. Background Art

[0002] Data calibration technology is a technical method used to ensure the accuracy and consistency of data. Data calibration technology applied to video acquisition calibration processing can effectively reduce the differences in multi-viewpoint video acquisition, thereby improving video authenticity.

[0003] The prior art has the following deficiencies:

[0004] In the past, when multiple viewpoints were synchronously calibrated, the viewpoint acquisition equipment was used to perform multi-view acquisition of the acquisition objects through the acquisition points set in the past through the historical database. However, most of the viewpoint acquisition devices were installed only considering the monitoring range, and ignored the impact of differences in the acquisition objects on different viewpoint acquisition devices, resulting in poor video file accuracy. Some of the acquired objects could not be matched with the label information stored in the database, and multi-viewpoint video file calibration was difficult. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a multi-viewpoint video acquisition and synchronous calibration method, which updates the acquisition point coordinates by analyzing the adaptability of the acquisition object and the viewpoint acquisition device model, and performs targeted data calibration after pixel aggregation of the video file to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] The multi-viewpoint video acquisition and synchronization calibration method comprises the following steps:

[0008] Step S1: 3D scan the collected object to obtain the total number of collected object planes, and determine whether to enter the collection point calibration mechanism by receiving the number of collection viewpoints transmitted by the user terminal and combining it with the total number of collected object planes;

[0009] Step S2: when entering the acquisition point calibration mechanism, the acquisition coverage of the viewpoint acquisition device is obtained, and the calibration rules are set based on the number of acquisition viewpoints and the acquisition coverage to obtain the calibrated acquisition point coordinates;

[0010] Step S3: according to the calibrated collection point coordinates or the preset collection point coordinates, a period of time is set to perform video collection of the collection object, the video files of each collection point are received for recording and the storage space occupied by different video files is analyzed, a feature screening mechanism is set to process the video files of each collection point and the number of aggregated pixels with different marks in each video file is obtained;

[0011] Step S4: The space occupied by different video files and the number of aggregated pixels of different tags are comprehensively considered and calibrated using a rocking balance algorithm. The calibration steps are recorded and applied to the current collected object in real time.

[0012] In a preferred embodiment, in step S1, a collection viewpoint quantity window is sent to the user terminal, the user is guided by the window to fill in the number of viewpoint collection devices installed for the current collection object, and the number filled in by the user is received as the collection viewpoint quantity.

[0013] In a preferred embodiment, in step S1, the total number of collected object planes is compared with the number of collected viewpoints. When the total number of collected object planes exceeds the number of collected viewpoints, the collection point calibration mechanism is not entered; when the total number of collected object planes is lower than the number of collected viewpoints, the collection point calibration mechanism is performed.

[0014] In a preferred embodiment, in step S2, when entering the acquisition point calibration mechanism, the model of the viewpoint acquisition device is identified, and the acquisition coverage of the viewpoint acquisition device is obtained from the model database of the corresponding viewpoint acquisition device;

[0015] The number of supplementary viewpoints is obtained by subtracting the number of collection viewpoints from the total number of collection item planes. In the collection items, each collection item plane is matched with a viewpoint collection device, and then the viewpoint collection devices of the supplementary viewpoints are randomly allocated to each collection item plane, and the difference in the number of viewpoint collection devices in each collection item plane does not exceed 1; the collection coverage of each viewpoint collection device in each collection item plane is accumulated to obtain the plane collection range of the corresponding collection item plane;

[0016] The number of collection viewpoints is the number of viewpoint collection devices installed in the collection item.

[0017] In a preferred embodiment, in step S2, the viewpoint acquisition devices allocated in each object acquisition plane are counted to obtain the number of viewpoint acquisition devices in each object acquisition plane and the following rules are set for processing:

[0018] Rule 1: When the number of viewpoint acquisition devices on the acquisition object plane is 1, the center position coordinates of the acquisition object plane are used as the acquisition point coordinates of the viewpoint acquisition device in the corresponding acquisition object plane;

[0019] Rule 2: When the number of viewpoint acquisition devices on the object acquisition plane is not 1 and the plane acquisition range is less than 180 degrees, multiple coordinate combinations that reach the plane acquisition range are screened out and a coordinate combination is randomly selected as the acquisition point coordinates of each viewpoint acquisition device in the corresponding object acquisition plane;

[0020] Rule 3: When the number of viewpoint acquisition devices on the object acquisition plane is not 1 and the plane acquisition range is higher than 180 degrees, the coordinate combination of multiple viewpoint acquisition devices reaching 180 degrees is randomly selected in the object acquisition plane, and the coordinates are randomly assigned to the remaining viewpoint acquisition devices.

[0021] In a preferred embodiment, in step S3, if it is determined that the collection point calibration mechanism is not entered, the video collection of the collection object is performed according to the preset collection point; otherwise, the collection point calibration mechanism is entered to screen each collection point;

[0022] The viewpoint acquisition devices at each acquisition point are used to acquire scene videos of the acquisition objects and obtain video files. The video files are connected to the operating system to obtain the storage space occupied by the video files of each viewpoint acquisition device.

[0023] In a preferred embodiment, in step S3, a feature screening mechanism is set to process the video files collected at each collection point. The specific steps are as follows:

[0024] Step A1: Initialize the video file and mark the video files collected at each collection point with different serial numbers;

[0025] Step A2: extracting pixel points, extracting pixel points of images in video files marked with different serial numbers at the same time point, and obtaining pixel values ​​corresponding to each pixel point;

[0026] Step A3: Merge pixels, compare the pixel values ​​of adjacent pixels, compare the pixel value difference of adjacent pixels with a preset pixel value benchmark, and when the pixel value difference of adjacent pixels is lower than the preset pixel value benchmark, merge the adjacent pixels into an aggregated pixel, and take the average pixel value of the adjacent pixels as the pixel value of the aggregated pixel and compare it with the pixel values ​​of other adjacent pixels to merge the pixels;

[0027] Step A4: Compare the pixel values ​​of each aggregated pixel point finally obtained in the video files marked with different serial numbers, compare the video files in pairs according to the serial numbers, mark the aggregated pixel points with the same pixel values ​​in each video file in the same way, and count the number of aggregated pixel points with different marks in each video file.

[0028] In a preferred embodiment, in step S4, the calibrated steps of processing using a rocking balance algorithm based on the amount of space occupied by different video files and the number of aggregated pixels with different tags are as follows:

[0029] Determine the balance node: sort the sequence numbers of the video files and fill them into the balance nodes in sequence, and classify and mark different balance nodes;

[0030] Balance load calculation: standardize the space occupied by video files in different balance nodes and the number of aggregated pixels of the corresponding video files and sum them to obtain the load coefficient of each balance node;

[0031] Task transfer: Calculate the load average of each balance node, and process the video files in the balance node whose load factor exceeds the load average;

[0032] Load processing: sort the balance nodes for load processing and process them in sequence. When processing the current balance node, calculate the area of ​​the aggregated pixels with different tags in the balance node, and traverse the aggregated pixels with the same tags in other balance nodes. If the area of ​​the aggregated pixels with the tags in the current balance node exceeds the area of ​​the aggregated pixels with the same tags in other balance nodes, retain the aggregated pixel tags in the current balance node. Otherwise, delete the aggregated pixels and traverse the marked balance nodes of the video files in the balance node in sequence for load processing.

[0033] File calibration: Match the same labeled aggregated pixels of the video files in each balance node, and the file calibration is completed.

[0034] A multi-view video acquisition and synchronous calibration system, used to implement the above-mentioned multi-view video acquisition and synchronous calibration method, including a data acquisition module, a coordinate acquisition module, a feature screening module and a calibration adjustment module;

[0035] The data acquisition module is used to perform 3D scanning on the collected objects, obtain the total number of collected object planes, the number of collected viewpoints, and the collection coverage of the viewpoint collection device, and transmit them to the coordinate acquisition module;

[0036] The coordinate acquisition module receives the total number of collected object planes and the number of collected viewpoints and determines whether to perform a collection point calibration mechanism. The calibration rules are set to filter the calibrated collection point coordinates and pass them into the feature screening module.

[0037] The feature screening module sets a period of time to collect videos of the collected objects according to the collection point coordinates and analyzes to obtain the space occupied by the video files, sets a feature screening mechanism to process the video files of each collection point to obtain the number of aggregated pixels with different marks in each video file, and transmits the space occupied by the video files and the number of aggregated pixels to the calibration adjustment module;

[0038] The calibration and adjustment module receives the amount of space occupied by the video file and the number of aggregated pixels, and uses the rocking balance algorithm to process and calibrate the video file and record it for application.

[0039] The technical effects and advantages of the multi-view video acquisition and synchronous calibration method and system of the present invention are as follows:

[0040] The present invention scans the collected objects to obtain the total number of collected object planes and receives the number of collected viewpoints transmitted by the user end to determine whether to enter the collection point calibration mechanism. When entering the collection point calibration mechanism, the collection coverage of the viewpoint collection device is obtained, and the collection point combination is screened in combination with the number of collected viewpoints. The collection device is set to collect video files according to the preset collection points or the collection points screened by the collection point calibration mechanism, which greatly reduces the differences between the viewpoints when the fixed collection point collects video files. The video files of each collection point are received and recorded, and the space occupied by different video files is analyzed. A feature screening mechanism is set to process the video files of each collection point and obtain the number of the same feature points of different video files. The space occupied by different video files and the number of the same feature points are comprehensively considered to perform calibration using a rocking balance algorithm and applied in real time, thereby improving the calibration efficiency of multi-viewpoint video processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a schematic diagram of the multi-viewpoint video acquisition and synchronous calibration method of the present invention.

[0042] Figure 2 This is a flow chart of the multi-viewpoint video acquisition and synchronous calibration system of the present invention. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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 creative work are within the scope of protection of the present invention.

[0044] The present invention scans the collected objects to obtain the total number of collected object planes and receives the number of collected viewpoints transmitted by the user end to determine whether to enter the collection point calibration mechanism. When entering the collection point calibration mechanism, the collection coverage of the viewpoint collection device is obtained, and the collection point combination is screened in combination with the number of collected viewpoints. The collection device is set to collect video files according to the preset collection points or the collection points screened by the collection point calibration mechanism. The video files of each collection point are received and recorded, and the occupied space of different video files is analyzed. A feature screening mechanism is set to process the video files of each collection point and obtain the number of the same feature points of different video files. The space occupied by different video files and the number of the same feature points are comprehensively considered to perform calibration using a rocking balance algorithm and applied in real time, thereby improving the calibration efficiency of multi-viewpoint video processing.

[0045] Embodiment 1, multi-view video acquisition and synchronous calibration method, such as Figure 1 As shown, the following steps are included:

[0046] Step S1: 3D scan the collected object to obtain the total number of collected object planes, and determine whether to enter the collection point calibration mechanism by receiving the number of collection viewpoints transmitted by the user terminal and combining it with the total number of collected object planes;

[0047] Step S2: when entering the acquisition point calibration mechanism, the acquisition coverage of the viewpoint acquisition device is obtained, and the calibration rules are set based on the number of acquisition viewpoints and the acquisition coverage to obtain the calibrated acquisition point coordinates;

[0048] Step S3: according to the calibrated collection point coordinates or the preset collection point coordinates, a period of time is set to perform video collection of the collection object, the video files of each collection point are received for recording and the storage space occupied by different video files is analyzed, a feature screening mechanism is set to process the video files of each collection point and the number of aggregated pixels with different marks in each video file is obtained;

[0049] Step S4: The space occupied by different video files and the number of aggregated pixels of different tags are comprehensively considered and calibrated using a rocking balance algorithm. The calibration steps are recorded and applied to the current collected object in real time.

[0050] The specific implementation is as follows:

[0051] In step S1, the object plane is a two-dimensional plane in a three-dimensional space, and the total number of object planes is the total number of two-dimensional planes of the collected objects. The collected objects are 3D scanned by a laser scanner to obtain the total number of collected object planes;

[0052] Send a collection viewpoint quantity window to the user end, and fill in the number of viewpoint collection devices installed for the current collection object according to the window's instructions, and receive the number filled in by the user as the collection viewpoint quantity;

[0053] Compare the total number of collected item planes with the number of collected viewpoints. When the total number of collected item planes exceeds the number of collected viewpoints, the collection point calibration mechanism is entered; when the total number of collected item planes is lower than the number of collected viewpoints, the collection point calibration mechanism is not performed.

[0054] It should be noted that a laser scanner is a device that uses laser technology to obtain three-dimensional information on the surface of an object. The total number of planes of collected objects can be obtained through the laser scanner. The user end is an interactive port for interacting with the user. Each set of scenes retains the coordinates of past collection points in the historical database before collecting the viewing angle. The number of collection viewpoints is obtained through the user end. If the number of collection viewpoints filled in by the user according to demand is inconsistent with the number of planes of the corresponding collected objects, it is determined after judgment whether to perform a collection point calibration mechanism. When the number of collected object planes is lower than the number of collection viewpoints, that is, the number of viewpoint collection devices is more than the number of collected object planes, each plane of the collected object can be assigned one or more viewpoint collection devices, and the video collection of the collected objects is more comprehensive, and there is no need to calibrate the coordinates of past collection points in the historical database.

[0055] In step S2, if it is determined that the collection point calibration mechanism is not entered, the video collection of the collection object is performed according to the preset collection point; otherwise, the collection point calibration mechanism is entered to screen each collection point;

[0056] When entering the acquisition point calibration mechanism, the model of the viewpoint acquisition device is identified, and the acquisition coverage of the viewpoint acquisition device is obtained from the model database of the corresponding viewpoint acquisition device;

[0057] It should be noted that the model database is a systematic data set used to store, manage and query relevant information under a specific model device, including the collection coverage of the viewpoint collection device; the collection coverage is the angle between the upper and lower edges of the viewpoint collection device. The larger the collection coverage, the wider the collection range of the collected objects, the more repeated content appears in the video file, and it is easier to calibrate the video file; the viewpoint collection device is a device that obtains the video file of the collected object, such as a camera, etc., which will not be elaborated here.

[0058] The number of supplementary viewpoints is obtained by subtracting the number of collected viewpoints from the total number of collected item planes. In the collected items, each collected item plane is matched with a viewpoint collection device, and then the viewpoint collection devices of the supplementary viewpoints are randomly assigned to each collected item plane, and the difference in the number of viewpoint collection devices in each collected item plane does not exceed 1; the collection coverage of each viewpoint collection device in each collected item plane is accumulated to obtain the plane collection range of the corresponding collected item plane. It should be explained that the number of collected viewpoints is the number of installed viewpoint collection devices in the collected items.

[0059] The viewpoint acquisition devices allocated in each object acquisition plane are counted to obtain the number of viewpoint acquisition devices in each object acquisition plane and the following rules are set for processing:

[0060] Rule 1: When the number of viewpoint acquisition devices on the acquisition object plane is 1, the center position coordinates of the acquisition object plane are used as the acquisition point coordinates of the viewpoint acquisition device in the corresponding acquisition object plane;

[0061] Rule 2: When the number of viewpoint acquisition devices on the object acquisition plane is not 1 and the plane acquisition range is less than 180 degrees, multiple coordinate combinations that reach the plane acquisition range are screened out and a coordinate combination is randomly selected as the acquisition point coordinates of each viewpoint acquisition device in the corresponding object acquisition plane;

[0062] Rule 3: When the number of viewpoint acquisition devices collecting the object plane is not 1 and the plane acquisition range is higher than 180 degrees, k coordinate combinations of viewpoint acquisition devices reaching 180 degrees are randomly selected in the object collection plane, and coordinates are randomly assigned to the remaining viewpoint acquisition devices.

[0063] It should be noted that the plane collection range of the object collection plane is the accumulation of the collection ranges of each viewpoint collection device in the corresponding object collection plane. The plane collection range is the maximum value of the collection range of each viewpoint collection device corresponding to the object collection plane without repeated monitoring. When the maximum collection range exceeds 180 degrees, viewpoint collection devices are added to capture only repeated images, and more calibration processing is required.

[0064] In step S3, if the collection point calibration mechanism is not entered, video collection of the collection object is performed according to a period of time set at the preset collection point coordinates. If the collection point calibration mechanism is entered, video collection of the collection object is performed according to a period of time set according to the calibrated collection point coordinates. It should be explained that the preset collection point coordinates are the default collection coordinates in the video collection system, which are set by professionals in this field and are not analyzed here.

[0065] The viewpoint acquisition device at each acquisition point is used to collect the scene video of the acquisition object and obtain the video file, and the video file is connected to the operating system to obtain the storage space occupied by the video file of each viewpoint acquisition device;

[0066] The specific steps for setting up a feature screening mechanism to process the video files collected at each collection point are as follows:

[0067] Step A1: Initialize the video file and mark the video files collected at each collection point with different serial numbers;

[0068] Step A2: extracting pixel points, extracting pixel points of images in video files marked with different serial numbers at the same time point, and obtaining pixel values ​​corresponding to each pixel point;

[0069] Step A3: Merge pixels, compare the pixel values ​​of adjacent pixels, compare the pixel value difference of adjacent pixels with a preset pixel value benchmark, and when the pixel value difference of adjacent pixels is lower than the preset pixel value benchmark, merge the adjacent pixels into an aggregated pixel, and take the average pixel value of the adjacent pixels as the pixel value of the aggregated pixel and compare it with the pixel values ​​of other adjacent pixels to merge the pixels;

[0070] Step A4: Compare the pixel values ​​of each aggregated pixel point finally obtained in the video files marked with different serial numbers, compare the video files in pairs according to the serial numbers, mark the aggregated pixel points with the same pixel values ​​in each video file in the same way, and count the number of aggregated pixel points with different marks in each video file.

[0071] It should be noted that in video files of the same scene, pixels with the same pixel values ​​may be the same object in the scene. After merging the pixels with the same pixel values, if the areas of the same pixel values ​​in two video files are similar after merging, they can be determined to be the same object in the scene.

[0072] In step S4, the occupancy amount of different video files and the number of aggregated pixels of different labels are comprehensively considered and the calibration steps are as follows using the rocking balance algorithm:

[0073] Determine the balance node: sort the sequence numbers of the video files and fill them into the balance nodes in sequence, and classify and mark different balance nodes;

[0074] Balance load calculation: The space occupied by video files in different balance nodes and the number of aggregated pixels of the corresponding video files are normalized using the Min-Max normalization method and summed to obtain the load coefficient of each balance node;

[0075] Task transfer: Calculate the load average of each balance node, and process the video files in the balance node whose load factor exceeds the load average;

[0076] Load processing: sort the balance nodes for load processing and process them in sequence. When processing the current balance node, calculate the area of ​​the aggregated pixels with different tags in the balance node, and traverse the aggregated pixels with the same tags in other balance nodes. If the area of ​​the aggregated pixels with the tags in the current balance node exceeds the area of ​​the aggregated pixels with the same tags in other balance nodes, retain the aggregated pixel tags in the current balance node. Otherwise, delete the aggregated pixels and traverse the marked balance nodes of the video files in the balance node in sequence for load processing.

[0077] File calibration: Match the same labeled aggregated pixels of the video files in each balance node, and the file calibration is completed.

[0078] Save the processing calibration steps after processing the video file according to the rocking balance algorithm.

[0079] It needs to be explained that the larger the amount of space occupied by the video file or the more aggregated pixels there are, the higher the load of the corresponding video file will be. The aggregated pixels with the same marks in the video files in each balance node will be matched, that is, the video files obtained from different acquisition points will be matched and scanned after load processing. When the areas of the aggregated pixels with the same marks in each video file are the same, it means that the part features of the collected objects in the video files of each viewpoint are accurate. The record processing calibration step is to record the updated acquisition point positions and the aggregated pixel records deleted in each video file.

[0080] It should be noted that the swing balance algorithm is an algorithm for solving constrained optimization problems. In this example, it is used to calculate the load average of video files in different balance nodes and judge the load conditions of different video files. Aggregate pixels in high-load video files are deleted to optimize video files acquired at different acquisition points. When the number of aggregate pixels in video files is standardized, the Min-Max standardization formula is: ,in, It is the amount of storage space or the number of aggregated pixels of the video file in each balance node. It is the minimum amount of space occupied by the video files or the number of aggregated pixels in each balance node. It is the amount of space occupied by the video files in each balance node or the maximum number of aggregated pixels. It is the normalized result of the amount of space occupied by the video files or the number of aggregated pixels in each balance node.

[0081] Embodiment 2, multi-view video acquisition and synchronous calibration system, such as Figure 2 As shown, it is used to implement a multi-viewpoint video acquisition and synchronous calibration method, including a data acquisition module, a coordinate acquisition module, a feature screening module and a calibration adjustment module;

[0082] The data acquisition module is used to perform 3D scanning on the collected objects, obtain the total number of collected object planes, the number of collected viewpoints, and the collection coverage of the viewpoint collection device, and transmit them to the coordinate acquisition module;

[0083] The coordinate acquisition module receives the total number of collected object planes and the number of collected viewpoints and determines whether to perform a collection point calibration mechanism. The calibration rules are set to filter the calibrated collection point coordinates and pass them into the feature screening module.

[0084] The feature screening module sets a period of time to collect videos of the collected objects according to the collection point coordinates and analyzes to obtain the space occupied by the video files, sets a feature screening mechanism to process the video files of each collection point to obtain the number of aggregated pixels with different marks in each video file, and transmits the space occupied by the video files and the number of aggregated pixels to the calibration adjustment module;

[0085] The calibration and adjustment module receives the amount of space occupied by the video file and the number of aggregated pixels, and uses the rocking balance algorithm to process and calibrate the video file and record it for application.

[0086] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0087] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application of the technical solution and the invention constraints. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0088] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0089] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0090] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A multi-view video acquisition and synchronization calibration method, characterized in that: The following steps are included: Step S1: 3D scan the collected object to obtain the total number of collected object planes, and determine whether to enter the collection point calibration mechanism by receiving the number of collection viewpoints transmitted by the user terminal and combining it with the total number of collected object planes; Step S2: when entering the acquisition point calibration mechanism, the acquisition coverage of the viewpoint acquisition device is obtained, and the calibration rules are set based on the number of acquisition viewpoints and the acquisition coverage to obtain the calibrated acquisition point coordinates; Step S3: according to the calibrated collection point coordinates or the preset collection point coordinates, a period of time is set to perform video collection of the collection object, the video files of each collection point are received for recording and the storage space occupied by different video files is analyzed, a feature screening mechanism is set to process the video files of each collection point and the number of aggregated pixels with different marks in each video file is obtained; Step S4: The space occupied by different video files and the number of aggregated pixels of different tags are comprehensively considered and calibrated using a rocking balance algorithm. The calibration steps are recorded and applied to the current collected object in real time.

2. The multi-view video acquisition and synchronization calibration method according to claim 1, characterized in that: In step S1, a collection viewpoint quantity window is sent to the user terminal, and the user is guided by the window to fill in the number of viewpoint collection devices installed for the current collection object, and the number filled in by the user is received as the collection viewpoint quantity.

3. The multi-view video acquisition and synchronization calibration method according to claim 1, characterized in that: In step S1, the total number of collected object planes is compared with the number of collected viewpoints. When the total number of collected object planes exceeds the number of collected viewpoints, the collection point calibration mechanism is entered; when the total number of collected object planes is lower than the number of collected viewpoints, the collection point calibration mechanism is not performed.

4. The multi-view video acquisition and synchronization calibration method according to claim 1, characterized in that: In step S2, when entering the acquisition point calibration mechanism, the model of the viewpoint acquisition device is identified, and the acquisition coverage of the viewpoint acquisition device is obtained from the model database of the corresponding viewpoint acquisition device; The number of supplementary viewpoints is obtained by subtracting the number of collection viewpoints from the total number of collection item planes. In the collection items, each collection item plane is matched with a viewpoint collection device, and then the viewpoint collection devices of the supplementary viewpoints are randomly allocated to each collection item plane, and the difference in the number of viewpoint collection devices in each collection item plane does not exceed 1; the collection coverage of each viewpoint collection device in each collection item plane is accumulated to obtain the plane collection range of the corresponding collection item plane; The number of collection viewpoints is the number of viewpoint collection devices installed in the collection item.

5. The multi-view video acquisition and synchronization calibration method according to claim 1, characterized in that: In step S2, the viewpoint acquisition devices allocated in each object acquisition plane are counted to obtain the number of viewpoint acquisition devices in each object acquisition plane and the following rules are set for processing: Rule 1: When the number of viewpoint acquisition devices on the acquisition object plane is 1, the center position coordinates of the acquisition object plane are used as the acquisition point coordinates of the viewpoint acquisition device in the corresponding acquisition object plane; Rule 2: When the number of viewpoint acquisition devices on the object acquisition plane is not 1 and the plane acquisition range is less than 180 degrees, multiple coordinate combinations that reach the plane acquisition range are screened out and a coordinate combination is randomly selected as the acquisition point coordinates of each viewpoint acquisition device in the corresponding object acquisition plane; Rule 3: When the number of viewpoint acquisition devices on the object acquisition plane is not 1 and the plane acquisition range is higher than 180 degrees, the coordinate combination of multiple viewpoint acquisition devices reaching 180 degrees is randomly selected in the object acquisition plane, and the coordinates are randomly assigned to the remaining viewpoint acquisition devices.

6. The multi-view video acquisition and synchronization calibration method according to claim 3, characterized in that: In step S3, if it is determined that the collection point calibration mechanism is not entered, the video collection of the collection object is performed according to the preset collection point; otherwise, the collection point calibration mechanism is entered to screen each collection point; The viewpoint acquisition devices at each acquisition point are used to acquire scene videos of the acquisition objects and obtain video files. The video files are connected to the operating system to obtain the storage space occupied by the video files of each viewpoint acquisition device.

7. The multi-view video acquisition and synchronization calibration method according to claim 6, characterized in that: In step S3, a feature screening mechanism is set to process the video files collected at each collection point. The specific steps are as follows: Step A1: Initialize the video file and mark the video files collected at each collection point with different serial numbers; Step A2: extracting pixel points, extracting pixel points of images in video files marked with different serial numbers at the same time point, and obtaining pixel values ​​corresponding to each pixel point; Step A3: Merge pixels, compare the pixel values ​​of adjacent pixels, compare the pixel value difference of adjacent pixels with a preset pixel value benchmark, and when the pixel value difference of adjacent pixels is lower than the preset pixel value benchmark, merge the adjacent pixels into an aggregated pixel, and take the average pixel value of the adjacent pixels as the pixel value of the aggregated pixel and compare it with the pixel values ​​of other adjacent pixels to merge the pixels; Step A4: Compare the pixel values ​​of each aggregated pixel point finally obtained in the video files marked with different serial numbers, compare the video files in pairs according to the serial numbers, mark the aggregated pixel points with the same pixel values ​​in each video file in the same way, and count the number of aggregated pixel points with different marks in each video file.

8. The multi-view video acquisition and synchronization calibration method according to claim 7, characterized in that: In step S4, the occupancy amount of different video files and the number of aggregated pixels of different labels are comprehensively considered and the calibration steps are as follows using the rocking balance algorithm: Determine the balance node: sort the sequence numbers of the video files and fill them into the balance nodes in sequence, and classify and mark different balance nodes; Balance load calculation: standardize the space occupied by video files in different balance nodes and the number of aggregated pixels of the corresponding video files and sum them to obtain the load coefficient of each balance node; Task transfer: Calculate the load average of each balance node, and process the video files in the balance node whose load factor exceeds the load average; Load processing: sort the balance nodes for load processing and process them in sequence. When processing the current balance node, calculate the area of ​​the aggregated pixels with different tags in the balance node, and traverse the aggregated pixels with the same tags in other balance nodes. If the area of ​​the aggregated pixels with the tags in the current balance node exceeds the area of ​​the aggregated pixels with the same tags in other balance nodes, retain the aggregated pixel tags in the current balance node. Otherwise, delete the aggregated pixels and traverse the marked balance nodes of the video files in the balance node in sequence for load processing. File calibration: Match the same-labeled aggregated pixels of the video files in each balance node, and the file calibration is completed; Match the same-marked aggregated pixel points in the video files of each balance node, that is, perform matching scans on the video files obtained from different acquisition points after load processing. When the areas of the same-marked aggregated pixel points in each video file are the same, it means that the part features of the captured objects in the video files of each viewpoint are accurate; The record processing calibration step is to record the updated acquisition point positions and the aggregated pixel point records deleted in each video file.

9. A multi-view video acquisition and synchronous calibration system, based on the multi-view video acquisition and synchronous calibration method according to any one of claims 1 to 8, characterized in that: It includes a data acquisition module, a coordinate acquisition module, a feature screening module and a calibration adjustment module; The data acquisition module is used to perform 3D scanning on the collected objects, obtain the total number of collected object planes, the number of collected viewpoints, and the collection coverage of the viewpoint collection device, and transmit them to the coordinate acquisition module; The coordinate acquisition module receives the total number of collected object planes and the number of collected viewpoints and determines whether to perform a collection point calibration mechanism. The calibration rules are set to filter the calibrated collection point coordinates and pass them into the feature screening module. The feature screening module sets a period of time to collect videos of the collected objects according to the collection point coordinates and analyzes to obtain the space occupied by the video files, sets a feature screening mechanism to process the video files of each collection point to obtain the number of aggregated pixels with different marks in each video file, and transmits the space occupied by the video files and the number of aggregated pixels to the calibration adjustment module; The calibration and adjustment module receives the amount of space occupied by the video file and the number of aggregated pixels, and uses the rocking balance algorithm to process and calibrate the video file and record it for application.

Images