A cloud platform camera continuous snapshot target identification early warning deduplication method and system, device and medium

Abstract

The present application relates to a kind of pan-tilt camera continuous snapshot target identification early warning deduplication method and system, equipment and medium, by obtaining the installation height of pan-tilt camera, latitude and longitude and the administrative region belonging to, generates position parameter;Get the snapshot image that pan-tilt camera is along the set snapshot track, snapshot time shoots, and the PTZ value of the pan-tilt camera corresponding to snapshot time, generate image parameter;Splice position parameter and image parameter generate image analysis parameter;After using target identification algorithm to identify target, generate early warning in combination with image analysis parameter;And simultaneously using time sequence rule and space rule to judge the effectiveness of early warning, solve the problem of repeated early warning, significantly reduce artificial review workload, greatly save manpower cost;While improving supervision response speed, early warning aggregation greatly shortens average disposal time, makes regional illegal construction containment rate increase compared with last year;Precise early warning makes natural resource encroachment behavior discovery window period advance, further strengthens ecological protection.

Description

Technical Field

[0001] This invention relates to the field of natural resource conservation technology based on visual detection, and in particular to a deduplication method, system, device, and medium for continuous target capture, identification, and early warning using a PTZ camera. Background Technology

[0002] In the dynamic monitoring and protection of natural resources, PTZ cameras based on machine vision are typically used to continuously capture images for several hours. Target recognition algorithms are then used to identify targets within the captured images, enabling automatic early warning of illegal construction activities. Specifically, a combination of spatiotemporal overlap detection (such as GPS coordinates and timestamps) and simple feature matching (such as construction machinery type) is used to determine and issue early warnings for illegal construction activities.

[0003] However, in large-scale monitoring scenarios, the high dispersion of monitoring targets, changes in shooting angles, and redundant detection between consecutive frames can lead to a single illegal event triggering alarms multiple times, resulting in duplicate warning data. Duplicate warnings can cause multiple verifications of the same area, increasing the costs of field investigations, data processing, and equipment scheduling; or they can exaggerate the severity of the problem, leading to excessive enforcement (such as repeatedly halting legal projects) or resource misallocation; simultaneously, redundant data consumes storage and computing resources, causing delays in the monitoring platform's response. Summary of the Invention

[0004] Based on this, the purpose of this invention is to provide a deduplication method for continuous target capture and early warning by a PTZ camera, so as to solve the problem of repeated alarms caused by a single illegal event being triggered multiple times in a large-scale monitoring scenario, resulting in duplicate early warning data.

[0005] A deduplication method for continuous target capture and early warning using a PTZ camera includes the following steps:

[0006] S10: Obtain the installation height, latitude and longitude, and administrative region of the PTZ camera, and generate location parameters;

[0007] S20: Acquire the captured image taken by the PTZ camera along the set capture trajectory and capture time, and generate image parameters by combining the PTZ value of the PTZ camera corresponding to the capture time;

[0008] S30: Concatenate the position parameters with the image parameters to generate image analysis parameters;

[0009] S40: Employ a target recognition algorithm to detect illegal construction targets in the captured image and determine whether any illegal construction targets exist in the captured image.

[0010] If so, then based on the mapping relationship between the pixel coordinates of the PTZ camera and the world coordinates and planar coordinates, the world coordinates and planar coordinates of the illegal construction target in the captured image are obtained; combined with the image analysis parameters, an early warning message with warning information and warning image is generated.

[0011] S50: Analyze the warning using the warning determination rules to determine whether the warning is valid:

[0012] If yes, then store the warning in the warning dataset;

[0013] If not, then delete the warning.

[0014] Compared with existing technologies, this invention generates location parameters by acquiring the installation height, latitude and longitude, and administrative region of the PTZ camera; it generates image parameters by acquiring images captured by the PTZ camera along a set capture trajectory and capture time, and the PTZ value of the PTZ camera corresponding to the capture time; it generates image analysis parameters by stitching together the location parameters and image parameters; it generates an early warning by using a target recognition algorithm to identify the target and combining the image analysis parameters; and it uses temporal and spatial rules to judge the effectiveness of the early warning, solving the problem of duplicate early warnings, significantly reducing the workload of manual review, and greatly saving labor costs; at the same time, it improves the speed of regulatory response, and the aggregation of early warnings greatly shortens the average handling time, resulting in a year-on-year increase in the rate of curbing illegal construction in the region; precise early warnings advance the window of opportunity for detecting natural resource encroachment, further strengthening ecological protection, and providing reliable quantitative evidence for law enforcement agencies.

[0015] Furthermore, the warning information includes: the warning image address. Warning occurrence time af-t ij , Target plane coordinates af-(x i ,y i ), types of early warning targets The administrative region (af-ID) and algorithm type (af-alg) of the PTZ camera; where θ i This represents the i-th capture trajectory point of the PTZ camera, where i ∈ (1, N), and N represents the N capture trajectory points set; t ij This indicates the j-th capture time point of the PTZ camera at the i-th capture trajectory point, where j∈(1,M), and M represents the j-th capture time point within the set capture time period;

[0016] S50 includes the following sub-steps:

[0017] S51A: Set timing determination rules (tr), and set core timing verification parameters including: the administrative region (af-ID) of the PTZ camera and the type of warning target. Warning occurrence time af-t ijAlgorithm type af-alg;

[0018] S52A: If the warning occurs at time af-t ij If the time series determination rule tr is greater than the time series, the early warning dataset is traversed based on the core time series verification parameters:

[0019] If old data corresponding to the warning exists, it means that the warning is a duplicate warning and should be deleted.

[0020] If no old data corresponding to the warning exists, then a time series dataset Ω that meets the conditions is obtained. t ;

[0021] Among them, the time-series judgment rule tr represents the set warning time duration threshold;

[0022] S51B: Set spatial determination rules (sr), and set core spatial verification parameters including: the administrative region (AF-ID) of the PTZ camera and the type of warning target. Early warning target plane coordinates af-(x) i ,y i Algorithm type af-alg, old plane coordinates of the warning target af-(x) 0i ,y 0i );

[0023] S52B: Calculates the plane coordinates of the early warning target using a spatial position calculation function af-(x) i ,y i ) and the old plane coordinates of the early warning target af-(x 0i ,y 0i If the distance d is less than or equal to the spatial decision rule sr, then the warning dataset is traversed according to the spatial core verification parameters.

[0024] If old data corresponding to the warning exists, it means that the warning is a duplicate warning and should be deleted.

[0025] If no old data corresponding to the warning exists, then a spatial dataset Ω that meets the conditions is obtained. s ;

[0026] Among them, the spatial determination rule sr represents the set warning distance threshold;

[0027] S53: For the time series dataset Ω t and spatial dataset Ω s Perform set operations to remove duplicates, and obtain the intersection ΔΩ:

[0028] If the intersection ΔΩ is empty, then the warning is a valid warning, and the warning is stored in the warning dataset;

[0029] If the intersection ΔΩ is not empty, then the warning is a duplicate warning and the warning is deleted.

[0030] Furthermore, the mapping relationship between the pixel coordinates of the PTZ camera and the world coordinates and planar coordinates is obtained in the following way:

[0031] SA1 finds a distinctive reference point P in an image from a PTZ camera. k Obtain reference point P k pixel coordinates P k (u Pk ,v Pk );

[0032] SA2 finds the reference point P in the map tool. k To obtain the reference point P corresponding to the actual location. k latitude and longitude coordinates P k (lat Pk ,lon Pk ), and set the reference point P k latitude and longitude coordinates P k (lat Pk ,lon Pk Convert to world coordinates P k (X W-Pk ,Y W-Pk Z W-Pk );

[0033] SA3 repeats steps SA1 and SA2 to obtain multiple sets of reference point pairs between pixel coordinates and world coordinates;

[0034] SA4 uses stochastic gradient descent to calculate the camera's intrinsic and extrinsic parameters based on multiple sets of reference points between pixel coordinates and world coordinates.

[0035]

[0036] In the formula: M is the intrinsic parameter matrix of the PTZ camera, [R, T] is the extrinsic parameter matrix of the PTZ camera, R is a 3×3 rotation matrix, T is a 3×1 translation matrix; k∈(1,L), and L≥5;

[0037] SA5 establishes a mapping relationship between pixel coordinates and world coordinates based on the camera's intrinsic and extrinsic parameters:

[0038]

[0039] SA6 obtains the planar coordinates (x, y) based on world coordinates:

[0040] (x, y) = (X W Y W ).

[0041] Furthermore, reference point P kThe conversion between latitude coordinates and world coordinates is as follows:

[0042] SC1 uses the Gaussian projection forward calculation formula to calculate the reference point P. k latitude and longitude coordinates Convert to planar coordinates

[0043] SC2 obtains reference point P from the elevation file DEM. k latitude and longitude The corresponding altitude h is used to obtain the world coordinates (X). W-Pk ,Y W-Pk Z W-Pk )=(x Pk ,y Pk ,h Pk ).

[0044] Furthermore, the mapping relationship between the pixel coordinates of the PTZ camera and the world coordinates and planar coordinates can also be obtained through the following method:

[0045] The SD1 acquires several calibration board images from different angles to calibrate the PTZ camera, thereby obtaining the intrinsic parameter matrix M of the PTZ camera.

[0046] SD2 uses an alignment algorithm to calculate the rotation and translation vectors of the pixel coordinates of the calibration board with known world coordinates, and obtains the rotation matrix R and the translation matrix T; where R is a 3×3 matrix and T is a 3×1 matrix.

[0047] SD3 establishes the mapping relationship between pixel coordinates and world coordinates based on the intrinsic parameter matrix M, rotation matrix R, and translation matrix T of the PTZ camera:

[0048]

[0049] SD4 obtains the planar coordinates (x, y) from the world coordinates:

[0050] (x, y) = (X W Y W ).

[0051] Furthermore, the image parameters are represented as I(I1,I2,...,I...). i ,...,I N ), where I i satisfy:

[0052]

[0053] Where: img represents the image address; P represents the horizontal azimuth angle of the PTZ camera; T represents the vertical pitch angle of the PTZ camera; Z represents the zoom level of the PTZ camera; θ iThis represents the i-th capture trajectory point of the PTZ camera, where i ∈ (1, N), and N represents the N capture trajectory points set; t ij This indicates the j-th capture time point of the PTZ camera at the i-th capture trajectory point, where j∈(1,M), and M represents the j-th capture time point within the set capture time period.

[0054] Furthermore, the image analysis parameters are represented as A(A1,A2,...,A...). i ,...,A N ), where A i satisfy:

[0055] A i =(I i ,H,lon,lat,af-ID)

[0056] Among them, I i The image parameters of the PTZ camera at the i-th capture trajectory point are represented by H, H represents the installation height of the PTZ camera, (lon,lat) represents the latitude and longitude of the PTZ camera, and af-ID represents the administrative region to which the PTZ camera belongs.

[0057] Meanwhile, this invention also provides a deduplication system for continuous target capture and early warning using a PTZ camera, comprising a PTZ camera and a deduplication device. The PTZ camera is used to capture images of the target monitoring area and transmits the captured images and their corresponding PTZ values ​​to the deduplication device. The deduplication device includes a position parameter acquisition module, an image parameter acquisition module, a parameter integration module, an early warning generation module, and an early warning deduplication module.

[0058] The location parameter acquisition module is used to acquire the installation height, latitude and longitude, and administrative region of the PTZ camera, and generate location parameters.

[0059] The image parameter acquisition module is used to acquire the captured images taken by the PTZ camera along the set capture trajectory and capture time, and generate image parameters by combining them with the PTZ value of the PTZ camera corresponding to the capture time.

[0060] The parameter integration module is used to concatenate position parameters and image parameters to generate image analysis parameters;

[0061] The warning generation module is used to detect illegal construction targets in the captured image using a target recognition algorithm, and to determine whether there are illegal construction targets in the captured image:

[0062] If so, then based on the mapping relationship between the pixel coordinates of the PTZ camera and the world coordinates and planar coordinates, the world coordinates and planar coordinates of the illegal construction target in the captured image are obtained; combined with the image analysis parameters, an early warning message with warning information and warning image is generated.

[0063] The early warning deduplication module is used to analyze early warnings using early warning judgment rules to determine whether an early warning is valid.

[0064] If yes, then store the warning in the warning dataset;

[0065] If not, then delete the warning.

[0066] Compared with the prior art, the beneficial effects of the deduplication system for continuous target capture and early warning of PTZ cameras of the present invention and the deduplication method for continuous target capture and early warning of PTZ cameras described above will not be repeated here.

[0067] To better understand and implement this invention, the following detailed description is provided in conjunction with the accompanying drawings. Attached Figure Description

[0068] Figure 1 This is a schematic diagram of the deduplication system for continuous target capture, identification, and early warning using a PTZ camera according to the present invention.

[0069] Figure 2 This is a flowchart illustrating the deduplication method for continuous target capture, identification, and early warning using a PTZ camera according to the present invention. Detailed Implementation

[0070] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings of the embodiments of the present invention.

[0071] To address the issue of duplicate warning data arising from the repeated triggering of alarms for a single illegal event during the dynamic monitoring and protection of natural resources in large-scale monitoring scenarios using machine vision-based PTZ cameras to continuously capture images of the monitoring area for several hours, and then using spatiotemporal overlap detection and simple feature matching to determine the early warning of illegal construction activities based on the continuous capture data, this invention proposes a deduplication method for target identification and early warning in continuous capture using PTZ cameras. This deduplication method for continuous target capture and early warning using PTZ cameras generates location parameters by acquiring the camera's installation height, latitude and longitude, and administrative region; it also generates image parameters by acquiring images captured by the PTZ camera along a set capture trajectory and at a set capture time, along with the PTZ value of the PTZ camera corresponding to the capture time; the method then stitches the location parameters and image parameters together to generate image analysis parameters; finally, it uses a target recognition algorithm to identify the target and combines the image analysis parameters to generate an early warning; and it simultaneously uses temporal and spatial rules to determine the effectiveness of the early warning, solving the problem of duplicate warnings, significantly reducing the workload of manual review, and greatly saving labor costs. It also improves the speed of regulatory response; the aggregation of early warnings greatly shortens the average processing time, resulting in a year-on-year increase in the rate of curbing illegal construction in the region. Precise early warnings advance the window of opportunity for detecting natural resource encroachment, further strengthening ecological protection.

[0072] Based on the deduplication method for continuous target capture and early warning using a PTZ camera, this invention also proposes a deduplication system for continuous target capture and early warning using a PTZ camera.

[0073] Specifically, please refer to Figure 1 The deduplication system for continuous target capture and early warning of PTZ camera according to the present invention includes a PTZ camera 100 and a deduplication device 200, wherein the PTZ camera 100 and the deduplication device 200 are communicatively connected.

[0074] The pan-tilt camera 100 is used to capture images of the target monitoring area and transmit the captured images and the corresponding PTZ values ​​to the deduplication device 200.

[0075] Specifically, the PTZ camera 100 can acquire and transmit its current PTZ value in real time. Specifically, the PTZ value includes the horizontal azimuth angle Pan, the vertical pitch angle Tilt, and the zoom level of the PTZ camera.

[0076] Please see Figure 1 and Figure 2 The deduplication device 200 is used to obtain the installation height, latitude and longitude and the administrative region of the PTZ camera 100, and to receive the captured images with PTZ values ​​taken by the PTZ camera 100, and to execute the deduplication method for continuous capture target recognition and early warning of the PTZ camera. It includes: a position parameter acquisition module 210, an image parameter acquisition module 220, a parameter integration module 230, an early warning generation module 240, and an early warning deduplication module 250.

[0077] The location parameter acquisition module 210 is used to perform step S10: acquire the installation height H, latitude and longitude (lon,lat) and administrative region af-ID of the PTZ camera, and generate location parameter L=(H,lon,lat,af-ID).

[0078] The image parameter acquisition module 220 is used to perform step S20: acquire the captured image taken by the PTZ camera along the set capture trajectory and capture time, and combine it with the PTZ value of the PTZ camera corresponding to the capture time to generate image parameters I(I1,I2,...,I...). i ,...,I N ), i∈(1,N).

[0079] The rotation trajectory and rotation time of the PTZ camera are set to obtain the camera's capture trajectory and capture time. Furthermore, the capture frequency can be set to obtain multiple capture images at the same capture trajectory point and capture time point.

[0080] Specifically, it records the captured images at each capture trajectory point and different capture time points, along with the corresponding PTZ value of the PTZ camera, and generates image parameters I(I1,I2,...,I...). i ,...,I N ), I i satisfy:

[0081]

[0082] Where: img represents the image address; P represents the horizontal azimuth angle of the PTZ camera; T represents the vertical pitch angle of the PTZ camera; Z represents the zoom level of the PTZ camera; θ i This represents the i-th capture trajectory point of the PTZ camera, where i ∈ (1, N), and N represents the N capture trajectory points set; t ij This indicates the j-th capture time point of the PTZ camera at the i-th capture trajectory point, where j∈(1,M), and M represents the j-th capture time point within the set capture time period.

[0083] The parameter integration module 230 is used to perform step S30: concatenating the position parameters and image parameters to generate image analysis parameters A(A1,A2,...,A1). i ,...,A N ), i∈(1,N).

[0084] Among them, A i =(I i ,H,lon,lat,af-ID).

[0085] The warning generation module 240 is used to execute step S40: using a target recognition algorithm to detect illegal construction targets in the captured image, and determining whether there are illegal construction targets in the captured image.

[0086] If so, then based on the mapping relationship between the pixel coordinates of the PTZ camera and the world coordinates and planar coordinates, the world coordinates and planar coordinates of the illegal construction target in the captured image are obtained; combined with the image analysis parameters, an early warning message with warning information and warning image is generated.

[0087] If no, then the process ends.

[0088] Specifically, the target recognition algorithm can be a YOLO series, R-CNN series, or DETR series, etc., and this application does not impose any restrictions. In this embodiment, the YOLOv9 model is used, and the YOLOv9 model is pre-trained using a dataset containing illegal construction targets. The illegal construction targets include, but are not limited to, engineering operation equipment, buildings, and building materials. Engineering operation equipment includes, for example, excavators, bulldozers, and drilling rigs; buildings include, for example, houses under construction and fences; and building materials include, for example, piles of sand and gravel.

[0089] The mapping relationship between the pixel coordinates of the pan-tilt camera and the world coordinates and planar coordinates is obtained in the following way:

[0090] SA1 finds a distinctive reference point P in an image from a PTZ camera. k Obtain reference point P k pixel coordinates P k (u Pk ,v Pk ).

[0091] Specifically, reference point P k Preferably, locations include street corners and building corners.

[0092] SA2 finds the reference point P in the map tool. k To obtain the reference point P corresponding to the actual location. k latitude and longitude coordinates P k (lat Pk ,lon Pk ), and set the reference point P k latitude and longitude coordinates P k (lat Pk ,lon Pk Convert to world coordinates P k (X W-Pk ,Y W-Pk Z W-Pk ).

[0093] Among them, reference point P k The conversion between latitude coordinates and world coordinates is as follows.

[0094] SC1 uses the Gaussian projection forward calculation formula to calculate the reference point P. k latitude and longitude coordinates Convert to planar coordinates

[0095] The specific calculation formula is as follows:

[0096]

[0097] In the formula: X is the arc length of the meridian; N is the radius of curvature of the meridian circle; l″ = lon - L0, where L0 is the longitude of the central meridian; t = tan(lat); η 2 =e′ 2 cos 2 (lat), e′ is the second eccentricity of the ellipsoid.

[0098] SC2 obtains reference point P from the elevation file DEM. k latitude and longitude The corresponding altitude h is used to obtain the world coordinates (X). W-Pk ,Y W-Pk Z W-Pk )=(x Pk ,y Pk ,h Pk ).

[0099] SA3 repeats steps SA1 and SA2 to obtain multiple sets of reference point pairs between pixel coordinates and world coordinates.

[0100] SA4 uses stochastic gradient descent to calculate the camera's intrinsic and extrinsic parameters based on multiple sets of reference points between pixel coordinates and world coordinates.

[0101]

[0102] In the formula: M is the intrinsic parameter matrix of the PTZ camera, [R, T] is the extrinsic parameter matrix of the PTZ camera, R is a 3×3 rotation matrix, and T is a 3×1 translation matrix; k∈(1,L), and L≥5.

[0103] SA5 establishes a mapping relationship between pixel coordinates and world coordinates based on the camera's intrinsic and extrinsic parameters:

[0104]

[0105] SA6 obtains the planar coordinates (x, y) based on world coordinates:

[0106] (x, y) = (X W Y W ).

[0107] The mapping relationship between the pixel coordinates of the pan-tilt camera and the world coordinates and planar coordinates can also be obtained through the following method:

[0108] The SD1 acquires several calibration board images from different angles to calibrate the PTZ camera, obtaining the intrinsic parameter matrix M of the PTZ camera.

[0109] SD2 uses an alignment algorithm to calculate the rotation and translation vectors of the pixel coordinates of the calibration board with known world coordinates, and obtains the rotation matrix R and the translation matrix T; where R is a 3×3 matrix and T is a 3×1 matrix.

[0110] SD3 establishes the mapping relationship between pixel coordinates and world coordinates based on the intrinsic parameter matrix M, rotation matrix R, and translation matrix T of the PTZ camera:

[0111]

[0112] SD4 obtains the planar coordinates (x, y) from the world coordinates:

[0113] (x, y) = (X W Y W ).

[0114] The warning image is an image containing an illegal construction target, and the illegal construction target is not located at the edge of the image. In this embodiment, an image with an area of ​​the illegal construction target larger than 64*64 pixels is selected.

[0115] The warning information includes: warning image address Warning occurrence time af-t ij , Target plane coordinates af-(x i ,y i ), types of early warning targets The administrative region (af-ID) and algorithm type (af-alg) of the PTZ camera are specified. The algorithm type is used to enable the deduplication rules to filter and determine the warning targets more precisely.

[0116] The warning target refers to any existing illegal construction targets.

[0117] The early warning deduplication module 250 is used to execute step S50: analyze the early warning using early warning determination rules to determine whether the early warning is a valid early warning.

[0118] If so, save the warning;

[0119] If not, then delete the warning.

[0120] The early warning determination rules include a time-series rule unit, a spatial rule unit, and a merging and deduplication unit.

[0121] Specifically, step S50 includes the following sub-steps:

[0122] The timing rule unit is used to execute steps S51A and S52A.

[0123] S51A: Set timing determination rules (tr), and set core timing verification parameters including: the administrative region (af-ID) of the PTZ camera and the type of warning target. Warning occurrence time af-t ij Algorithm type af-alg;

[0124] S52A: If the warning occurs at time af-t ij If the time series determination rule tr is greater than the time series, the early warning dataset is traversed based on the core time series verification parameters:

[0125] If old data corresponding to the warning exists, it means that the warning is a duplicate warning and should be deleted.

[0126] If no old data corresponding to the warning exists, then a time series dataset Ω that meets the conditions is obtained.t .

[0127] The time-series judgment rule 'tr' represents the set threshold for the duration of the warning. Illegal activities typically have a persistent characteristic over a certain period; setting the time-series judgment rule 'tr' avoids generating a large number of identical warnings within a short period.

[0128] Spatial rule unit, used to execute steps S51B and S52B.

[0129] S51B: Set spatial determination rules (sr), and set core spatial verification parameters including: the administrative region (AF-ID) of the PTZ camera and the type of warning target. Early warning target plane coordinates af-(x) i ,y i Algorithm type af-alg, old plane coordinates of the warning target af-(x) 0i ,y 0i );

[0130] S52B: Calculates the plane coordinates of the early warning target using a spatial position calculation function af-(x) i ,y i ) and the old plane coordinates of the early warning target af-(x 0i ,y 0i If the distance d is less than or equal to the spatial decision rule sr, then the warning dataset is traversed according to the spatial core verification parameters.

[0131] If old data corresponding to the warning exists, it means that the warning is a duplicate warning and should be deleted.

[0132] If no old data corresponding to the warning exists, then a spatial dataset Ω that meets the conditions is obtained. s .

[0133] Among them, the spatial determination rule sr represents the set warning distance threshold.

[0134] The deduplication unit is merged and used to execute step S53.

[0135] S53: For the time series dataset Ω t and spatial dataset Ω s Perform set operations to remove duplicates, and obtain the intersection ΔΩ:

[0136] If the intersection ΔΩ is empty, then the warning is a valid warning, and the warning is stored in the warning dataset;

[0137] If the intersection ΔΩ is not empty, then the warning is a duplicate warning and the warning is deleted.

[0138] The spatial location calculation function is ST_Distance_Sphere, a spatial calculation function provided by MySQL / MariaDB.

[0139] Example: If the time series dataset Ω t ={1, 2, 3}, spatial dataset Ω s If ={1,4}, then the intersection ΔΩ={1} indicates that the warning has existed.

[0140] The aforementioned deduplication device is stored in an electronic device and is executed by this electronic device using the deduplication method for continuous target capture and early warning by the PTZ camera. The electronic device includes, but is not limited to, memory, processor, and network interface that can be interconnected via a system bus.

[0141] The electronic device can be a rack server, blade server, tower server, or cabinet server, or other computing device. The electronic device can be a standalone server or a server cluster composed of multiple servers.

[0142] The memory includes at least one type of readable storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. The memory can be an internal storage unit of the electronic device, such as the hard disk or RAM of the electronic device. The memory can also be an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc. The memory may also include both internal storage units and external storage devices of the electronic device.

[0143] The processor can be a central processing unit (CPU), controller, microcontroller, microprocessor, or other data processing chip. This processor is typically used to control the overall operation of the electronic device, such as performing control and processing related to data interaction or communication with the electronic device. The processor is used to run program code stored in the memory or process data, such as running the deduplication method for continuous camera capture, target recognition, and early warning.

[0144] The network interface may include a wireless network interface or a wired network interface, which is typically used to establish communication connections between the electronic device and other electronic devices. For example, the network interface is used to connect the electronic device to an external data platform via a network, establishing a data transmission channel and communication connection between the electronic device and the external data platform. The network may be an intranet, the Internet, Global System for Mobile communication (GSM), Wideband Code Division Multiple Access (WCDMA), 4G network, 5G network, Bluetooth, Wi-Fi, or other wireless or wired networks.

[0145] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to limit the embodiments of this application. The singular forms “a,” “the,” and “the” used in the embodiments and claims of this application are also intended to include the plural forms, unless the context clearly indicates otherwise. It should also be understood that, unless otherwise stated, “a plurality” and “several” refer to two or more; “and / or” refers to and includes any or all possible combinations of one or more associated listed items; “first,” “second,” “third,” etc., are used only to distinguish and not to describe a particular order or sequence, nor should they be construed as indicating or implying relative importance. When the above description relates to drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. In the description of this application, those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0146] The embodiments described above are merely examples of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and the present invention also intends to include these modifications and variations.

Claims

1. A deduplication method for continuous target capture, identification, and early warning using a PTZ camera, characterized in that, Includes the following steps: S10: Obtain the installation height, latitude and longitude, and administrative region of the PTZ camera, and generate location parameters; S20: Acquire the captured image taken by the PTZ camera along the set capture trajectory and capture time, and generate image parameters by combining the PTZ value of the PTZ camera corresponding to the capture time; S30: Concatenate the position parameters with the image parameters to generate image analysis parameters; S40: Employ a target recognition algorithm to detect illegal construction targets in the captured image and determine whether any illegal construction targets exist in the captured image. If so, then based on the mapping relationship between the pixel coordinates of the PTZ camera and the world coordinates and planar coordinates, the world coordinates and planar coordinates of the illegal construction target in the captured image are obtained; combined with the image analysis parameters, an early warning message with warning information and warning image is generated. The warning information includes: warning image address Warning time Target plane coordinates Types of early warning targets Administrative regions where PTZ cameras belong and algorithm type ;in, This represents the i-th capture trajectory point of the PTZ camera, where i∈(1,N), and N represents the N capture trajectory points set. This indicates the j-th capture time point of the PTZ camera at the i-th capture trajectory point, where j∈(1,M), and M represents the j-th capture time point within the set capture time period; S50: Analyze the warning using the warning determination rules to determine whether the warning is valid: If yes, then store the warning in the warning dataset; If not, then delete the warning; Specifically, it includes: S51A: Setting Timing Determination Rules The core verification parameters for setting the timing include: the administrative region to which the PTZ camera belongs. Types of early warning targets Warning time Algorithm type ; S52A: If the warning occurs at the time Greater than timing determination rules The early warning dataset is traversed based on the time-series core verification parameters: If old data corresponding to the warning exists, it means that the warning is a duplicate warning and should be deleted. If no old data corresponding to the warning exists, then a time series dataset that meets the conditions is obtained. ; Among them, the timing judgment rule This indicates the set threshold for the duration of the warning. S51B: Setting Space Determination Rules The core verification parameters for the space include: the administrative region to which the PTZ camera belongs. Types of early warning targets Target plane coordinates Algorithm type Old plane coordinates of the early warning target ; S52B: Calculates the planar coordinates of early warning targets using a spatial position calculation function. old plane coordinates of the early warning target distance If the distance Less than or equal to space determination rules Then, the early warning dataset is traversed based on the spatial core verification parameters: If old data corresponding to the warning exists, it means that the warning is a duplicate warning and should be deleted. If no old data corresponding to the warning exists, then a spatial dataset that meets the conditions is obtained. ; Among them, spatial determination rules This indicates the set warning distance threshold; S53: For time series datasets and spatial datasets Perform set operations to remove duplicates and obtain the intersection. : If the intersection If the value is empty, the warning is a valid warning, and the warning is stored in the warning dataset; If the intersection If the value is not empty, the warning is a duplicate warning and should be deleted.

2. The deduplication method for continuous target capture and early warning using a PTZ camera according to claim 1, characterized in that, The mapping relationship between the pixel coordinates of the pan-tilt camera and the world coordinates and planar coordinates is obtained in the following way: SA1 finds a distinctive reference point in an image from a PTZ camera. Obtain reference points pixel coordinates ; SA2 finds reference points in the map tools. Obtain reference points corresponding to the actual locations. latitude and longitude coordinates and reference point latitude and longitude coordinates Convert to world coordinates ; SA3 Repeat steps SA1 and SA2 to obtain multiple sets of reference point pairs between pixel coordinates and world coordinates; SA4 uses stochastic gradient descent to calculate the camera's intrinsic and extrinsic parameters based on multiple sets of reference points between pixel coordinates and world coordinates. In the formula: M is the intrinsic parameter matrix of the PTZ camera, [R, T] is the extrinsic parameter matrix of the PTZ camera, R is a 3×3 rotation matrix, and T is a 3×1 translation matrix; k∈(1,L), and L≥5; SA5 establishes a mapping relationship between pixel coordinates and world coordinates based on the camera's intrinsic and extrinsic parameters: ； SA6 obtains planar coordinates based on world coordinates. : 。 3. The deduplication method for continuous target capture and early warning using a PTZ camera according to claim 2, characterized in that, Reference point The conversion between latitude coordinates and world coordinates is as follows: SC1 uses the Gaussian projection forward calculation formula to calculate the reference point. latitude and longitude coordinates Convert to planar coordinates ; SC2 Obtains Reference Points from Elevation File DEM latitude and longitude The corresponding altitude h is used to obtain world coordinates. .

4. The deduplication method for continuous target capture and early warning using a PTZ camera according to claim 1, characterized in that, The mapping relationship between the pixel coordinates of the pan-tilt camera and the world coordinates and planar coordinates can also be obtained through the following method: SD1 acquires several calibration board images from different angles to calibrate the PTZ camera and obtains the intrinsic parameter matrix M of the PTZ camera; SD2 uses an alignment algorithm to calculate the rotation and translation vectors of the pixel coordinates of the calibration board with known world coordinates, and obtains the rotation matrix R and the translation matrix T; where R is a 3×3 matrix and T is a 3×1 matrix. SD3 establishes the mapping relationship between pixel coordinates and world coordinates based on the intrinsic parameter matrix M, rotation matrix R, and translation matrix T of the PTZ camera: ； SD4 obtains planar coordinates based on world coordinates. : 。 5. The deduplication method for continuous target capture and early warning using a PTZ camera according to any one of claims 1-4, characterized in that, The image parameters are represented as follows: ,in, satisfy: Where: img represents the image address; P represents the horizontal azimuth angle of the PTZ camera; T represents the vertical pitch angle of the PTZ camera; Z represents the zoom level of the PTZ camera; This represents the i-th capture trajectory point of the PTZ camera, where i∈(1,N), and N represents the N capture trajectory points set. This indicates the j-th capture time point of the PTZ camera at the i-th capture trajectory point, where j∈(1,M), and M represents the j-th capture time point within the set capture time period.

6. The deduplication method for continuous target capture and early warning using a PTZ camera according to claim 5, characterized in that, The image analysis parameters are expressed as follows: ,in, satisfy: in, Let H represent the image parameters of the PTZ camera at the i-th capture trajectory point, H represent the installation height of the PTZ camera, and (lon,lat) represent the latitude and longitude of the PTZ camera. This indicates the administrative region to which the PTZ camera belongs.

7. A deduplication system for continuous target capture and early warning using a PTZ camera, comprising a PTZ camera and a deduplication device, wherein the PTZ camera is used to capture images of a target monitoring area and transmits the captured images and their corresponding PTZ values ​​to the deduplication device; characterized in that, The deduplication device includes a position parameter acquisition module, an image parameter acquisition module, a parameter integration module, an early warning generation module, and an early warning deduplication module. The location parameter acquisition module is used to acquire the installation height, latitude and longitude, and administrative region of the PTZ camera, and generate location parameters. The image parameter acquisition module is used to acquire the captured images taken by the PTZ camera along the set capture trajectory and capture time, and generate image parameters by combining them with the PTZ value of the PTZ camera corresponding to the capture time. The parameter integration module is used to concatenate position parameters and image parameters to generate image analysis parameters; The warning generation module is used to detect illegal construction targets in the captured image using a target recognition algorithm, and to determine whether there are illegal construction targets in the captured image: If so, then based on the mapping relationship between the pixel coordinates of the PTZ camera and the world coordinates and planar coordinates, the world coordinates and planar coordinates of the illegal construction target in the captured image are obtained; combined with the image analysis parameters, an early warning message with warning information and warning image is generated. The warning information includes: warning image address Warning time Target plane coordinates Types of early warning targets Administrative regions where PTZ cameras belong and algorithm type ;in, This represents the i-th capture trajectory point of the PTZ camera, where i∈(1,N), and N represents the N capture trajectory points set. This indicates the j-th capture time point of the PTZ camera at the i-th capture trajectory point, where j∈(1,M), and M represents the j-th capture time point within the set capture time period; The early warning deduplication module is used to analyze early warnings using early warning judgment rules to determine whether an early warning is valid. If yes, then store the warning in the warning dataset; If not, then delete the warning; Specifically, it includes: S51A: Setting Timing Determination Rules The core verification parameters for setting the timing include: the administrative region to which the PTZ camera belongs. Types of early warning targets Warning time Algorithm type ; S52A: If the warning occurs at the time Greater than timing determination rules The early warning dataset is traversed based on the time-series core verification parameters: If old data corresponding to the warning exists, it means that the warning is a duplicate warning and should be deleted. If no old data corresponding to the warning exists, then a time series dataset that meets the conditions is obtained. ; Among them, the timing judgment rule This indicates the set threshold for the duration of the warning. S51B: Setting Space Determination Rules The core verification parameters for the space setting include: the administrative region to which the PTZ camera belongs. Types of early warning targets Target plane coordinates Algorithm type Old plane coordinates of the early warning target ; S52B: Calculates the planar coordinates of early warning targets using a spatial position calculation function. old plane coordinates of the early warning target distance If the distance Less than or equal to space determination rules Then, the early warning dataset is traversed based on the spatial core verification parameters: If old data corresponding to the warning exists, it means that the warning is a duplicate warning and should be deleted. If no old data corresponding to the warning exists, then a spatial dataset that meets the conditions is obtained. ; Among them, spatial determination rules This indicates the set warning distance threshold; S53: For time series datasets and spatial datasets Perform set operations to remove duplicates and obtain the intersection. : If the intersection If the value is empty, the warning is a valid warning, and the warning is stored in the warning dataset; If the intersection If the value is not empty, the warning is a duplicate warning and should be deleted.

8. An electronic device, characterized in that, The electronic device includes a processor and a memory, the memory storing programs or instructions that can run on the processor, the programs or instructions being executed by the processor to implement the steps of the method as described in any one of claims 1 to 6.

9. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the method as described in any one of claims 1 to 6.

Images