Property project data fusion and inspection method and device based on unmanned aerial vehicle aerial photography
By optimizing the drone's flight path and parameters, and taking into account battery power and geographical environment, the problems of rapid battery consumption and insufficient image clarity during property area inspections have been solved, enabling efficient and clear multi-point inspections and assessments.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGZHOU HANZHISOFT CO LTD
- Filing Date
- 2026-03-26
- Publication Date
- 2026-06-12
Smart Images
Figure CN122195034A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the fields of unmanned aerial vehicles (UAVs) and data processing technology, specifically to a method and apparatus for data fusion and inspection of property projects based on UAV aerial photography. Background Technology
[0002] Green spaces and trees are not only an important part of the Earth's ecosystem, but also a core focus of property inspection and management. Whether in residential communities or parks, grass and trees play a vital role in absorbing carbon dioxide and releasing oxygen, thereby maintaining the atmospheric environment and ecological balance of the site.
[0003] The property management company needs to regularly irrigate, spray pesticides, and remove weeds from the green spaces and trees. Because the grass in the green spaces is low and the area is wide, and the trees have branches and leaves, it is easy to miss some areas by manual inspection. In addition, the property may have multiple facilities in the same location, such as multiple buildings and multiple parks, which increases the cost of manual inspection.
[0004] In recent years, drones have been widely used in surveying, agriculture, environmental monitoring and emergency rescue. For example, drones take aerial photos of the ground during inspections to achieve target monitoring. However, there are still some shortcomings in multi-point inspections by drones: (1) Drones inspect multiple buildings and parks one by one, which consumes a lot of power and may not be able to cover all the areas to be inspected. In some cases, the power may run out during the inspection and the drone may not be able to return. (2) In mountainous areas, when flying from one park to another for inspection, the drone cannot fly at low altitude the whole time and needs to adjust the flight altitude according to the height of the ridge. (3) When drones take aerial photos of the green areas in buildings and parks, if the drone is too far from the ground, the drone cannot clearly take pictures of the grass and trees on the green areas, which makes it impossible for property staff to judge the growth status of the grass and trees, such as the color of the grass and trees, whether there are many yellow leaves and grass piles, and whether there are many weeds on the green areas.
[0005] In summary, there is an urgent need for a method that comprehensively considers the dynamic changes in drone battery power, the geographical environment of the area to be inspected, and the height of tree canopies in the property area, and controls the drone's dynamic ascent and descent over long cruising distances, thereby increasing the clarity of images when taking pictures of green spaces and trees in the property area from multiple points and reducing the drone's energy consumption. Summary of the Invention
[0006] To overcome the problems existing in related technologies, the purpose of this invention is to provide a method and device for data fusion and inspection of property projects based on drone aerial photography. The method comprehensively considers the dynamic changes in the drone's battery power, the geographical environment of the area to be inspected, and the height of the tree canopy in the property area. It controls the drone's dynamic ascent and descent over long cruising distances, thereby increasing the clarity of the images when taking pictures of green spaces and trees in the property area from multiple points and reducing the drone's energy consumption.
[0007] Data fusion and inspection methods for property projects based on drone aerial photography include: Obtain the property area and design constraints based on the property area; By combining constraints and a multi-objective loss function, a flight path and flight parameters are designed; wherein, the flight parameters include the UAV's dynamic flight altitude and dynamic flight speed. Control the drone to perform inspections according to the stated flight path and flight parameters; The relative position between the drone and the property area is monitored in real time. If the relative position is the target position, the drone is controlled to take pictures of the property area from different angles to obtain drone images. The maintenance status of the property area is determined based on the detection results obtained by integrating the drone images.
[0008] In a preferred embodiment of the present invention, the step of applying the design constraints of the property area includes: The first inequality is constructed using the drone's battery power, the inspected locations, and the power consumption per unit time; Construct a second inequality based on the elevation curves of the ridges surrounding the property area; A third inequality is constructed based on the tree canopy height of the property area.
[0009] In a preferred embodiment of the present invention, before combining the constraint conditions and the multi-objective loss function, the method further includes: Design a multi-objective loss function based on the following formula: ; Among them, L final Let L represent the multi-objective loss function. cross L represents the cross-entropy loss function, which characterizes the difference between the predicted drone inspection probability distribution and the actual property area distribution; energy L represents the energy loss function, which characterizes the remaining battery power and the rate of battery consumption of the UAV; GIOU The target overlap loss function is used to characterize the degree of overlap between the UAV's flight area and the corresponding area of obstacles, including buildings, ridges, and forests.
[0010] In a preferred embodiment of the present invention, the calculation formula for the cross-entropy loss function is as follows: ; Among them, y i This represents the label of the i-th inspection location of the drone. i=1, which means that the i-th inspection location is a real property area; if y i =0, which means the i-th inspection location is in another area; p i Let represent the probability that the drone will pass through the i-th inspection location, log represents the logarithmic function with the natural constant as the base, and N represents the total number of inspection locations.
[0011] In a preferred embodiment of the present invention, the formula for calculating the energy loss function is as follows: ; Where A represents the drone's current battery level, exp represents the exponential function, k represents the energy loss factor (k > 0), and x represents the remaining flight distance. This represents the offset term caused by drag during the flight of the drone. It is positively correlated with the drag experienced by the drone.
[0012] In a preferred embodiment of the present invention, the calculation formula for the target overlap loss function is as follows: ; ; Where IoU represents the intersection-union ratio loss function, and c represents the overlap coefficient. Represents the three-dimensional center distance. Indicates the prediction box. Represents the true bounding box. This represents the three-dimensional center distance between the predicted bounding box and the ground truth bounding box; This represents the intersection operation. This indicates the union operation.
[0013] In a preferred embodiment of the present invention, the step of designing the flight path and flight parameters by combining constraints and a multi-objective loss function includes: The first inequality is used to constrain the total number N of inspection locations in the cross-entropy loss function; The energy loss factor in the energy loss function is constrained by the second inequality. The overlap coefficient in the third inequality constraint is adopted.
[0014] In a preferred embodiment of the present invention, the step of constraining the total number N of inspection locations in the cross-entropy loss function using the first inequality includes: The first inequality is constructed based on the following formula: ; Where QUA represents the drone's battery level, N0 represents the inspected locations, inter represents the time interval between two adjacent inspected locations (inter is a fixed value), and PC represents the power consumption per unit time (QUA is expressed as a percentage). The first inequality is used as a constraint condition for the cross-entropy loss function.
[0015] In a preferred embodiment of the present invention, the step of fusing the detection results of the UAV images and determining the maintenance status of the property area based on the detection results includes: Determine the range of pixel values for the green area; Extract the green area from the drone image, which is an RGB image; Calculate the ratio of the green area to the entire drone image to obtain the green area percentage; The maintenance status of the property area is determined based on the percentage of green areas.
[0016] This invention also provides a property project data fusion and inspection device based on drone aerial photography, comprising: The constraint design module is used to obtain the property area and design constraints based on the property area. The UAV parameter design module is used to design flight routes and flight parameters by combining constraints and multi-objective loss functions; wherein, the flight parameters include the UAV dynamic flight altitude and the UAV dynamic flight speed; The drone inspection module is used to control the drone to perform inspections according to the flight route and flight parameters. The drone shooting module is used to monitor the relative position between the drone and the property area in real time. If the relative position is the target position, the drone is controlled to shoot the property area from different angles to obtain drone images. The detection result fusion module is used to fuse the detection results of the UAV images and determine the maintenance status of the property area based on the detection results.
[0017] The beneficial effects of this invention are as follows: The present invention provides a method for data fusion and inspection of property projects based on drone aerial photography, which includes acquiring the property area and designing constraints based on the property area. Combining the constraints and a multi-objective loss function, a flight route and flight parameters are designed, including the drone's dynamic flight altitude and dynamic flight speed. The constraints include three inequalities, and the multi-objective loss function consists of three loss functions: cross-entropy loss function, energy loss function, and target overlap loss function. The first inequality constrains the cross-entropy loss function to predict the distribution differences between the locations that the drone can inspect and the locations to be inspected, taking into account the drone's battery level. The second inequality constrains the energy loss function to adjust the drone's ascent and descent based on the ridge height, controlling the drone's optimal flight altitude in stages, thereby saving the drone's remaining battery power and reducing the rate of battery consumption. The third inequality constrains the target overlap loss function to minimize the distance between the drone and the trees and green areas of the inspected property site, while avoiding collisions between the drone and trees, and preventing drone components from becoming entangled in branches. The system controls drones to patrol the property area according to flight routes and parameters, monitoring their relative position to the property area in real time. If the relative position matches the target location, the drone is controlled to capture images of the property area from different angles. The detection results from the drone images are then fused to determine the maintenance status of the property area. The maintenance status is further determined by the proportion of green areas; the less yellow and the more green, the better the growth of green spaces and trees, and thus the better the maintenance status of the property area. Attached Figure Description
[0018] Figure 1 This is a flowchart of the property project data fusion and inspection method based on drone aerial photography of the present invention; Figure 2 This is a flowchart of the present invention based on the design constraints of the property area. Detailed Implementation
[0019] Preferred embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While preferred embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that the invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.
[0020] Example 1 like Figure 1 As shown, this embodiment provides a method for data fusion and inspection of property projects based on drone aerial photography, including: S1: Obtain the property area and design constraints based on the property area; S2: Combining constraints and multi-objective loss functions, design flight routes and flight parameters; wherein, the flight parameters include the UAV's dynamic flight altitude and the UAV's dynamic flight speed; S3: Control the drone to perform inspections according to the stated flight route and flight parameters; S4: Monitor the relative position between the drone and the property area in real time. If the relative position is the target position, control the drone to take pictures of the property area from different angles to obtain drone images. S5: Combine the detection results of the drone images and determine the maintenance status of the property area based on the detection results.
[0021] The design constraints based on the property area include: S12: Construct the first inequality using the drone's battery power, the inspected locations, and the power consumption per unit time; S13: Construct a second inequality based on the elevation curves of the ridges surrounding the property area; S14: Construct a third inequality based on the canopy height of the property area.
[0022] Before step S12, step S11 is included: obtaining the property area. The step of using the first inequality to constrain the total number N of inspection locations in the cross-entropy loss function includes: The first inequality is constructed based on the following formula: (1) Where QUA represents the drone's battery level, N0 represents the inspected locations, inter represents the time interval between two adjacent inspected locations (inter is a fixed value), and PC represents the power consumption per unit time (QUA is expressed as a percentage). The first inequality is used as a constraint condition for the cross-entropy loss function.
[0023] This invention assumes that all property areas to be inspected are uniformly distributed, meaning that the distance between adjacent property areas is within a certain range. Therefore, a fixed value can be used to describe the time interval between two adjacent inspection locations. N-N0 represents the total number of locations to be inspected. This represents the power required for the current inspection location. An additional 10% of the power is reserved to ensure sufficient power for the drone. N is dynamically changing; if formula (1) is not satisfied, i.e. If the value of N is not satisfied, then the value of N is reduced by 1. If the value of N is still not satisfied, then the value of N is reduced by 2, and so on, until the value of N satisfies the formula (1), so as to inspect more property areas while ensuring that the drone returns to base.
[0024] Construct the second inequality based on the following formula: (2) Where k represents the energy loss factor, adj1 represents the energy adjustment parameter, exp represents the exponential function, and x represents the horizontal coordinate of the ridge on the ground. Let A represent the horizontal coordinate of the i-th peak of the ridge. i This represents the peak height of the i-th mountain peak. This represents the width parameter of the i-th peak, used to control the slope of the ridge.
[0025] Construct the third inequality based on the following formula: (3) Where row represents the horizontal coordinate of the ground, col represents the vertical coordinate of the ground, DSM represents the digital surface model, and DEM represents the digital elevation model. This represents the canopy height at the location with the horizontal coordinate "row" and the vertical coordinate "col". `adj2` represents the target overlap adjustment parameter; `adj2` is greater than 1 and less than 2, used to appropriately increase the canopy height. The overlap coefficient `c` after the third inequality constraint is too large, causing the target overlap loss function `L` to... GIOU The distance is too small, so the 3D center distance between the predicted box and the ground truth box does not need to be too small, nor does the IoU need to be too large. This prevents the drone from getting too close to the tree canopy, avoids the drone from colliding with the tree, and prevents drone parts such as propellers from getting tangled in the branches.
[0026] Before combining the constraints and the multi-objective loss function, the following is also included: Design a multi-objective loss function based on the following formula: (4) Among them, L final Let L represent the multi-objective loss function. cross L represents the cross-entropy loss function, which characterizes the difference between the predicted drone inspection probability distribution and the actual property area distribution; energy L represents the energy loss function, which characterizes the remaining battery power and the rate of battery consumption of the UAV; GIOU The target overlap loss function is used to characterize the degree of overlap between the UAV's flight area and the corresponding area of obstacles, including buildings, ridges, and forests.
[0027] The formula for calculating the cross-entropy loss function is as follows: (5) Among them, y i This represents the label of the i-th inspection location of the drone. i=1, which means that the i-th inspection location is a real property area; if y i =0, which means the i-th inspection location is in another area; p i Let represent the probability that the drone will pass through the i-th inspection location, log represents the logarithmic function with the natural constant as the base, and N represents the total number of inspection locations.
[0028] The labels for inspection locations include both actual property areas and other areas, when y i When y = 0, the corresponding term is 0 and is not included in the measurement of the cross-entropy loss function. i When =1, the corresponding term is - Because p i >0, therefore - >0. p i and- The cross-entropy loss function is negatively correlated, so the higher the probability that the drone will pass through each inspection location, the smaller the cross-entropy loss function, and the greater the probability that the drone will inspect all property areas.
[0029] The formula for calculating the energy loss function is as follows: (6) Where A represents the drone's current battery level, exp represents the exponential function, k represents the energy loss factor (k > 0), and x represents the remaining flight distance. This represents the offset term caused by drag during the flight of the drone. It is positively correlated with the drag experienced by the drone.
[0030] The energy consumption of the drone is positively correlated with its flight speed, air density during flight, and surface roughness of its components. Formula (6) is used to measure the drone's energy consumption during flight. The greater the remaining flight distance and drag, the lower the drone's current battery level, and the higher the energy loss function L. energy The larger the value of L, the greater the energy loss function L. energy This reflects the extent of the property area that the drone can patrol under current battery power, weather conditions, and component condition. The drone's flight altitude and speed are adjusted based on an energy loss function, allowing for controlled ascent and descent according to ridge height. This phased control of the drone's optimal flight altitude adapts to different terrains, such as mountains and plains, thereby reducing energy consumption in different flight areas.
[0031] The formula for calculating the target overlap loss function is as follows: (7) (8) Where IoU represents the intersection-union ratio loss function, and c represents the overlap coefficient. Represents the three-dimensional center distance. Indicates the prediction box. Represents the true bounding box. This represents the three-dimensional center distance between the predicted bounding box and the ground truth bounding box; This represents the intersection operation. This indicates the union operation.
[0032] The Intersection over Union (IoU) loss function is used to measure the distance deviation between the drone and the location to be inspected on a plane. This is used to measure the distance deviation between the drone and the inspection site in the vertical direction. The higher the overlap between the drone and the inspection site in the horizontal plane, the higher the target overlap loss function. The smaller the value, the closer the drone is to the inspection location in the altitude direction, and the lower the target overlap loss function. The smaller.
[0033] The first inequality constrains the total number N of inspection locations in the cross-entropy loss function, the second inequality constrains the energy loss factor in the energy loss function, and the third inequality constrains the overlap coefficient. Flight routes and parameters are generated based on the multi-objective loss function. The tool used in this invention to generate UAV flight routes is DJI Terra or QGIS. The multi-objective loss function consists of a cross-entropy loss function, an energy loss function, and a target overlap loss function. The flight route is continuously adjusted and flight parameters are optimized, including flight speed and altitude. The value of the multi-objective loss function is calculated for each flight route and each set of flight parameters. After calculating the first batch of 20 sets of multi-objective loss function values, the smallest multi-objective loss function value in the first batch is selected. If the smallest multi-objective loss function value is less than or equal to the multi-objective loss function threshold, the corresponding UAV flight route and flight parameters are used as the final values, and the UAV is controlled to inspect according to the stated flight route and flight parameters. If the smallest multi-objective loss function value is greater than the multi-objective loss function threshold, a second batch of 20 sets of multi-objective loss function values is calculated.
[0034] Collecting and fusing drone data is used to calculate the usage of equipment, manpower, and materials, ultimately serving as a reference standard cost for comparison and control with the company's actual operating costs.
[0035] This embodiment provides a data fusion and inspection method for property projects based on UAV aerial photography. The method includes acquiring the property area and designing constraints based on the property area. Combining these constraints with a multi-objective loss function, a flight route and flight parameters are designed. These flight parameters include the UAV's dynamic flight altitude and dynamic flight speed. The constraints include three inequalities, and the multi-objective loss function consists of three loss functions: cross-entropy loss, energy loss, and target overlap loss. The first inequality constrains the cross-entropy loss function to predict the distribution differences between the locations the UAV can inspect and the locations to be inspected, taking into account the UAV's battery level. The second inequality constrains the energy loss function to adjust the UAV's ascent and descent based on the ridge height, controlling the optimal flight altitude in stages to save remaining battery power and reduce the rate of battery consumption. The third inequality constrains the target overlap loss function to minimize the distance between the UAV and the trees and green areas of the inspected property site, while avoiding collisions between the UAV and trees, and preventing UAV components from becoming entangled in branches. The system controls drones to patrol the property area according to flight routes and parameters, monitoring their relative position to the property area in real time. If the relative position matches the target location, the drone is controlled to capture images of the property area from different angles. The detection results from the drone images are then fused to determine the maintenance status of the property area. The maintenance status is further determined by the proportion of green areas; the less yellow and the more green, the better the growth of green spaces and trees, and thus the better the maintenance status of the property area.
[0036] Example 2 like Figure 1 As shown, this embodiment describes the differences from Embodiment 1, based on Embodiment 1. The step of fusing the detection results of the UAV images and determining the maintenance status of the property area based on the detection results includes: S51: Determine the range of pixel values for the green area; S52: Extract the green area from the drone image, where the drone image is an RGB image; S53: Calculate the ratio of the green area to the entire UAV image to obtain the green area percentage; S54: Determine the maintenance status of the property area based on the proportion of the green area.
[0037] Drone images consist of R, G, and B channels. This invention selects the G channel of the drone image, where the values range from 0 to 255. A green threshold of 120 is set, and areas in the G channel of the drone image with values greater than 120 are designated as green areas.
[0038] The total number of pixels within the green area is counted, and then divided by the total number of pixels in the drone image to obtain the percentage of the green area.
[0039] Since the RGB camera on the drone is aimed at the property area, such as green space, which contains grass and trees, it can be divided into multiple zones to assess the maintenance status of the property area. If the green area accounts for less than 40%, the maintenance status of the property area is judged as poor; if the green area accounts for more than or equal to 40% but less than 80%, the maintenance status of the property area is judged as good; if the green area accounts for more than or equal to 80%, the maintenance status of the property area is judged as excellent.
[0040] This embodiment also provides a property project data fusion and inspection device based on drone aerial photography, including: The constraint design module is used to obtain the property area and design constraints based on the property area. The UAV parameter design module is used to design flight routes and flight parameters by combining constraints and multi-objective loss functions; wherein, the flight parameters include the UAV dynamic flight altitude and the UAV dynamic flight speed; The drone inspection module is used to control the drone to perform inspections according to the flight route and flight parameters. The drone shooting module is used to monitor the relative position between the drone and the property area in real time. If the relative position is the target position, the drone is controlled to shoot the property area from different angles to obtain drone images. The detection result fusion module is used to fuse the detection results of the UAV images and determine the maintenance status of the property area based on the detection results.
[0041] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.
[0042] The above description is only a preferred embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural changes made based on the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.
Claims
1. A method for data fusion and inspection of property projects based on drone aerial photography, characterized in that, include: Obtain the property area and design constraints based on the property area; By combining constraints and a multi-objective loss function, a flight path and flight parameters are designed; wherein, the flight parameters include the UAV's dynamic flight altitude and dynamic flight speed. Control the drone to perform inspections according to the stated flight path and flight parameters; The relative position between the drone and the property area is monitored in real time. If the relative position is the target position, the drone is controlled to take pictures of the property area from different angles to obtain drone images. The maintenance status of the property area is determined based on the detection results obtained by integrating the drone images.
2. The method for data fusion and inspection of property projects based on UAV aerial photography according to claim 1, characterized in that, The design constraints based on the property area include: The first inequality is constructed using the drone's battery power, the inspected locations, and the power consumption per unit time; Construct a second inequality based on the elevation curves of the ridges surrounding the property area; A third inequality is constructed based on the tree canopy height of the property area.
3. The method for data fusion and inspection of property projects based on UAV aerial photography according to claim 2, characterized in that, Before combining the constraints and the multi-objective loss function, the following is also included: Design a multi-objective loss function based on the following formula: ； Among them, L final Let L represent the multi-objective loss function. cross L represents the cross-entropy loss function, which characterizes the difference between the predicted drone inspection probability distribution and the actual property area distribution; energy L represents the energy loss function, which characterizes the remaining battery power and the rate of battery consumption of the UAV; GIOU The target overlap loss function is used to characterize the degree of overlap between the UAV's flight area and the corresponding area of obstacles, including buildings, ridges, and forests.
4. The method for data fusion and inspection of property projects based on UAV aerial photography according to claim 3, characterized in that, The formula for calculating the cross-entropy loss function is as follows: ； Among them, y i This represents the label of the i-th inspection location of the drone. i =1, which means that the i-th inspection location is a real property area; if y i =0, which means the i-th inspection location is in another area; p i Let represent the probability that the drone will pass through the i-th inspection location, log represents the logarithmic function with the natural constant as the base, and N represents the total number of inspection locations.
5. The method for data fusion and inspection of property projects based on UAV aerial photography according to claim 3, characterized in that, The formula for calculating the energy loss function is as follows: ； Where A represents the drone's current battery level, exp represents the exponential function, k represents the energy loss factor (k > 0), and x represents the remaining flight distance. This represents the offset term caused by drag during the flight of the drone. It is positively correlated with the drag experienced by the drone.
6. The method for data fusion and inspection of property projects based on UAV aerial photography according to claim 3, characterized in that, The formula for calculating the target overlap loss function is as follows: ； ； Where IoU represents the intersection-union ratio loss function, and c represents the overlap coefficient. Represents the three-dimensional center distance. Indicates the prediction box. Represents the true bounding box. This represents the three-dimensional center distance between the predicted bounding box and the ground truth bounding box; This represents the intersection operation. This indicates the union operation.
7. The method for data fusion and inspection of property projects based on UAV aerial photography according to claim 3, characterized in that, The design of flight routes and parameters, combining constraints and multi-objective loss functions, includes: The first inequality is used to constrain the total number N of inspection locations in the cross-entropy loss function; The energy loss factor in the energy loss function is constrained by the second inequality. The overlap coefficient in the third inequality constraint is adopted.
8. The method for data fusion and inspection of property projects based on UAV aerial photography according to claim 7, characterized in that, The constraint of the first inequality on the total number N of inspection locations in the cross-entropy loss function includes: The first inequality is constructed based on the following formula: ； Where QUA represents the drone's battery level, N0 represents the inspected locations, inter represents the time interval between two adjacent inspected locations (inter is a fixed value), and PC represents the power consumption per unit time (QUA is expressed as a percentage). The first inequality is used as a constraint condition for the cross-entropy loss function.
9. The method for data fusion and inspection of property projects based on UAV aerial photography according to claim 1, characterized in that, The detection results fused from the drone images are used to determine the maintenance status of the property area, including: Determine the range of pixel values for the green area; Extract the green area from the drone image, which is an RGB image; Calculate the ratio of the green area to the entire drone image to obtain the green area percentage; The maintenance status of the property area is determined based on the percentage of green areas.
10. A property project data fusion and inspection device based on drone aerial photography, characterized in that, include: The constraint design module is used to obtain the property area and design constraints based on the property area. The UAV parameter design module is used to design flight routes and flight parameters by combining constraints and multi-objective loss functions; wherein, the flight parameters include the UAV dynamic flight altitude and the UAV dynamic flight speed; The drone inspection module is used to control the drone to perform inspections according to the flight route and flight parameters. The drone shooting module is used to monitor the relative position between the drone and the property area in real time. If the relative position is the target position, the drone is controlled to shoot the property area from different angles to obtain drone images. The detection result fusion module is used to fuse the detection results of the UAV images and determine the maintenance status of the property area based on the detection results.