An intelligent monitoring system and method based on digital twins and drones

Abstract

The present invention discloses an intelligent monitoring system and method based on digital twins and unmanned aerial vehicles, and relates to the field of unmanned aerial vehicle monitoring technology. The system obtains image data taken by an unmanned aerial vehicle; determines a monitoring location with fixed obstructions based on the image data taken by the unmanned aerial vehicle; adjusts the speed of the unmanned aerial vehicle according to the information of the monitoring location with fixed obstructions, controls the unmanned aerial vehicle to pass through the monitoring location with fixed obstructions at a set speed and takes pictures; monitors the monitoring location based on the image data taken by the unmanned aerial vehicle; through the monitoring by the unmanned aerial vehicle, the monitoring range is wider and the safety is higher; when the unmanned aerial vehicle passes through an area with rapidly changing lighting conditions, the flight speed of the unmanned aerial vehicle is appropriately reduced to leave more time for the camera system to adapt to the changes in lighting conditions, thereby alleviating the influence of the speed on the captured image when the unmanned aerial vehicle passes through the area with rapidly changing lighting conditions to a certain extent, and higher quality image data can be obtained.

Description

Technical Field

[0001] The present invention relates to the field of drone monitoring technology, and in particular to an intelligent monitoring system and method based on digital twins and drones. Background Art

[0002] Drone monitoring is a method of remote monitoring using drone technology. It has a wide range of applications, including but not limited to security monitoring, environmental monitoring, traffic management, agricultural monitoring, etc. The basic principle of drone monitoring is to use various sensors and equipment carried by drones to conduct real-time observation and data collection of target areas. Drones can provide safer, more efficient and accurate monitoring services. The performance of drone monitoring systems depends largely on the quality of the collected data. Shadow areas caused by cloud cover, buildings and other obstructions have a significant impact on drone image acquisition. When a drone quickly passes through a shadow area, the camera's automatic exposure system has no time to adjust, resulting in inconsistent image exposure. The image captured by the camera becomes blurred due to rapidly changing lighting conditions, affecting the performance of the drone monitoring system. Therefore, how to improve the quality of image data collected by drones has become an urgent problem to be solved. Summary of the invention

[0003] The purpose of the present invention is to provide an intelligent monitoring system and method based on digital twins and drones to solve the problems raised in the above background technology.

[0004] In one aspect of the present invention, an intelligent monitoring method based on digital twins and drones is provided, comprising:

[0005] S11, obtaining image data taken by the drone;

[0006] S12, determining a monitoring location with fixed obstructions based on the image data captured by the drone;

[0007] S13, adjusting the speed of the drone according to the information of the monitoring location with fixed obstructions, controlling the drone to pass through the monitoring location with fixed obstructions at a set speed and take photos;

[0008] S14, monitoring the monitoring location based on the image data taken by the drone.

[0009] In step S12, the step of determining the monitoring location where fixed obstructions exist further includes the following steps:

[0010] Suppose the existing image data with inconsistent exposure problem is the first type of image data, and the existing image data without inconsistent exposure problem is the second type of image data. The first type of data and the second type of data are evenly divided into k regions, and the first type of image data and the second type of image data are judged whether there is inconsistent exposure problem by regional contrast; if the first type of image data is judged to have inconsistent exposure problem, the correct judgment number x1 of the first type of image data is increased by one, otherwise the correct judgment number x1 of the first type of image data remains unchanged; if the second type of image data is judged to have no inconsistent exposure problem, the correct judgment number x2 of the second type of image data is increased by one, otherwise the correct judgment number x2 of the second type of image data remains unchanged; calculate the correct judgment rate x1 of the first type of image data and the correct judgment rate x2 of the second type of image data, Where sx1 and sx2 are the total number of the first type of image data and the second type of image data;

[0011] Calculate the comprehensive accuracy y, y=k1x1+k2x2, k1 and k2 are preset coefficients, and determine the value of the number of regions k by the grid search method: set the initial value of k; set the step size to 1, and increase the value of k after each iteration. At the same time, determine the value of the comprehensive accuracy after each iteration. When the comprehensive accuracy reaches the maximum value, the search is completed, and the value of k is the final number of regions.

[0012] The image is divided into multiple regions and the average brightness or contrast of each region is calculated. If the brightness or contrast between adjacent regions is very different, it indicates that there may be exposure inconsistency problems.

[0013] Judging whether there is an exposure inconsistency problem between the first type of image data and the second type of image data by using the regional contrast also includes the following steps:

[0014] The first and second image data are converted into RGB images, and the pixel value G(i,j) of each pixel is calculated, G(i,j) = 0.299R(i,j) + 0.587G(i,j) + 0.114B(i,j), where (i,j) represents the position in the image, R(i,j), G(i,j) and B(i,j) are the red, green and blue channel values ​​at the position (i,j) respectively; the difference between the average pixel value of each region and the average pixel value of the adjacent region is calculated. If the difference between the average pixel values ​​of adjacent regions in the RGB image is greater than or equal to the first threshold, it is judged that the RGB image has an inconsistent exposure problem, and the corresponding first or second image data also has an inconsistent exposure problem; the first threshold is determined by an image known to have inconsistent exposure problems. For RGB images, the formula is used to convert to grayscale space. Since the brightness of the grayscale image is the weighted sum of the brightness of the three RGB color channels, the brightness of the pixel point is obtained.

[0015] Optionally, based on an image known to have inconsistent exposure problems, after dividing it into multiple regions, the difference between the average pixel value of each region and the adjacent pixel values ​​can be obtained, and the maximum of all differences is taken; each image with inconsistent exposure problems has a maximum value, and the minimum value between all maximum values ​​is taken to obtain the first threshold.

[0016] In step S12, the step of determining the monitoring location where fixed obstructions exist further includes the following steps:

[0017] S41, obtaining the image data taken by the drone and the location information of the drone when taking the image, extracting the image data of the monitoring location a taken at time b on different dates, judging the exposure inconsistency problem of the image data of the monitoring location a taken at time b on different dates by regional contrast, and calculating the probability P1(a,b) of the exposure inconsistency problem. Where n(a,b) represents the number of dates with inconsistent exposure, and sn(a,b) represents the total number of dates;

[0018] S42, in all cases of shooting time b, take the maximum value of P1(a,b) to obtain the probability P1(a) of inconsistent exposure problem occurring at monitoring location a, If P1(a) is greater than or equal to the second threshold, and the second threshold must be greater than 1-x2, then it is determined that there is fixed occlusion at the monitoring location a; traverse all values ​​of a to determine the fixed occlusion conditions of all monitoring locations.

[0019] Cloud occlusion and shadow areas caused by buildings and other obstructions have a significant impact on drone image acquisition. Shadow areas caused by clouds and ground obstructions will cause uneven lighting. Rapidly crossing the shadow area may cause the camera's automatic exposure system to not have time to adjust, resulting in inconsistent image exposure, too dark images in the shadow area, and too bright images in the sun area. Among the causes of shadow areas, cloud occlusion is not fixed, while buildings and other obstructions are relatively fixed. Cloud occlusion is dynamic because clouds move with the wind and constantly change their positions in the sky; relative to clouds, buildings and other ground obstructions are relatively fixed, and their positions do not change rapidly over time. Therefore, when crossing the shadow areas caused by these fixed obstructions, the flight speed of the drone needs to be adjusted. These fixed obstructions will produce continuous shadow areas on the ground. The size and shape of these shadows will change with the change of the sun's position, but the change speed is usually slow. When the drone monitors regularly according to the set monitoring route, the time when the drone arrives at the monitoring location is relatively consistent, and the size and shape of the shadow are also relatively consistent.

[0020] In order to determine the area with fixed occlusion, it is first necessary to determine the monitoring locations with inconsistent exposure problems. Then, in these determined monitoring locations, exclude the influence of movable occlusions such as clouds. To this end, the probability of inconsistent exposure problems at each location is judged. If there is only the influence of movable occlusions such as clouds, inconsistent exposure problems will only occur when movable occlusions such as clouds produce shadows, while areas with fixed occlusions will always have inconsistent exposure problems. In particular, there are monitoring locations that are affected by movable occlusions such as clouds for a long time. At this time, movable occlusions such as clouds are equivalent to fixed occlusions. By comparing the probability of inconsistent exposure problems at each location with the threshold, it is determined whether there is a fixed occlusion at the location. Since there is a 1-x2 probability that images without inconsistent exposure problems are misidentified, the second threshold must be greater than 1-x2. Optionally, the second threshold can be taken as the minimum value of the probability of inconsistent exposure problems at all monitoring locations with fixed occlusions based on the image data of the monitoring locations with fixed occlusions.

[0021] In step S13, the speed adjustment of the drone according to the monitoring location information with fixed obstructions also includes the following steps:

[0022] The monitoring location with fixed obstructions is calibrated, and the drone monitors according to the set driving route. When the distance between the drone and the calibration location decreases to the target distance, the drone starts to decelerate from the normal flight speed according to the deceleration curve. After passing the calibration location, the drone accelerates back to the normal flight speed. The target distance dis is calculated by the following formula: Where v(t) represents the deceleration curve of the UAV, t1 is the time it takes for the UAV to reduce its speed from normal to target according to the deceleration curve, and the target flight speed is the time required for the camera to respond to changes in illumination. Through field tests, the image quality of the UAV crossing the shadow area at different flight speeds is recorded, and the target flight speed is determined based on the test results.

[0023] The deceleration curve of the drone is determined by the following steps: inputting meteorological data and drone parameter information into the digital twin platform, constructing deceleration scenarios of the drone under different meteorological conditions, setting deceleration indicators, and selecting the deceleration curve with the best deceleration indicator. Optionally, the deceleration indicator is set to the energy consumption of the drone, and the deceleration process with the lowest energy consumption is selected directly according to the drone's instructions.

[0024] In another aspect of the present invention, an intelligent monitoring system based on digital twins and drones is provided, comprising: a drone monitoring module, a data analysis module, a digital twin platform, a data storage module and a data acquisition module; the output end of the drone monitoring module is connected to the input end of the data storage module to obtain image data of the monitoring location; the output end of the digital twin platform is connected to the input end of the data analysis module to construct a deceleration scenario for the drone under different meteorological conditions and determine the deceleration curve of the drone; the digital twin platform also simulates the scene of the monitoring location based on the monitoring image taken by the drone and monitors the monitoring location; the output end of the data storage module is connected to the input end of the data analysis module to store the monitoring images taken by the drone; the output end of the data acquisition module is connected to the input end of the digital twin platform to obtain meteorological data on the day of drone monitoring; the data analysis module is used to analyze the monitoring images taken by the drone and monitor the monitoring location.

[0025] The data analysis module also includes an image conversion unit, an image segmentation unit, an exposure detection unit, a fixed occlusion recognition unit and an image analysis unit; the image conversion unit is used to convert the image taken by the drone into an RGB image, the image segmentation unit is used to divide the image taken by the drone into several areas on average; the exposure detection unit is used to identify the exposure of the image taken by the drone; the fixed occlusion recognition unit determines the fixed occlusion of the monitoring location based on the exposure of the image taken by the drone; the image analysis unit is used to analyze the image taken by the drone and monitor the monitoring location. The image segmentation unit obtains the optimization target based on the correct judgment rate of the inconsistent exposure problem of the image data, and determines the number of areas after the image is segmented by a grid search method. The drone monitoring module also includes an information transmission unit, a positioning unit, an image acquisition unit and a control unit; the information transmission unit is used to realize information transmission between the drone and the data analysis module; the positioning unit is used to determine the position of the drone; the image acquisition unit is used to obtain image data of the monitoring location; the control unit is used to control the drone to go to the monitoring location to take images according to the set route.

[0026] Compared with the prior art, the beneficial effects achieved by the present invention are: monitoring by drones can fly to complex areas or high-risk environments that are difficult for traditional monitoring equipment to reach, with a wider monitoring range and higher safety; when the drone passes through an area where lighting conditions change rapidly, the flight speed of the drone is appropriately reduced, leaving more time for the camera system to adapt to changes in lighting conditions, which to a certain extent reduces the impact of the speed of the drone on the captured image when the drone passes through an area where lighting conditions change rapidly, and can obtain higher quality image data. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0028] Figure 1 It is a structural schematic diagram of an intelligent monitoring system based on digital twins and drones according to an embodiment of the present invention. DETAILED DESCRIPTION

[0029] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0030] In the embodiments of the present invention, please refer to Figure 1 , provides an intelligent monitoring system based on digital twins and drones, including: a drone monitoring module, a data analysis module, a digital twin platform, a data storage module and a data acquisition module; the output end of the drone monitoring module is connected to the input end of the data storage module to obtain image data of the monitoring location; the output end of the digital twin platform is connected to the input end of the data analysis module to construct a deceleration scenario of the drone under different meteorological conditions and determine the deceleration curve of the drone; the digital twin platform also simulates the scene of the monitoring location according to the monitoring image taken by the drone, and monitors the monitoring location; the output end of the data storage module is connected to the input end of the data analysis module to store the monitoring images taken by the drone; the output end of the data acquisition module is connected to the input end of the digital twin platform to obtain the meteorological data of the drone monitoring day; the data analysis module is used to analyze the monitoring images taken by the drone and monitor the monitoring location.

[0031] The data analysis module also includes an image conversion unit, an image segmentation unit, an exposure detection unit, a fixed occlusion recognition unit and an image analysis unit; the image conversion unit is used to convert the image taken by the drone into an RGB image, the image segmentation unit is used to divide the image taken by the drone into several areas on average; the exposure detection unit is used to identify the exposure of the image taken by the drone; the fixed occlusion recognition unit determines the fixed occlusion of the monitoring location based on the exposure of the image taken by the drone; the image analysis unit is used to analyze the image taken by the drone and monitor the monitoring location. The image segmentation unit obtains the optimization target based on the correct judgment rate of the inconsistent exposure problem of the image data, and determines the number of areas after the image is segmented by a grid search method. The drone monitoring module also includes an information transmission unit, a positioning unit, an image acquisition unit and a control unit; the information transmission unit is used to realize information transmission between the drone and the data analysis module; the positioning unit is used to determine the position of the drone; the image acquisition unit is used to obtain image data of the monitoring location; the control unit is used to control the drone to go to the monitoring location to take images according to the set route.

[0032] In an embodiment of the present invention, an intelligent monitoring method based on digital twins and drones is provided, comprising:

[0033] S11, obtaining image data taken by the drone;

[0034] S12, determining a monitoring location with fixed obstructions based on the image data captured by the drone;

[0035] Suppose the existing image data with inconsistent exposure problem is the first type of image data, and the existing image data without inconsistent exposure problem is the second type of image data. The first type of data and the second type of data are evenly divided into k regions, and the first type of image data and the second type of image data are judged whether there is inconsistent exposure problem by regional contrast; if the first type of image data is judged to have inconsistent exposure problem, the correct judgment number x1 of the first type of image data is increased by one, otherwise the correct judgment number x1 of the first type of image data remains unchanged; if the second type of image data is judged to have no inconsistent exposure problem, the correct judgment number x2 of the second type of image data is increased by one, otherwise the correct judgment number x2 of the second type of image data remains unchanged; calculate the correct judgment rate x1 of the first type of image data and the correct judgment rate x2 of the second type of image data, Where sx1 and sx2 are the total number of the first type of image data and the second type of image data;

[0036] Specifically, determine whether there is an exposure inconsistency problem through the following steps:

[0037] The first type of image data and the second type of image data are converted into RGB images, and the pixel value G(i,j) of each pixel is calculated, G(i,j)=0.299R(i,j)+0.587G(i,j)+0.114B(i,j), where (i,j) represents the position in the image, R(i,j), G(i,j) and B(i,j) are the red, green and blue channel values ​​at the position (i,j), respectively; the difference between the average pixel value of each area and the average pixel value of the adjacent area is calculated, and if the difference between the average pixel values ​​of adjacent areas in the RGB image is greater than or equal to the first threshold, it is determined that the RGB image has an exposure inconsistency problem, and the corresponding first type of image data or the second type of image data also has an exposure inconsistency problem; the first threshold is determined by an image known to have an exposure inconsistency problem.

[0038] Optionally, based on an image known to have inconsistent exposure problems, after dividing it into multiple regions, the difference between the average pixel value of each region and the adjacent pixel values ​​can be obtained, and the maximum of all differences is taken; each image with inconsistent exposure problems has a maximum value, and the minimum value between all maximum values ​​is taken to obtain the first threshold.

[0039] When partitioning an image, first set a smaller initial value for the number of blocks in the horizontal and vertical directions of the image. Then, keep the number of blocks in the vertical direction unchanged and increase the number of blocks in the horizontal direction by one each time to obtain a comprehensive accuracy. Then, increase the number of blocks in the vertical direction by one each time. When the number of blocks in the vertical direction changes, the number of blocks in the horizontal direction increases from the initial value with 1 as compensation. Take the maximum value of all comprehensive accuracy rates. The product of the number of blocks in the vertical direction and the number of blocks in the horizontal direction corresponding to the maximum value is the number of regions into which the image is divided.

[0040] Calculate the comprehensive accuracy y, y=k1x1+k2x2, k1 and k2 are preset coefficients, and determine the value of the number of regions k by the grid search method: set the initial value of k; set the step size to 1, and increase the value of k after each iteration. At the same time, determine the value of the comprehensive accuracy after each iteration. When the comprehensive accuracy reaches the maximum value, the search is completed, and the value of k is the final number of regions.

[0041] Specifically, the monitoring locations with fixed obstructions are determined through the following steps:

[0042] S41, obtaining the image data taken by the drone and the location information of the drone when taking the image, extracting the image data of the monitoring location a taken at time b on different dates, judging the exposure inconsistency problem of the image data of the monitoring location a taken at time b on different dates by regional contrast, and calculating the probability P1(a,b) of the exposure inconsistency problem. Where n(a,b) represents the number of dates with inconsistent exposure, and sn(a,b) represents the total number of dates;

[0043] S42, in all cases of shooting time b, take the maximum value of P1(a,b) to obtain the probability P1(a) of inconsistent exposure problem occurring at monitoring location a, If P1(a) is greater than or equal to the second threshold, and the second threshold must be greater than 1-x2, then it is determined that there is fixed occlusion at the monitoring location a; traverse all values ​​of a to determine the fixed occlusion conditions of all monitoring locations.

[0044] Optionally, the second threshold may be determined based on the image data of the monitoring location with fixed obstructions, and the minimum value of the probability of inconsistent exposure problems occurring at all monitoring locations with fixed obstructions.

[0045] S13, adjusting the speed of the drone according to the information of the monitoring location with fixed obstructions, controlling the drone to pass through the monitoring location with fixed obstructions at a set speed and take photos;

[0046] The monitoring location with fixed obstructions is calibrated, and the drone monitors according to the set driving route. When the distance between the drone and the calibration location decreases to the target distance, the drone starts to decelerate from the normal flight speed according to the deceleration curve. After passing the calibration location, the drone accelerates back to the normal flight speed. The target distance dis is calculated by the following formula: In the formula, v(t) represents the deceleration curve of the drone, t1 is the time it takes for the drone to reduce its normal flight speed to the target flight speed according to the deceleration curve, and the target flight speed is the time required for the camera to respond to changes in illumination. Through field tests, the image quality of the drone crossing the shadow area at different flight speeds is recorded, and the target flight speed is determined based on the test results. The meteorological data and the parameter information of the drone are input into the digital twin platform, the deceleration scenes of the drone under different meteorological conditions are constructed, the deceleration index is set, and the deceleration curve with the best deceleration index is selected.

[0047] S14, monitoring the monitoring location based on the image data taken by the drone.

[0048] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0049] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An intelligent monitoring method based on digital twins and drones, characterized in that: The following steps are involved: S11, obtaining image data taken by the drone; S12, based on the image data taken by the drone, determining the monitoring location with fixed obstructions, including the following steps S41 and S42: S41, obtaining the image data taken by the drone and the location information of the drone when taking the image, extracting the image data of the monitoring location a taken at time b on different dates, judging the exposure inconsistency problem of the image data of the monitoring location a taken at time b on different dates by regional contrast, and calculating the probability P1(a,b) of the exposure inconsistency problem. Where n(a,b) represents the number of dates with inconsistent exposure, and sn(a,b) represents the total number of dates; S42, in all cases of shooting time b, take the maximum value of P1(a,b) to obtain the probability P1(a) of inconsistent exposure problem occurring at monitoring location a, If P1(a) is greater than or equal to the second threshold, and the second threshold must be greater than 1-x2, it is determined that there is fixed occlusion at the monitoring location a; traverse all values ​​of a to determine the fixed occlusion of all monitoring locations; where x2 is the correct judgment rate of the second type of image data, and the existing image data without exposure inconsistency is the second type of image data; S13, adjusting the speed of the drone according to the information of the monitoring location with fixed obstructions, controlling the drone to pass through the monitoring location with fixed obstructions at a set speed and take photos; adjusting the speed of the drone also includes the following steps: The monitoring location with fixed obstructions is calibrated, and the drone monitors according to the set driving route. When the distance between the drone and the calibration location decreases to the target distance, the drone starts to decelerate from the normal flight speed according to the deceleration curve. After passing the calibration location, the drone accelerates back to the normal flight speed. The target distance dis is calculated by the following formula: In the formula, v(t) represents the deceleration curve of the UAV, t1 is the time it takes for the UAV to reduce its normal flight speed to the target flight speed according to the deceleration curve, and the target flight speed is the time required for the camera to respond to changes in illumination. Through field tests, the image quality of the UAV crossing the shadow area at different flight speeds is recorded, and the target flight speed is determined based on the test results; S14, monitoring the monitoring location based on the image data taken by the drone.

2. According to claim 1, the intelligent monitoring method based on digital twins and drones is characterized in that: In step S12, the step of determining the monitoring location where fixed obstructions exist further includes the following steps: Suppose the existing image data with inconsistent exposure problem is the first type of image data, and the existing image data without inconsistent exposure problem is the second type of image data. The first type of data and the second type of data are evenly divided into k regions, and the first type of image data and the second type of image data are judged whether there is inconsistent exposure problem by regional contrast; if the first type of image data is judged to have inconsistent exposure problem, the correct judgment number x1 of the first type of image data is increased by one, otherwise the correct judgment number x1 of the first type of image data remains unchanged; if the second type of image data is judged to have no inconsistent exposure problem, the correct judgment number x2 of the second type of image data is increased by one, otherwise the correct judgment number x2 of the second type of image data remains unchanged; calculate the correct judgment rate x1 of the first type of image data and the correct judgment rate x2 of the second type of image data, Where sx1 and sx2 are the total number of the first type of image data and the second type of image data; Calculate the comprehensive accuracy y, y=k1x1+k2x2, k1 and k2 are preset coefficients, and determine the value of the number of regions k by the grid search method: set the initial value of k; set the step size to 1, and increase the value of k after each iteration. At the same time, determine the value of the comprehensive accuracy after each iteration. When the comprehensive accuracy reaches the maximum value, the search is completed, and the value of k is the final number of regions.

3. According to claim 2, the intelligent monitoring method based on digital twins and drones is characterized in that: Judging whether there is an exposure inconsistency problem between the first type of image data and the second type of image data by using the regional contrast also includes the following steps: The first type of image data and the second type of image data are converted into RGB images, and the pixel value G(i,j) of each pixel is calculated, G(i,j)=0.299R(i,j)+0.587G(i,j)+0.114B(i,j), where (i,j) represents the position in the image, R(i,j), G(i,j) and B(i,j) are the red, green and blue channel values ​​at the position (i,j), respectively; the difference between the average pixel value of each area and the average pixel value of the adjacent area is calculated, and if the difference between the average pixel values ​​of adjacent areas in the RGB image is greater than or equal to the first threshold, it is determined that the RGB image has an exposure inconsistency problem, and the corresponding first type of image data or the second type of image data also has an exposure inconsistency problem; the first threshold is determined by an image known to have an exposure inconsistency problem.

4. According to claim 1, the intelligent monitoring method based on digital twins and drones is characterized in that: The deceleration curve of the UAV is determined by the following steps: The meteorological data and the parameter information of the UAV are input into the digital twin platform, the deceleration scenarios of the UAV under different meteorological conditions are constructed, the deceleration index is set, and the deceleration curve with the best deceleration index is selected.

5. An intelligent monitoring system based on digital twins and drones, using an intelligent monitoring method based on digital twins and drones as described in any one of claims 1 to 4, characterized in that: include: UAV monitoring module, data analysis module, digital twin platform, data storage module and data acquisition module; The output end of the drone monitoring module is connected to the input end of the data storage module to obtain image data of the monitoring location; the output end of the digital twin platform is connected to the input end of the data analysis module to construct the deceleration scene of the drone under different meteorological conditions and determine the deceleration curve of the drone; the digital twin platform also simulates the scene of the monitoring location based on the monitoring image taken by the drone and monitors the monitoring location; the output end of the data storage module is connected to the input end of the data analysis module to store the monitoring image taken by the drone; the output end of the data acquisition module is connected to the input end of the digital twin platform to obtain the meteorological data of the drone monitoring day; The data analysis module is used to analyze the monitoring images taken by the drone and monitor the monitoring location.

6. The intelligent monitoring system based on digital twins and drones according to claim 5 is characterized in that: The data analysis module also includes an image conversion unit, an image blocking unit, an exposure detection unit, a fixed occlusion recognition unit and an image analysis unit; the image conversion unit is used to convert the image taken by the drone into an RGB image, and the image blocking unit is used to divide the image taken by the drone into a number of areas on average; the exposure detection unit is used to identify the exposure condition of the image taken by the drone; the fixed occlusion recognition unit determines the fixed occlusion condition of the monitoring location based on the exposure condition of the image taken by the drone; the image analysis unit is used to analyze the image taken by the drone and monitor the monitoring location.

7. The intelligent monitoring system based on digital twins and drones according to claim 6 is characterized in that: The image segmentation unit obtains an optimization target based on the correct judgment rate of the exposure inconsistency problem of the image data, and determines the number of regions after the image is segmented by a grid search method.

8. The intelligent monitoring system based on digital twins and drones according to claim 5 is characterized in that: The drone monitoring module also includes an information transmission unit, a positioning unit, an image acquisition unit and a control unit; the information transmission unit is used to realize information transmission between the drone and the data analysis module; the positioning unit is used to determine the position of the drone; the image acquisition unit is used to obtain image data of the monitoring location; The control unit is used to control the drone to go to the monitoring location to capture images along a set route.

Images