A natural resource digital law enforcement video image intelligent coordinate association system

Abstract

The present application relates to the technical field of image recognition, in particular to a video image intelligent coordinate correlation system in natural resource digital law enforcement, which comprises a law enforcement video acquisition module for generating patrol image space reference information, a pixel offset generation module for generating pixel offset information according to a central position, a line of sight angle calculation module for generating line of sight expression information in combination with attitude correction, a ground projection generation module for mapping the line of sight to generate a ground projection position, and a coordinate correlation generation module for comparing a graph spot contour to construct an intelligent correlation result of pixels and geographic coordinates. The present application integrates longitude and latitude and attitude to form an image space reference, establishes a target pixel horizontal and vertical offset relationship for collaborative expression, corrects the offset and corrects the line of sight direction in combination with a camera field of view, calculates the ground projection according to the device position and height, compares with illegal land occupation and mining contour to establish a correlation relationship, realizes stable mapping of video pixels and geographic coordinates, and improves the law enforcement image space positioning accuracy.

Description

Technical Field

[0001] This invention relates to the field of image recognition technology, and in particular to an intelligent coordinate association system for video images in digital law enforcement of natural resources. Background Technology

[0002] The field of image recognition technology includes digital image acquisition, image preprocessing, feature extraction, target detection, target classification and spatial information representation. It uses digital images as the main information carrier, acquires continuous or discrete image data through image acquisition devices, performs grayscale processing, edge extraction, texture feature extraction, target contour recognition and spatial location annotation on the images, and combines geographic information representation methods to describe the shape, position and spatial relationship of target objects in the image.

[0003] This technology is widely used in scenarios such as remote sensing image analysis, video surveillance analysis, natural resource monitoring, and geospatial information analysis. Its core content lies in expressing the correspondence between image information and geographical location data by associating image pixel features with spatial coordinate information, and forming a data system that can be used for spatial analysis and target positioning.

[0004] Among them, the intelligent coordinate association system for video images in digital law enforcement of natural resources refers to an image coordinate association processing system that associates and expresses video image information from law enforcement sites with geospatial coordinate information during the supervision and law enforcement of natural resources. The technical aspects involved include video image acquisition from law enforcement sites, video frame image extraction, contour recognition of ground objects in images, extraction of image pixel positions, acquisition of geographic coordinates, establishment of the correspondence between image pixel positions and geographic coordinates, and association and expression of law enforcement images with spatial location data. It generally acquires on-site video data through law enforcement recording equipment or fixed video acquisition equipment, extracts images frame by frame from the video, performs contour recognition of target ground objects in the images and extracts pixel position data, and simultaneously acquires the longitude and latitude coordinates of the corresponding shooting location. Based on the correspondence between image pixel coordinates and geographic coordinates, image target position data is established, thereby forming an association data system between video image information and geospatial coordinates.

[0005] Existing technologies typically use the video capture location as the basis for spatial information in video image representation. In continuous patrol or high-altitude evidence collection scenarios, the device's attitude and angle continuously change, and the shooting direction deflects with the flight trajectory. Different areas of the video image form significant positional differences in pixel space. However, the coordinates of a single capture location are insufficient to reflect the actual ground spatial orientation corresponding to different areas of the image. For example, when drones patrol mining areas or occupied areas, the camera's overhead angle and flight direction change, resulting in significant differences in the ground positions corresponding to the edge and center areas of the image. When image spatial representation still uses the capture location as a unified reference, the geographical location representation of the target area is prone to deviation, leading to a lack of stable mapping between the target area in the image and the actual spatial location. This affects the accuracy of spatial positioning of law enforcement images and the reliability of subsequent spatial comparison and analysis results. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of existing technologies by proposing an intelligent coordinate association system for video images in digital law enforcement of natural resources.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: an intelligent coordinate association system for video images in digital law enforcement of natural resources, the system comprising: The law enforcement video acquisition module uses drone patrol cameras to capture video footage of the patrol area, simultaneously read longitude, latitude, heading angle, pitch angle, and flight altitude, perform pixel analysis to identify the center position, record it in the image data record, and generate patrol image spatial reference information; The pixel offset generation module reads the target pixel coordinates based on the horizontal and vertical center pixel positions in the spatial reference information of the patrol image, performs a comparison with the center pixel position as a reference, determines the direction based on the results, determines the pixel offset position, and generates image pixel offset expression information. The line-of-sight angle calculation module, based on the heading angle and pitch angle in the patrol image spatial reference information, combined with the image pixel offset expression information, reads the camera's horizontal and vertical field of view, performs matching judgment, and performs correction direction in combination with the heading angle and pitch angle to determine the image pixel line-of-sight direction and generate image pixel line-of-sight expression information. The ground projection generation module determines the line-of-sight pointing position based on the image pixel line-of-sight representation information and the patrol image spatial reference information, combined with the longitude and latitude of the patrol equipment, performs spatial position association processing and ground position mapping processing, and generates image ground projection position representation information.

[0008] As a further aspect of the present invention, the system further includes: The coordinate association generation module reads the outline positions of illegal land occupation and mining activities image patches based on the image ground projection position expression information, compares the image ground projection positions with the outline positions of illegal land occupation and mining activities image patches respectively, establishes corresponding geographical location association relationships, constructs the correspondence between the target pixel positions of the video image and geographical coordinates, and generates intelligent coordinate association results for digital law enforcement video images of natural resources.

[0009] As a further aspect of the present invention, the law enforcement video acquisition module includes a patrol image acquisition submodule, a pixel matrix parsing submodule, and a spatial reference recording submodule; The patrol video acquisition submodule acquires video footage of the patrol area from the drone patrol camera, the longitude and latitude from the satellite positioning unit, the heading angle and pitch angle from the attitude sensor, and the flight altitude from the altimeter. It performs time-sequence processing on the video frame sequence and the longitude, latitude, heading angle, pitch angle, and flight altitude, performs a correspondence judgment between the video frame sequence and the equipment attitude record, and completes the unified processing of the video footage and the patrol equipment attitude record to generate the patrol video frame sequence. The pixel matrix parsing submodule reads the pixel matrix of the video frame based on the patrol video frame sequence, performs horizontal and vertical pixel sequence scanning processing of the pixel matrix, identifies the center pixel position of the video frame, performs corresponding arrangement of the center pixel position and the video frame pixel position, completes the identification of video frame pixel position and pixel coordinate arrangement, and generates the video center pixel coordinates. The spatial reference recording submodule, based on the coordinates of the video center pixel, calls longitude, latitude, heading angle, pitch angle and flight altitude to perform spatial position correspondence organization, performs horizontal center pixel position and vertical center pixel position matching judgment with the spatial position of the inspection equipment, completes the unified organization of equipment spatial position recording and image pixel position, establishes the correspondence record between image pixel position and inspection equipment spatial position, and generates inspection image spatial reference information.

[0010] As a further embodiment of the present invention, the pixel offset generation module includes a center coordinate calling submodule, a pixel position comparison submodule, and a direction determination processing submodule; The center coordinate calling submodule reads the horizontal center pixel position and the vertical center pixel position according to the patrol image spatial reference information, reads the horizontal pixel coordinates and vertical pixel coordinates corresponding to the target position in the video frame, organizes the corresponding records of the horizontal center pixel position and the horizontal pixel coordinates, organizes the corresponding records of the vertical center pixel position and the vertical pixel coordinates, completes the unified recording of the target pixel position and the center pixel position in the video frame, and obtains the target pixel coordinate record. The pixel position comparison submodule calls the target pixel coordinate record, reads the horizontal and vertical pixel coordinates, calls the horizontal center pixel position to perform horizontal pixel position comparison processing, calls the vertical center pixel position to perform vertical pixel position comparison processing, and organizes the horizontal and vertical pixel position comparison results to obtain the pixel position comparison record. The orientation determination processing submodule, based on the pixel position comparison record, reads the horizontal pixel position comparison result and the vertical pixel position comparison result, performs horizontal pixel position orientation determination processing, performs vertical pixel position orientation determination processing, sorts the horizontal pixel orientation determination results and the vertical pixel orientation determination results, forms the pixel offset position corresponding to the center pixel position of the video image, and generates image pixel offset expression information.

[0011] As a further embodiment of the present invention, the line-of-sight angle calculation module includes a field-of-sight range reading submodule, an offset matching determination submodule, and a line-of-sight direction correction submodule; The field of view reading submodule acquires the camera's horizontal and vertical field of view, calls the image pixel offset expression information, reads the horizontal and vertical pixel offset directions, calls the patrol image spatial reference information to read the heading and pitch angles, organizes the corresponding records of the horizontal and vertical pixel offset directions, horizontal and vertical field of view, completes the unified organization of image pixel offset directions and camera field of view, and generates a field of view matching preparation record. The offset matching determination submodule reads the horizontal pixel offset direction and the horizontal field of view based on the field of view matching preparation record, performs horizontal pixel offset direction matching determination processing, reads the vertical pixel offset direction and the vertical field of view to perform vertical pixel offset direction matching determination processing, organizes the horizontal pixel offset direction matching determination record and the vertical pixel offset direction matching determination record, and generates a pixel offset matching record. The line-of-sight direction correction submodule, based on the pixel offset matching record, calls the heading angle and pitch angle to carry out correction and organization of the lateral pixel offset direction and the longitudinal pixel offset direction, completes the unified organization of the pixel offset direction and the attitude direction of the inspection equipment, forms the observation direction record corresponding to the image pixel position, and generates the image pixel line-of-sight expression information.

[0012] As a further embodiment of the present invention, the ground projection generation module includes a line-of-sight pointing analysis submodule, a spatial position association submodule, and a ground mapping processing submodule; The line-of-sight analysis submodule calls the line-of-sight expression information of the image pixels, reads the line-of-sight direction of the image pixels, calls the spatial reference information of the patrol image to read the longitude, latitude and flight altitude of the patrol equipment, organizes the corresponding information of the line-of-sight direction of the image pixels and the spatial position of the patrol equipment, and performs line-of-sight direction recognition processing of the image pixels in combination with the longitude and latitude of the patrol equipment, and completes the unified organization of the line-of-sight direction of the image pixels and the position of the patrol equipment to generate the line-of-sight result; The spatial location association submodule reads the image pixel line-of-sight pointing position based on the image line-of-sight pointing result, calls the longitude and latitude of the patrol equipment to perform spatial location correspondence organization, organizes the image pixel line-of-sight pointing position and the longitude and latitude of the patrol equipment to form the spatial positioning result of the image pixel line-of-sight pointing position, and generates the line-of-sight pointing spatial position; The ground mapping processing submodule, based on the line-of-sight spatial position, calls the flight altitude to carry out ground position mapping and sorting, sorts out the corresponding information of the line-of-sight position of image pixels and the ground spatial position, forms a unified expression of the line-of-sight position of image pixels and the ground spatial position, and generates the image ground projection position expression information.

[0013] As a further embodiment of the present invention, the coordinate association generation module includes a projection position reading submodule, a contour pixel comparison submodule, and a coordinate association construction submodule; The projection position reading submodule reads the image ground projection position based on the image ground projection position expression information, reads the image ground projection position, reads the outline pixel position of the illegal land occupation boundary patch and the outline pixel position of the mining activity image patch, organizes the corresponding information of the image ground projection position and the outline pixel position of the illegal land occupation boundary patch, organizes the corresponding information of the image ground projection position and the outline pixel position of the mining activity image patch, and completes the unified organization of the image ground projection position and the target outline pixel position to obtain the image projection outline position set. The contour pixel comparison submodule calls the image projection contour position set, reads the image ground projection position and the contour pixel position of the illegal land occupation boundary patch, performs comparison processing between the image ground projection position and the contour pixel position of the illegal land occupation boundary patch, reads the contour pixel position of the mining activity image patch and performs comparison processing between the image ground projection position and the contour pixel position of the mining activity image patch, sorts out the two types of contour pixel position comparison results, and obtains the contour pixel position comparison result. The coordinate association construction submodule, based on the contour pixel position comparison results, reads the corresponding positions of the contour pixel positions of the illegal land occupation boundary patches and the contour pixel positions of the mining activity image patches, organizes the spatial correspondence information between the image ground projection position and the target contour pixel position, forms the organized result of the correspondence between the target pixel position of the video screen and the geographic coordinates, establishes the expression of the correspondence between the target pixel position of the video screen and the geographic coordinates, and generates the intelligent coordinate association result of the digital law enforcement video image of natural resources.

[0014] As a further aspect of the present invention, the information on the correspondence between the ground projection position of the image and the outline pixel position of the illegally occupied land boundary patch is specifically as follows: The illegal land occupation boundary patches are arranged into closed contours according to connectivity and divided into continuous contour segments. For each image ground projection position, the corresponding contour segment is determined and the contour segment identifier and the position of adjacent contour pixels within the contour segment are recorded. The specific process of comparing the image ground projection position with the outline pixel position of the illegally occupied land boundary patch is as follows: The comparison range is defined by the contour segment identifier. Within the contour segment, point-by-point matching is performed in the order of adjacent contour pixel positions to generate matching pairs. The specific steps for organizing the comparison results of the two types of contour pixel positions are as follows: Conflict resolution is performed on the matching pairs. When the ground projection position of the same image corresponds to the outline pixel position of multiple illegal land occupation boundary patches, the matching pairs with the same outline segment order are retained. When the outline pixel position of the same illegal land occupation boundary patch corresponds to multiple ground projection positions of the image, the matching pairs with continuous adjacent relationships within the outline segment are retained.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by synchronously integrating the device's longitude, latitude, heading angle, pitch angle, and flight altitude, and identifying the position of the video's center pixel to form an image spatial reference, a horizontal and vertical pixel offset relationship is established between the target pixel coordinates and the center pixel, and the pixel direction distribution is determined, so that the pixel position and the device's attitude information form a coordinated expression. Combining the camera's horizontal and vertical fields of view, the pixel offset direction is matched and corrected, and the line-of-sight direction is corrected based on the attitude information, so that the spatial line-of-sight direction corresponding to the image pixel is clearly expressed. Based on the device's spatial position and flight altitude, the line-of-sight direction is calculated by ground projection, so that the video pixel is definitely expressed on the ground spatial position. Furthermore, a spatial position comparison is performed with the contour of the illegal land occupation boundary and the contour of the mining activity image to establish a correlation, thereby realizing a stable spatial mapping relationship between the video target pixel and the geographic coordinates and improving the spatial positioning accuracy and spatial correlation expression capability of law enforcement images. Attached Figure Description

[0016] Figure 1 This is a system flowchart of the present invention; Figure 2 This is a flowchart illustrating the acquisition process of the law enforcement video acquisition module of the present invention. Figure 3 This is a flowchart illustrating the acquisition process of the pixel offset generation module of the present invention. Figure 4 This is a flowchart illustrating the acquisition process of the line-of-sight angle calculation module of the present invention. Figure 5 This is a flowchart illustrating the acquisition process of the ground projection generation module of the present invention. Figure 6 This is a flowchart of the coordinate association generation module of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0018] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0019] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0020] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0021] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0022] Please see Figure 1 This invention provides a technical solution: an intelligent coordinate association system for video images in digital law enforcement of natural resources, the system comprising: The law enforcement video acquisition module continuously collects video footage of the patrol area through the drone patrol camera, simultaneously reads the longitude and latitude output by the satellite positioning unit, reads the heading angle and pitch angle output by the attitude sensor, and the flight altitude recorded by the altimeter. It performs pixel matrix analysis on the video footage, organizes the pixel positions of the video footage and identifies the center position of the video footage, and records the longitude, latitude, heading angle, pitch angle, flight altitude, and the horizontal and vertical center pixel positions into the image data record, constructs the patrol image spatial reference relationship, and generates patrol image spatial reference information. The pixel offset generation module reads the horizontal and vertical pixel coordinates corresponding to the target position in the video frame based on the horizontal and vertical center pixel positions in the spatial reference information of the patrol image. It performs horizontal pixel position comparison with the horizontal center pixel position as a reference and vertical pixel position comparison with the vertical center pixel position as a reference. It performs direction determination by combining the horizontal and vertical pixel position comparison results and determines the pixel offset position of the target pixel position relative to the center pixel position of the video frame based on the direction determination result, and generates image pixel offset expression information. The line-of-sight angle calculation module, based on the heading and pitch angles in the spatial reference information of the patrol image, and combined with the lateral and vertical pixel offset directions in the image pixel offset expression information, reads the camera's lateral and vertical field of view. Using the lateral field of view as a reference, it performs lateral pixel offset direction matching and determination; using the vertical field of view as a reference, it performs vertical pixel offset direction matching and determination. It then combines the lateral and vertical pixel offset direction matching and determination results with the heading and pitch angles to perform direction correction. Based on the direction correction results, it determines the image pixel line-of-sight direction and generates image pixel line-of-sight expression information. The ground projection generation module calls the image pixel line-of-sight direction from the image pixel line-of-sight representation information and the longitude, latitude, and flight altitude of the patrol equipment from the patrol image spatial reference information. It combines the longitude and latitude of the patrol equipment to determine the line-of-sight pointing position corresponding to the image pixel line-of-sight direction, performs spatial position association processing between the image pixel line-of-sight pointing position and the longitude and latitude of the patrol equipment, and performs ground position mapping processing on the image pixel line-of-sight pointing position in combination with the flight altitude to form a correspondence between the image pixel line-of-sight pointing position and the ground spatial position, and generates the image ground projection position representation information. The coordinate association generation module reads the corresponding pixel contour positions of the illegal land occupation boundary patches and mining activity image patches based on the image ground projection position expression information. It then compares the image ground projection positions with the contour pixel positions of the illegal land occupation boundary patches and mining activity image patches respectively, establishes the geographical location association relationship between the contour pixel positions of the illegal land occupation boundary patches and the contour pixel positions of the mining activity image patches, constructs the correspondence between the target pixel positions in the video image and the geographical coordinates, and generates intelligent coordinate association results for digital law enforcement video images of natural resources.

[0023] Please see Figure 2 The law enforcement video acquisition module includes a patrol image acquisition submodule, a pixel matrix parsing submodule, and a spatial reference recording submodule; The patrol video acquisition submodule acquires video footage of the patrol area from the drone patrol camera, the longitude and latitude from the satellite positioning unit, the heading angle and pitch angle from the attitude sensor, and the flight altitude from the altimeter. It performs time-sequence processing on the video frame sequence and the longitude, latitude, heading angle, pitch angle, and flight altitude, performs a correspondence judgment between the video frame sequence and the equipment attitude record, and completes the unified processing of the video footage and the patrol equipment attitude record to generate the patrol video frame sequence. The drone uses a CMOS image sensor to read real-time video streams captured in the area where illegal land occupation is being patrolled, and a multi-constellation GNSS receiver integrated into the flight control system to obtain the current longitude value. The latitude value is 116.3971°. The current heading angle is 39.9087°, obtained from the output of the MEMS three-axis gyroscope and accelerometer. 45.0° and pitch angle The current relative flight altitude is -30.0°, extracted and recorded using a laser altimeter or barometer. The distance is 120.0 meters. The device clock pulse sequence number corresponding to each frame of the video image is recorded as... The sensor data acquisition time stamps corresponding to each parameter, including longitude, latitude, heading angle, pitch angle, and flight altitude, are recorded as follows: The video frame sequence is time-axis aligned with each pose parameter, and the absolute deviation of their time labels is calculated. The time deviation is defined as follows: 0-50 milliseconds is considered high synchronization; 51-100 milliseconds is considered medium synchronization; and above 101 milliseconds is considered low synchronization and is deemed an invalid frame that needs to be discarded or resampled. A synchronization offset baseline value is also set. The baseline value is 30 milliseconds. This value is determined based on the reciprocal of the least common multiple of the sensor data sampling frequency and the image frame rate. For example, when the sampling rate is 20Hz and the frame rate is 25fps, the reciprocal difference is 10 milliseconds. Combining this with the inherent 20-millisecond delay compensation from the hardware transmission link, we arrive at 30 milliseconds. Time-series data retrieval is then performed, and the selected data is... Less than or equal to The parameter group is matched with the corresponding video frame. When the time difference between the time sequence number and the time label of parameters such as latitude and longitude of the 500th frame image is found to be 15 milliseconds, it is determined that the frame and the attitude record belong to the same spatial state snapshot. The successfully matched longitude 116.3971°, latitude 39.9087°, heading 45.0°, pitch -30.0° and altitude 120.0 meters are sequentially pushed into the attribute metadata of the corresponding video frame, completing the one-to-one mapping and association of each frame with the spatial attitude data of the device at the moment of shooting, and generating the patrol video frame sequence.

[0024] The pixel matrix parsing submodule reads the pixel matrix of the video frame based on the patrol video frame sequence, performs horizontal and vertical pixel sequence scanning processing, identifies the center pixel position of the video frame, performs corresponding arrangement of the center pixel position and the video frame pixel position, completes the video frame pixel position recognition and pixel coordinate arrangement, and generates the video center pixel coordinates. Based on the patrol video frame sequence, the original RGB pixel matrix of a single frame image is retrieved to obtain the total horizontal pixel count of the video frame. 3840 and total vertical pixels Given 2160, establish a point with the top left corner as the origin. Using a two-dimensional pixel coordinate system, perform a horizontal scan of pixel grayscale values ​​from column 1 to column 3840, and a vertical scan of pixel grayscale values ​​from row 1 to row 2160 to extract the row and column index position information of each pixel in the matrix. Calculate the horizontal center pixel index value of the video frame Calculate the vertical center pixel index value Set the center offset determination threshold The threshold is 0.5 pixels. If the calculated center point value obtained from the scan deviates from the theoretical value by more than this threshold, a recentering calibration operation is performed, and the coordinates of all scanned pixels are adjusted accordingly. With the calculated Perform logical index allocation one by one, defining the four-quadrant assignment relationship of each pixel relative to the center point within the coordinate system, such as pixel... The pixel is marked as being located to the lower right of the center point with a horizontal deviation of 80 pixels and a vertical deviation of 120 pixels. The matrix arrangement and reconstruction of all pixels in memory are performed, and the normalized coordinate value corresponding to each pixel index is stored to generate the coordinates of the video center pixel.

[0025] The spatial reference recording submodule, based on the coordinates of the video center pixel, calls longitude, latitude, heading angle, pitch angle and flight altitude to perform spatial position correspondence organization, performs horizontal center pixel position and vertical center pixel position matching judgment with the spatial position of the inspection equipment, completes the unified organization of equipment spatial position recording and image pixel position, establishes the correspondence record between image pixel position and inspection equipment spatial position, and generates inspection image spatial reference information; Based on the center pixel coordinates of the video, call the function in the video center pixel coordinates. The frame's longitude (116.3971°), latitude (39.9087°), heading (45.0°), pitch (-30.0°), and altitude (120.0 meters) are retrieved, and a spatial reference correlation weighting coefficient is set. The coefficient is set to 1.0. This coefficient is based on the physical optical axis alignment of the image sensor's photosensitive unit and the lens. The optical axis deviation is measured using a high-precision theodolite and collimator in an indoor precision optical calibration device, and the value is taken between 0.95 and 1.05. Here, 1.0 indicates no physical assembly eccentricity. The horizontal center pixel position 1920 and the vertical center pixel position 1080 are merged and packaged with the spatial position of the inspection equipment. The heading angle of 45.0° is used as the initial value for the rotation of the pixel matrix in the geographical orientation, the pitch angle of -30.0° is used as the tilt parameter of the image plane relative to the horizontal plane, and the flight altitude of 120.0 meters is used as the reference operator for the scaling calculation of pixel size and actual ground size. The image spatial reference matrix is ​​calculated. The matrix and image file header are written in binary format, and the structured encapsulation of each set of dynamic attitude records and static center pixel coordinates is completed. This establishes a multi-dimensional data chain covering geographic coordinates, flight attitude, and the center point of the image, generating spatial reference information for the patrol image.

[0026] Please see Figure 3 The pixel offset generation module includes a center coordinate calling submodule, a pixel position comparison submodule, and a direction determination processing submodule; The center coordinate calling submodule reads the horizontal center pixel position and the vertical center pixel position based on the spatial reference information of the patrol image, reads the horizontal pixel coordinates and vertical pixel coordinates corresponding to the target position in the video frame, organizes the corresponding records of the horizontal center pixel position and the horizontal pixel coordinates, organizes the corresponding records of the vertical center pixel position and the vertical pixel coordinates, completes the unified recording of the target pixel position and the center pixel position in the video frame, and obtains the target pixel coordinate record. Based on the spatial reference information of the patrol images, the horizontal center pixel position index 1920 and the vertical center pixel position index 1080 stored in the spatial reference information of the patrol images are retrieved. A deep learning convolutional neural network target detection algorithm is used to obtain the horizontal pixel coordinates corresponding to the center of the suspected illegal land occupation bulldozer identified in the video. 2500 and vertical pixel coordinates Set the target extraction confidence threshold to 1500. The threshold is 0.85. This threshold was determined by statistically analyzing the dispersion of target edge sharpness and background contrast in 500 historical natural resource law enforcement sample images. When the calculated target feature response value of 0.92 is greater than 0.85, the values ​​of the horizontal center pixel position 1920 and the horizontal pixel coordinate 2500 are paired, and the values ​​of the vertical center pixel position 1080 and the vertical pixel coordinate 1500 are paired. The above two sets of coordinate data are pushed onto a temporary data stack, and the Euclidean pixel distance between the target point and the center point is calculated. For each pixel, the distance to the target pixel is determined to be within the effective recognition range of the image. This range is defined as between 0 and 2200 pixels (i.e., not exceeding half the length of the image diagonal). The target pixel coordinates are recorded by merging and encapsulating the target pixel index value and the center reference position index value.

[0027] The pixel position comparison submodule calls the target pixel coordinate record, reads the horizontal and vertical pixel coordinates, calls the horizontal center pixel position to perform horizontal pixel position comparison processing, calls the vertical center pixel position to perform vertical pixel position comparison processing, and organizes the horizontal and vertical pixel position comparison results to obtain the pixel position comparison record. Call the target pixel coordinate record stored in The reference horizontal center pixel position 1920 and the reference vertical center pixel position 1080 are retrieved. A subtraction operation is performed between the horizontal pixel coordinate 2500 and the horizontal center pixel position 1920 to obtain the horizontal pixel difference. The value is +580. A subtraction operation is performed between the vertical pixel coordinate of 1500 and the vertical center pixel position of 1080 to obtain the vertical pixel difference. Set the offset significance judgment weight coefficient to +420. The coefficient is set to 1.2. This coefficient is set to the radial stretch rate of the lens distortion correction parameter. Using the Zhang Zhengyou calibration method, the deformation rate of 0.2 in the 5% area at the edge of the image is compensated and summed to obtain 1.2. This coefficient is called and the difference is multiplied to obtain the corrected horizontal comparison result of 696 pixels and the vertical comparison result of 504 pixels. The variation range of the comparison result is set. When the absolute value of the difference is within 0 to 10 pixels, it is judged as a slight shift. When it is within 11 to 500 pixels, it is judged as a significant shift. When it is above 501 pixels, it is judged as a severe shift and Kalman filtering smoothing process needs to be started. Here, 580 and 420 correspond to severe shift and significant shift, respectively. The above numerical difference and shift degree labels are sorted to obtain the pixel position comparison record.

[0028] The orientation determination processing submodule, based on the pixel position comparison record, reads the horizontal pixel position comparison result and the vertical pixel position comparison result, performs horizontal pixel position orientation determination processing, performs vertical pixel position orientation determination processing, sorts the horizontal pixel orientation determination results and vertical pixel orientation determination results, forms the pixel offset position corresponding to the center pixel position of the video image, and generates image pixel offset expression information. Based on the horizontal difference of +580 and the vertical difference of +420 stored in the pixel position comparison record, the horizontal pixel position direction determination process is performed. +580 is compared with the value 0. According to the logic that the value is greater than 0, the horizontal offset direction is determined to be to the right. The vertical pixel position direction determination process is then performed. +420 is compared with the value 0. According to the logic that the value is greater than 0, the vertical offset direction is determined to be downwards. The quadrant determination logic value is then set. The (right, bottom) combination corresponds to the fourth quadrant of the image coordinate system, and the (left, top) combination corresponds to the first quadrant. By using a logic switch to select the operator, the target's location is determined to be in the lower right corner of the image. The current pixel clock step data is read, and the horizontal pixel offset of 580 and the vertical pixel offset of 420 are concatenated with the location marker "lower right." An offset validity verification threshold is then set. The threshold is 1.5 pixels. This threshold is determined with reference to the thermal noise crosstalk voltage of adjacent pixels in the camera's CMOS photosensitive array. When the actual offset 580 is greater than 1.5, the direction data is considered valid. If it is less than 1.5, it is considered background noise and no direction update is performed. The matrix mapping between the direction determination result and the pixel difference is performed to form the pixel offset position corresponding to the center pixel position of the video image, and the image pixel offset expression information is generated.

[0029] Please see Figure 4 The line-of-sight angle calculation module includes a field-of-sight range reading submodule, an offset matching determination submodule, and a line-of-sight direction correction submodule. The field of view reading submodule acquires the camera's horizontal and vertical field of view, calls up image pixel offset expression information, reads the horizontal and vertical pixel offset directions, calls up the patrol image spatial reference information to read the heading and pitch angles, organizes the corresponding records of horizontal pixel offset direction, vertical pixel offset direction, horizontal field of view, and vertical field of view, completes the unified organization of image pixel offset direction and camera field of view, and generates a field of view matching preparation record; Extracting the inherent optical parameters of the drone camera: lateral field of view. With a longitudinal field of view of 84° The angle is 62°. The horizontal pixel offset value +580 and the vertical pixel offset value +420 stored in the image pixel offset representation information are retrieved. Simultaneously, the UAV's heading angle of 45.0° and pitch angle of -30.0° at this moment are read from the spatial reference information of the survey image. The field-of-view sampling weighting coefficient is then set. The coefficient is set to 0.98. This coefficient is determined by referencing the light attenuation coefficient at the edge of the lens. The illuminance ratio from the center to the edge is measured using an integrating sphere photometric meter and is taken between 0.9 and 1.0. Here, 0.98 is chosen to correct the field of view resolution accuracy. The horizontal pixel offset value +580 is matched with the horizontal field of view angle of 84°, and the vertical pixel offset value +420 is matched with the vertical field of view angle of 62°. The offset direction attribute "right" is bound to the memory address of the horizontal field of view data, and the offset direction attribute "down" is bound to the memory address of the vertical field of view data. The yaw angle of 45.0° and the pitch angle of -30.0° are used as global reference benchmarks for stacking. This completes the structured alignment of each visual deviation parameter with the physical optical range of the camera and generates a field of view matching preparation record.

[0030] The offset matching determination submodule reads the horizontal pixel offset direction and the horizontal field of view based on the field of view matching preparation record, performs horizontal pixel offset direction matching determination processing, reads the vertical pixel offset direction and the vertical field of view to perform vertical pixel offset direction matching determination processing, organizes the horizontal pixel offset direction matching determination record and the vertical pixel offset direction matching determination record, and generates pixel offset matching record. Based on the field of view matching preparation record, the horizontal pixel offset +580 and the horizontal field of view angle 84° stored in the field of view matching preparation record are retrieved. The total pixel width of a single frame is 3840, and the horizontal angular displacement (radians converted to degrees) corresponding to a unit pixel is calculated. °, perform a positive and negative interval matching judgment between the angular displacement and the horizontal field of view range of 84°. When 12.68° is within the range of -42° to +42°, it is determined to be a valid offset. Call the vertical pixel offset +420 and the vertical field of view angle of 62°, and retrieve the total pixel height of the single frame of 2160. Calculate the vertical angular displacement corresponding to a unit pixel. The system performs a positive and negative interval matching judgment between the angular displacement and the longitudinal field of view of 62°. When 12.05° is within the range of -31° to +31°, it is determined to be a valid offset. If the calculated result exceeds this range, it is determined that the target has exceeded the frame boundary and an alarm is triggered. The offset matching sensitivity threshold is set. The threshold is 0.01°, which is determined with reference to the minimum division value of the gimbal encoder. When the calculated angular displacement value of 12.68° is greater than 0.01°, the offset direction is determined to be successfully matched. The logical record of the horizontal matching result "rotate to the right by 12.68°" and the vertical matching result "rotate downward by 12.05°" is executed to generate a pixel offset matching record.

[0031] The line-of-sight direction correction submodule, based on pixel offset matching records, calls the heading angle and pitch angle to carry out correction and organization of the lateral and longitudinal pixel offset directions, completes the unified organization of pixel offset directions and the attitude direction of the inspection equipment, forms a record of the observation direction corresponding to the image pixel position, and generates image pixel line-of-sight expression information. Based on pixel offset matching records, the heading angle of 45.0° and the pitch angle of -30.0° are used to perform an additive correction operation on the heading angle. The initial heading angle of 45.0° is then summed with the horizontal offset angle of 12.68° to obtain the actual azimuth angle of the target. °, perform a subtraction correction operation on the pitch angle, summing the initial pitch of -30.0° (negative for downward) with the vertical offset angle of 12.05° to obtain the actual downward view of the target. °, Set attitude correction compensation coefficient The coefficient is set to 1.02. This coefficient is set to correct for long-distance line-of-sight deflection based on the Earth's curvature. Within a 2-kilometer patrol radius, it is determined according to the standard atmospheric refraction model. A vector synthesis is performed between the actual azimuth angle of 57.68° and the actual downward angle of -42.05°. The observation direction deviation level is set: 0° to 5° is considered slight deviation, 6° to 15° is significant deviation, and above 16° is extreme deviation. Here, 12.68° is determined to be significant deviation. The corrected angle value is then compared with the target pixel index. The final address association mapping forms a record of the observation direction corresponding to the image pixel position, generating image pixel line-of-sight expression information.

[0032] Please see Figure 5 The ground projection generation module includes a line-of-sight analysis submodule, a spatial location association submodule, and a ground mapping processing submodule. The line-of-sight analysis submodule calls the line-of-sight representation information of image pixels, reads the line-of-sight direction of image pixels, calls the spatial reference information of the patrol image to read the longitude, latitude and flight altitude of the patrol equipment, organizes the corresponding information of the line-of-sight direction of image pixels and the spatial position of the patrol equipment, and performs line-of-sight direction recognition processing of image pixels in combination with the longitude and latitude of the patrol equipment, and completes the unified organization of the line-of-sight direction of image pixels and the location of the patrol equipment to generate the line-of-sight result; Using the azimuth angle of 57.68° and the downward angle of -42.05° generated from the image pixel line-of-sight representation information, the longitude of the patrol equipment (116.3971°), latitude (39.9087°), and flight altitude (120.0 meters) recorded by the differential positioning terminal (RTK) in the patrol image spatial reference information are extracted, and the spatial line-of-sight resolution step size coefficient is set. The coefficient is set to 0.0001. This coefficient is determined by the resolution accuracy of the satellite positioning differential correction. It is obtained by taking values ​​between 0.00005 and 0.00015 based on the standard deviation of the carrier phase observations. The projection vector of the azimuth angle 57.68° in the north-northeast direction is extracted. The unit pointing vector decomposition of the downward angle of -42.05° in the three-dimensional Euclidean coordinate system is performed. The steepness judgment interval is set: angles between -90° and -60° are judged as near-vertical angles; between -59° and -30° as large tilt angles; and between -29° and 0° as level angles. Here, -42.05° is judged as a large tilt angle. The projection component of the line of sight on the horizontal plane is extracted by calling the cosine function calculation. The vertical height component is extracted by calling the sine function. The projected components are integrated with the polar coordinates of the azimuth angle of 57.68°. The starting coordinates of the UAV are used as the vector starting point to complete the initial topological alignment of the image pixel line of sight with the current latitude and longitude of the device, and generate the image line of sight pointing result.

[0033] The spatial location association submodule reads the image pixel line-of-sight pointing position based on the image line-of-sight pointing result, calls the longitude and latitude of the patrol equipment to perform spatial location correspondence processing, organizes the image pixel line-of-sight pointing position and the longitude and latitude of the patrol equipment to form the spatial positioning result of the image pixel line-of-sight pointing position, and generates the line-of-sight pointing spatial location. Based on the image line-of-sight results, the unit directional derivative of the image pixel line-of-sight direction is read, and the spatial location of the inspection equipment (longitude 116.3971°, latitude 39.9087°) is retrieved for spatial correspondence analysis. A geodetic coordinate transformation benchmark value is then set. The reference value is 6,378,137.0 meters. This reference value is determined with reference to the semi-major axis radius of the WGS84 ellipsoid. The distance mapping between the horizontal projection length of the line of sight and the intersection point with the ground is performed using the spherical trigonometric sine theorem. The longitude deviation of the line of sight on the Earth's tangent plane is then calculated. Deviation value of latitude Set location association weight The weight is 0.998, determined by the ratio of the difference between the average altitude of the current patrol area and the height of the reference surface. Within the altitude range of 50 meters to 500 meters, 0.998 is obtained through linear interpolation. Linear weighted calculation of the pointing vector in the longitude and latitude directions is performed. The sine component corresponding to the pointing azimuth of 57.68° is accumulated with the longitude coordinate, and the cosine component corresponding to the pointing azimuth of 57.68° is accumulated with the latitude coordinate. The spatial positioning accuracy level is set: high precision positioning is within 0.5 meters, standard positioning is between 0.6 and 2.0 meters, and low precision positioning is above 2.1 meters. If the positioning level is low precision, multi-frame data weighted smoothing correction is triggered to generate the spatial position of the line of sight pointing.

[0034] The ground mapping processing submodule, based on the line-of-sight spatial position, calls the flight altitude to carry out ground position mapping and sorting, sorts out the corresponding information of the line-of-sight position of image pixels and the ground spatial position, forms a unified expression of the line-of-sight position of image pixels and the ground spatial position, and generates the image ground projection position expression information. Based on the horizontal azimuth mapping point determined in the spatial location of the line of sight, the ground position mapping is processed using the flight altitude of 120.0 meters. The ground projection distance is then calculated using the trigonometric ratio between the altitude of 120.0 meters and the downward angle of 42.05°. Meters, with the unit of measurement in meters, conform to engineering logic; a projection correction threshold is set. The threshold value is 0.25. This threshold is determined by referring to the terrain slope coefficient between the UAV take-off and landing point and the patrol target point. The difference ratio between the highest and lowest points of the digital elevation model (DEM) along the patrol path is read and set between 0.1 and 0.5. The actual geographical location coordinates at a projection distance of 133.02 meters and an azimuth angle of 57.68° are reconstructed. The longitude of the ground projection point is calculated to be 116.3983°, and the latitude is calculated to be 39.9093°. The longitude and latitude values ​​are then compared with the image pixel index. The final address is stored, the ground projection coverage area is set, the diameter of the projection spot formed by the target pixel on the ground is recorded as 1.5 meters, the latitude and longitude data of the projection point are bound to the index of the original video frame number 500, the physical association between the image pixels and the ground coordinates is completed, and the image ground projection position expression information is generated.

[0035] Please see Figure 6 The coordinate association generation module includes a projection position reading submodule, a contour pixel comparison submodule, and a coordinate association construction submodule; The projection position reading submodule reads the image ground projection position based on the image ground projection position information, reads the image ground projection position, reads the outline pixel position of the illegal land occupation boundary patch and the outline pixel position of the mining activity image patch, organizes the corresponding information of the image ground projection position and the outline pixel position of the illegal land occupation boundary patch, and completes the unified organization of the image ground projection position and the target outline pixel position to obtain the image projection outline position set. The specific information regarding the correspondence between the ground projection position of the image and the pixel position of the illegal land occupation boundary patch is as follows: The illegal land occupation boundary patches are arranged into closed contours according to connectivity and divided into continuous contour segments. For each image ground projection position, the corresponding contour segment is determined and the contour segment identifier and the position of adjacent contour pixels within the contour segment are recorded. Based on the image's ground projection location information, the longitude (116.3983°) and latitude (39.9093°) of the image's ground projection are read. Then, the set of pixel locations of illegal land occupation boundary patches and the set of pixel locations of mining activity image patches, extracted using a hierarchical classification algorithm from a pre-set natural resource monitoring database, are retrieved. A spatial topology tolerance threshold is then set. The threshold is set at 0.00001 degrees (approximately 1.1 meters). This threshold is determined by referring to the coordinate accuracy level of the geographic information vector map features. A Euclidean distance scan is performed between the projected location and the pixel sequence of the database map feature outlines. Using an eight-neighbor search technique, the outline pixels of the illegally occupied land boundary map features are arranged into closed polygons according to connectivity. The polygon boundaries are then divided into continuous outline segments of 20 pixels in length. Calculate the contour segment from the projection point (116.3983, 39.9093). The perpendicular distance to the normal, when the distance value is less than When the contour segment to which it belongs is identified as ID-005, the coordinates of the adjacent contour pixels within the contour segment (116.3982, 39.9092) and (116.3984, 39.9094) are recorded. The structured memory pool allocation of the image ground projection position and the contour pixel positions of multiple types of targets is completed, and the image projection contour position set is obtained.

[0036] The contour pixel comparison submodule calls the image projection contour position set, reads the image ground projection position and the contour pixel position of the illegal land occupation boundary patch, performs comparison processing between the image ground projection position and the contour pixel position of the illegal land occupation boundary patch, reads the contour pixel position of the mining activity image patch and performs comparison processing between the image ground projection position and the contour pixel position of the mining activity image patch, sorts out the two types of contour pixel position comparison results, and obtains the contour pixel position comparison result. The specific process of comparing the image ground projection location with the pixel location of the illegal land occupation boundary patch is as follows: The comparison range is defined by the contour segment identifier. Within the contour segment, point-by-point matching is performed in the order of adjacent contour pixel positions to generate matching pairs. The specific results of comparing the pixel positions of the two types of contours are as follows: Conflict resolution is performed on the matching pairs. When the conflict resolution results in multiple illegal land occupation boundary patch outline pixel positions corresponding to the same image ground projection position, matching pairs with consistent outline segment order are retained. When the same illegal land occupation boundary patch outline pixel position corresponds to multiple image ground projection positions, matching pairs with continuous adjacent relationships within the outline segment are retained. The coordinates (116.3983, 39.9093) stored in the image projection contour position set are called. All discrete pixel coordinates of contour segment ID-005 within the limited range are retrieved. The difference operation between the projection position coordinates and the contour pixel sequence is performed. Point-by-point matching is performed within the contour segment in the order of adjacent contour pixel positions. The pixel pair with the smallest distance difference is selected to generate the initial matching pair. The conflict judgment logic interval is set. When the same projection point corresponds to multiple contour pixels, the distance difference between 0 and 0.000005 degrees is judged as a high-frequency conflict, and the distance difference between 0.000006 and 0.00002 degrees is judged as a normal conflict. Conflict resolution processing is performed on the matching pair. If the conflict type is detected as high-frequency, the matching pair with the contour segment index number consistent with the direction of the previous matching frame is retained and outliers are removed. If the same patch pixel corresponds to multiple projection points, the matching pair with the most continuous relationship between adjacent pixels is retained to maintain the boundary topological integrity. The comparison results of the two types of contour pixel positions are sorted to obtain the contour pixel position comparison results.

[0037] The coordinate association construction submodule reads the corresponding positions of the contour pixel positions of the illegal land occupation boundary patches and the contour pixel positions of the mining activity image patches based on the contour pixel position comparison results. It organizes the spatial correspondence information between the image ground projection position and the target contour pixel position, forms the organized result of the correspondence between the target pixel position of the video screen and the geographic coordinates, establishes the expression of the correspondence between the target pixel position of the video screen and the geographic coordinates, and generates the intelligent coordinate association result of the digital law enforcement video image of natural resources. Based on the comparison results of contour pixel positions, the contour pixel positions (116.39825, 39.90928) of the illegal land occupation boundary patch are read and their corresponding positions are matched with those of the mining activity image patch. The target pixel coordinates (2500, 1500) identified by the target edge detection operator in the original video frame are extracted, and the coordinate mapping linear regression coefficients are set. The coefficient is set to 1.0002. This coefficient is determined by referencing the cumulative error between image distortion residuals and geographic projection deformation. It is calculated by comparing the offset ratio between the measured control points and the calculated points in the field and taking a value between 1.0 and 1.001. The target pixel coordinates are then used. Geographic latitude and longitude coordinates The matrix concatenation operation performs a one-to-one attribute fusion of pixel domain feature vectors and spatial domain geographic vectors, sets the association confidence level, and considers an association deviation within 0.3 meters as extremely reliable, 0.4 to 1.5 meters as highly reliable. Here, a deviation of 0.2 meters is considered extremely reliable. The final persistent storage of the spatial correspondence information between the image ground projection position and the target contour pixels is performed, establishing a structured correspondence expression between the target pixel position and geographic coordinates in the video image, and generating intelligent coordinate association results for digital law enforcement video images of natural resources.

[0038] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A video image intelligent coordinate association system for digital law enforcement in natural resources, characterized in that, The system includes: The law enforcement video acquisition module uses drone patrol cameras to capture video footage of the patrol area, simultaneously read longitude, latitude, heading angle, pitch angle, and flight altitude, perform pixel analysis to identify the center position, record it in the image data record, and generate patrol image spatial reference information; The pixel offset generation module reads the target pixel coordinates based on the horizontal and vertical center pixel positions in the spatial reference information of the patrol image, performs a comparison with the center pixel position as a reference, determines the direction based on the results, determines the pixel offset position, and generates image pixel offset expression information. The line-of-sight angle calculation module, based on the heading angle and pitch angle in the patrol image spatial reference information, combined with the image pixel offset expression information, reads the camera's horizontal and vertical field of view, performs matching judgment, and performs correction direction in combination with the heading angle and pitch angle to determine the image pixel line-of-sight direction and generate image pixel line-of-sight expression information. The ground projection generation module determines the line-of-sight pointing position based on the image pixel line-of-sight representation information and the patrol image spatial reference information, combined with the longitude and latitude of the patrol equipment, performs spatial position association processing and ground position mapping processing, and generates image ground projection position representation information.

2. The intelligent coordinate association system for video images in digital law enforcement of natural resources according to claim 1, characterized in that: The system also includes: The coordinate association generation module reads the outline positions of illegal land occupation and mining activities image patches based on the image ground projection position expression information, compares the image ground projection positions with the outline positions of illegal land occupation and mining activities image patches respectively, establishes corresponding geographical location association relationships, constructs the correspondence between the target pixel positions of the video image and geographical coordinates, and generates intelligent coordinate association results for digital law enforcement video images of natural resources.

3. The intelligent coordinate association system for video images in digital law enforcement of natural resources according to claim 1, characterized in that: The law enforcement video acquisition module includes a patrol image acquisition submodule, a pixel matrix parsing submodule, and a spatial reference recording submodule; The patrol video acquisition submodule acquires video footage of the patrol area from the drone patrol camera, the longitude and latitude from the satellite positioning unit, the heading angle and pitch angle from the attitude sensor, and the flight altitude from the altimeter. It performs time-sequence processing on the video frame sequence and the longitude, latitude, heading angle, pitch angle, and flight altitude, performs a correspondence judgment between the video frame sequence and the equipment attitude record, and completes the unified processing of the video footage and the patrol equipment attitude record to generate the patrol video frame sequence. The pixel matrix parsing submodule reads the pixel matrix of the video frame based on the patrol video frame sequence, performs horizontal and vertical pixel sequence scanning processing of the pixel matrix, identifies the center pixel position of the video frame, performs corresponding arrangement of the center pixel position and the video frame pixel position, completes the identification of video frame pixel position and pixel coordinate arrangement, and generates the video center pixel coordinates. The spatial reference recording submodule, based on the coordinates of the video center pixel, calls longitude, latitude, heading angle, pitch angle and flight altitude to perform spatial position correspondence organization, performs horizontal center pixel position and vertical center pixel position matching judgment with the spatial position of the inspection equipment, completes the unified organization of equipment spatial position recording and image pixel position, establishes the correspondence record between image pixel position and inspection equipment spatial position, and generates inspection image spatial reference information.

4. The intelligent coordinate association system for video images in digital law enforcement of natural resources according to claim 1, characterized in that: The pixel offset generation module includes a center coordinate calling submodule, a pixel position comparison submodule, and a direction determination processing submodule; The center coordinate calling submodule reads the horizontal center pixel position and the vertical center pixel position according to the patrol image spatial reference information, reads the horizontal pixel coordinates and vertical pixel coordinates corresponding to the target position in the video frame, organizes the corresponding records of the horizontal center pixel position and the horizontal pixel coordinates, organizes the corresponding records of the vertical center pixel position and the vertical pixel coordinates, completes the unified recording of the target pixel position and the center pixel position in the video frame, and obtains the target pixel coordinate record. The pixel position comparison submodule calls the target pixel coordinate record, reads the horizontal and vertical pixel coordinates, calls the horizontal center pixel position to perform horizontal pixel position comparison processing, calls the vertical center pixel position to perform vertical pixel position comparison processing, and organizes the horizontal and vertical pixel position comparison results to obtain the pixel position comparison record. The orientation determination processing submodule, based on the pixel position comparison record, reads the horizontal pixel position comparison result and the vertical pixel position comparison result, performs horizontal pixel position orientation determination processing, performs vertical pixel position orientation determination processing, sorts the horizontal pixel orientation determination results and the vertical pixel orientation determination results, forms the pixel offset position corresponding to the center pixel position of the video image, and generates image pixel offset expression information.

5. The intelligent coordinate association system for video images in digital law enforcement of natural resources according to claim 1, characterized in that: The line-of-sight angle calculation module includes a field-of-sight range reading submodule, an offset matching determination submodule, and a line-of-sight direction correction submodule; The field of view reading submodule acquires the camera's horizontal and vertical field of view, calls the image pixel offset expression information, reads the horizontal and vertical pixel offset directions, calls the patrol image spatial reference information to read the heading and pitch angles, organizes the corresponding records of the horizontal and vertical pixel offset directions, horizontal and vertical field of view, completes the unified organization of image pixel offset directions and camera field of view, and generates a field of view matching preparation record. The offset matching determination submodule reads the horizontal pixel offset direction and the horizontal field of view based on the field of view matching preparation record, performs horizontal pixel offset direction matching determination processing, reads the vertical pixel offset direction and the vertical field of view to perform vertical pixel offset direction matching determination processing, organizes the horizontal pixel offset direction matching determination record and the vertical pixel offset direction matching determination record, and generates a pixel offset matching record. The line-of-sight direction correction submodule, based on the pixel offset matching record, calls the heading angle and pitch angle to carry out correction and organization of the lateral pixel offset direction and the longitudinal pixel offset direction, completes the unified organization of the pixel offset direction and the attitude direction of the inspection equipment, forms the observation direction record corresponding to the image pixel position, and generates the image pixel line-of-sight expression information.

6. The intelligent coordinate association system for video images in digital law enforcement of natural resources according to claim 1, characterized in that: The ground projection generation module includes a line-of-sight analysis submodule, a spatial location association submodule, and a ground mapping processing submodule; The line-of-sight analysis submodule calls the line-of-sight expression information of the image pixels, reads the line-of-sight direction of the image pixels, calls the spatial reference information of the patrol image to read the longitude, latitude and flight altitude of the patrol equipment, organizes the corresponding information of the line-of-sight direction of the image pixels and the spatial position of the patrol equipment, and performs line-of-sight direction recognition processing of the image pixels in combination with the longitude and latitude of the patrol equipment, and completes the unified organization of the line-of-sight direction of the image pixels and the position of the patrol equipment to generate the line-of-sight result; The spatial location association submodule reads the image pixel line-of-sight pointing position based on the image line-of-sight pointing result, calls the longitude and latitude of the patrol equipment to perform spatial location correspondence organization, organizes the image pixel line-of-sight pointing position and the longitude and latitude of the patrol equipment to form the spatial positioning result of the image pixel line-of-sight pointing position, and generates the line-of-sight pointing spatial position; The ground mapping processing submodule, based on the line-of-sight spatial position, calls the flight altitude to carry out ground position mapping and sorting, sorts out the corresponding information of the line-of-sight position of image pixels and the ground spatial position, forms a unified expression of the line-of-sight position of image pixels and the ground spatial position, and generates the image ground projection position expression information.

7. The intelligent coordinate association system for video images in digital law enforcement of natural resources according to claim 2, characterized in that: The coordinate association generation module includes a projection position reading submodule, a contour pixel comparison submodule, and a coordinate association construction submodule; The projection position reading submodule reads the image ground projection position based on the image ground projection position expression information, reads the image ground projection position, reads the outline pixel position of the illegal land occupation boundary patch and the outline pixel position of the mining activity image patch, organizes the corresponding information of the image ground projection position and the outline pixel position of the illegal land occupation boundary patch, organizes the corresponding information of the image ground projection position and the outline pixel position of the mining activity image patch, and completes the unified organization of the image ground projection position and the target outline pixel position to obtain the image projection outline position set. The contour pixel comparison submodule calls the image projection contour position set, reads the image ground projection position and the contour pixel position of the illegal land occupation boundary patch, performs comparison processing between the image ground projection position and the contour pixel position of the illegal land occupation boundary patch, reads the contour pixel position of the mining activity image patch and performs comparison processing between the image ground projection position and the contour pixel position of the mining activity image patch, sorts out the two types of contour pixel position comparison results, and obtains the contour pixel position comparison result. The coordinate association construction submodule, based on the contour pixel position comparison results, reads the corresponding positions of the contour pixel positions of the illegal land occupation boundary patches and the contour pixel positions of the mining activity image patches, organizes the spatial correspondence information between the image ground projection position and the target contour pixel position, forms the organized result of the correspondence between the target pixel position of the video screen and the geographic coordinates, establishes the expression of the correspondence between the target pixel position of the video screen and the geographic coordinates, and generates the intelligent coordinate association result of the digital law enforcement video image of natural resources.

8. The intelligent coordinate association system for video images in digital law enforcement of natural resources according to claim 7, characterized in that: The specific information regarding the correspondence between the ground projection position of the sorted image and the pixel position of the illegal land occupation boundary patch is as follows: The illegal land occupation boundary patches are arranged into closed contours according to connectivity and divided into continuous contour segments. For each image ground projection position, the corresponding contour segment is determined and the contour segment identifier and the position of adjacent contour pixels within the contour segment are recorded. The specific process of comparing the image ground projection position with the outline pixel position of the illegally occupied land boundary patch is as follows: The comparison range is defined by the contour segment identifier. Within the contour segment, point-by-point matching is performed in the order of adjacent contour pixel positions to generate matching pairs. The specific steps for organizing the comparison results of the two types of contour pixel positions are as follows: Conflict resolution is performed on the matching pairs. When the ground projection position of the same image corresponds to the outline pixel position of multiple illegal land occupation boundary patches, the matching pairs with the same outline segment order are retained. When the outline pixel position of the same illegal land occupation boundary patch corresponds to multiple ground projection positions of the image, the matching pairs with continuous adjacent relationships within the outline segment are retained.

Images