Farmland area measurement method based on unmanned aerial vehicle

Abstract

The invention discloses a farmland area measurement method based on an unmanned aerial vehicle, and the method comprises the following steps: 1, carrying out the flight and data collection of the unmanned aerial vehicle: enabling the unmanned aerial vehicle to fly according to a preset path, and collecting the coordinate data of a farmland boundary point through a carried GPS module and a camera; step 2, boundary point data preprocessing: using a boundary point sampling algorithm to preprocess the collected boundary point data, the preprocessing of the boundary point data including removing abnormal points and supplementing areas with insufficient sampling density; 3, generating a closed polygon graph: processing the boundary point data to generate the closed polygon graph; 4, calculating the area: calculating the area of the closed polygon graph through a Gaussian area formula; 5, data transmission: transmitting the measurement result and the boundary point coordinate data to terminal equipment through a wireless communication module; and step 6, generating a visual report: generating a visual farmland area report in the terminal equipment.

Description

Technical Field

[0001] The invention belongs to the technical field of agricultural measurement, and in particular relates to a method for measuring farmland area based on an unmanned aerial vehicle. Background Art

[0002] Existing methods for measuring farmland area mainly rely on manual measurement or the use of traditional measuring equipment, such as acre meters and radars. These methods have the following problems in operation: low measurement efficiency, especially in large-scale, irregularly shaped farmland, manual measurement requires a lot of time and manpower; radar measurement accuracy is greatly affected by weather and terrain conditions, and errors are difficult to control; traditional methods are difficult to generate intuitive area reports, data processing is cumbersome, and it is difficult to meet the needs of modern agricultural management. Although some measurement methods have introduced advanced measurement equipment, such as RTK surveying instruments, they are complex to operate and require a high level of professionalism from the users.

[0003] Therefore, there is a need for an efficient and accurate method for measuring farmland area. Summary of the invention

[0004] The purpose of the present invention is to solve the technical problems of low measurement efficiency and complex data processing, and to provide a method for measuring farmland area based on drones.

[0005] In order to solve the above problems, the present invention is implemented according to the following technical solutions:

[0006] The method for measuring farmland area based on a drone of the present invention comprises the following steps:

[0007] Step 1: Drone flight and data collection: The drone flies along a preset path and collects coordinate data of farmland boundary points through the GPS module and camera on board;

[0008] Step 2: preprocessing the boundary point data: preprocessing the collected boundary point data using a boundary point sampling algorithm, wherein the preprocessing of the boundary point data includes removing abnormal points and supplementing areas with insufficient sampling density;

[0009] Step 3, generating a closed polygonal figure: processing the boundary point data to generate a closed polygonal figure;

[0010] Step 4, calculate the area: calculate the area of ​​the closed polygon using the Gaussian area formula. The area calculation formula is:

[0011]

[0012] Among them, (x i ,y i) represents the coordinates of the boundary point, A is the area, n is the total number of vertices of the farmland boundary, i represents the vertex index from 1 to n, is the addition term from vertex 1 to n-1, (x n y1-x1y n ) is the calculation item of the last point, the nth point and the first point;

[0013] Step 5: Data transmission: The measurement results and boundary point coordinate data are transmitted to the terminal device through the wireless communication module;

[0014] Step 6: Generate a visualization report: Generate a visualization report of the farmland area in the terminal device.

[0015] Preferably, in step 1, the drone is provided with a real-time terrain perception module to identify obstacles in the farmland and adjust the flight path. The terrain perception module includes a laser radar, an ultrasonic sensor and a camera. The obstacle identification formula is:

[0016]

[0017] Among them, D(x,y) represents the distance from the current position of the drone to the obstacle position, (x0,y0) represents the coordinate position of the obstacle, and d min is the safety distance. When the distance from the current position of the drone D(x,y) to the obstacle position (x0,y0) is less than the preset safety distance d min is considered an obstacle.

[0018] Preferably, in step 2, the boundary point sampling algorithm uses a Kalman filter algorithm to process the GPS coordinate data, and the formula is:

[0019] x k =x k-1 +K k (z k -Hx k-1 );

[0020] Among them, x k represents the current estimated value, z k is the observed value, K k is the Kalman gain, and H is the measurement matrix.

[0021] Preferably, in step 3, the generated closed polygonal figure reduces redundant vertices through a polygon vertex optimization algorithm, and the optimization rule is:

[0022]

[0023] Among them, θ i-1,i,i+1 represents the angle between the three vertices, ∈ is the optimization threshold, when cosθ i-1,i,i+1 >∈, remove vertex Pi .

[0024] Preferably, in step 4, based on the closed polygon generated in step 3, when the closed polygon is irregular, the Gaussian area formula and curve fitting technology are used to process the irregular boundary, and the equation of the fitting curve used is:

[0025] y=a n x n +a n-1 x n-1 +···+a1x+a0;

[0026] Among them, y represents the ordinate value on the fitting curve, x represents the abscissa value on the fitting curve, n represents the highest power of the fitting curve, [a0, a1, ..., a n ] are fitting coefficients, which are determined by the least squares method.

[0027] Preferably, in step five, the measurement results received by the terminal device include original coordinate data, boundary graphics and area calculation results.

[0028] Preferably, during the flight of the drone in step 1, the boundary points are image-verified by combining optical imaging technology with GPS coordinates, and the image verification matching formula is:

[0029]

[0030] Among them, S is the matching degree, I1 and I2 are the pixel matrices of the two images.

[0031] Preferably, in step 1 to step 4, after the drone completes a measurement, the drone replans the path to perform multiple flight measurements, and superimposes the multiple measurement results to calculate the final value, and the superposition formula is:

[0032]

[0033] Among them, A i is the area measured for the i-th time, and m is the number of measurements.

[0034] Preferably, in step six, the farmland area report includes the total farmland area and corresponding boundary graphics.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] The method for measuring farmland area based on drone of the present invention comprises the following steps: step 1, drone flight and data collection: the drone flies according to a preset path, and collects coordinate data of farmland boundary points through the GPS module and camera carried by the drone; step 2, boundary point data preprocessing: the collected boundary point data are preprocessed by using the boundary point sampling algorithm, and the preprocessing of the boundary point data includes removing abnormal points and supplementing areas with insufficient sampling density; step 3, generating closed polygonal figures: the boundary point data are processed and used to generate closed polygonal figures; step 4, calculating the area: the area of ​​the closed polygonal figure is calculated by using the Gaussian area formula, and the area calculation formula is: Among them, (x i ,y i ) represents the coordinates of the boundary point, A is the area, n is the total number of vertices of the farmland boundary, i represents the vertex index from 1 to n, is the addition term from vertex 1 to n-1, (x n y1-x1y n ) is the calculation item of the nth point and the 1st point of the last item; Step 5, data transmission: the measurement results and boundary point coordinate data are transmitted to the terminal device through the wireless communication module; Step 6, generate a visualization report: generate a visualized farmland area report in the terminal device.

[0037] The farmland area measurement method based on drones in the present invention collects farmland boundary point data by drone flight, removes abnormal points and supplements insufficient areas by algorithm preprocessing, thereby generating accurate closed polygon graphics. The Gaussian area formula is used to calculate the polygon area to ensure accurate and efficient results. The measurement data is transmitted to the terminal device through a wireless module, and a visual report containing the farmland area and boundary graphics is generated. The method is easy to operate, accurate in data, and highly efficient. It is suitable for various farmland shapes and agricultural management needs, and provides technical support for modern agriculture. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The specific embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings, wherein:

[0039] Figure 1 The present invention is a flow chart of a method for measuring farmland area based on unmanned aerial vehicle. DETAILED DESCRIPTION

[0040] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0041] This embodiment uses a method for measuring farmland area based on a drone as an example for illustration. It should be understood that the present invention should not be limited to this method, and other similar methods / structures are within the protection scope of the present invention.

[0042] like Figure 1 As shown, a method for measuring farmland area based on a drone of the present invention comprises the following steps:

[0043] Step 1: Drone flight and data collection: The drone flies along a preset path and collects coordinate data of farmland boundary points through the GPS module and camera on board;

[0044] Step 2: preprocessing of boundary point data: using the boundary point sampling algorithm, preprocessing the collected boundary point data, preprocessing the boundary point data includes removing abnormal points and supplementing areas with insufficient sampling density;

[0045] Step 3, generating a closed polygonal figure: processing the boundary point data to generate a closed polygonal figure;

[0046] Step 4: Calculate the area: Calculate the area of ​​the closed polygon using the Gaussian area formula. The area calculation formula is:

[0047]

[0048] Among them, (x i ,y i ) represents the coordinates of the boundary point, A is the area, n is the total number of vertices of the farmland boundary, i represents the vertex index from 1 to n, is the addition term from vertex 1 to n-1, (x n y1-x1y n ) is the calculation item of the last point, the nth point and the first point;

[0049] Step 5: Data transmission: The measurement results and boundary point coordinate data are transmitted to the terminal device through the wireless communication module;

[0050] Step 6: Generate a visualization report: Generate a visualization report of the farmland area in the terminal device.

[0051] In step 1 of the present invention: the drone flies according to a preset flight path, and the flight path can be planned according to the actual shape and size of the farmland. The drone is equipped with a high-precision GPS module and a high-definition camera. During the flight, the GPS module collects the coordinate data of the farmland boundary in real time, and the camera helps determine the boundary of the farmland through image recognition technology to provide more accurate boundary information. During the data collection process, the drone will automatically adjust the flight altitude and angle according to the real-time GPS data and image analysis to ensure the collection of boundary point data.

[0052] The boundary point data collected in step 2 may contain some abnormal points, for example, erroneous points caused by GPS signal interference or image misrecognition.

[0053] Therefore, the boundary point sampling algorithm is used to preprocess the data. The specific methods include:

[0054] Remove outliers: Identify and remove outliers that deviate too much from surrounding data points by setting thresholds to ensure data accuracy.

[0055] In step 3, a closed polygon is generated by using the processed boundary point data and polygon generation algorithms such as Delaunay triangulation and convex hull algorithm to ensure that all collected boundary points are connected into a complete closed loop. The purpose of this step is to accurately represent the boundary of the farmland and avoid open or irregular polygons. The polygon generation process can use boundary point sorting, angle calculation and other technologies to ensure that the generated polygon can completely cover the farmland area.

[0056] In the fourth implementation step, the Gauss area formula is used to process the closed polygon and calculate its area. The formula is as follows: i ,y i are the coordinates of the vertices of the polygon.

[0057]

[0058] Boundary point data input:

[0059] Let the coordinates of the boundary points of the polygon be (x1,y1),(x2,y2),...,(x n ,y n ), these coordinates are collected and pre-processed by drones.

[0060] Polygon closure processing:

[0061] To ensure that the boundary points form a closed polygon, connect the nth vertex to the first vertex:

[0062] (x n+1 ,y n+1 )=(x1,y1);

[0063] n is the total number of boundary points, (x n+1 ,y n+1 ) are the coordinates of the first point, i.e. the starting point of the closed polygon;

[0064] Definition of n: n is the total number of vertices of the farmland boundary.

[0065] Definition of i: In the formula, i represents the vertex index from 1 to n.

[0066] Gaussian area formula to calculate area:

[0067] The area calculation formula is based on the vertex coordinates of the polygon and is performed using the following formula:

[0068] is the addition term from vertex 1 to n-1, (x n y1-x1y n ) is the calculation item of the last point, the nth point and the first point;

[0069]

[0070] Where (x i ,y i ) represents the coordinates of the i-th vertex, x i+1 ,y i+1 is the coordinate of the i+1th vertex and (x1, y1) when i=n, ​​x i+1 =x1,y i+1 =y1.

[0071] This formula directly calculates the directed area enclosed by vertices and is applicable to any polygon.

[0072] Note that the coordinates must be given in clockwise or counterclockwise order. Polygons cannot intersect themselves.

[0073] During the measurement process of step 5, the area data and boundary point coordinate data collected in real time are transmitted to the remote terminal device through the wireless communication module. The wireless communication module uses standard wireless transmission protocols such as Wi-Fi, 4G / 5G and LoRa to ensure the stability and accuracy of data transmission. The terminal device can be a PC, mobile device or dedicated farmland management terminal, and the received data can be displayed to the operator in real time for subsequent processing and analysis.

[0074] Step 6: Generate a farmland area report on the terminal device through data processing and analysis. The report includes:

[0075] Total farmland area: Displays the total area of ​​farmland based on the calculation results.

[0076] Boundary graphics: Graphics showing the boundaries of farmland for easy visual viewing by operators. Graphics can be displayed in the form of maps or CAD drawings, and interactive operations such as zooming in and out are provided.

[0077] Export function: The report can be exported in multiple formats, including PDF, GeoJSON, etc., so that operators can further analyze or print the report.

[0078] In some specific embodiments,

[0079] Drone path planning: The drone flight path can be dynamically adjusted according to the specific shape of the farmland to ensure coverage of the entire farmland area and avoid omissions.

[0080] Precision control: In step 2, the boundary point sampling algorithm can dynamically adjust the sampling frequency according to the GPS signal strength and the number of satellites to ensure that the collected coordinate data has sufficient accuracy.

[0081] Multiple measurements and data fusion: In order to improve measurement accuracy, multiple flight measurements can be performed at different times or at different angles, and data fusion technology can be used to perform weighted averaging of multiple measurement results to improve the accuracy of the final result.

[0082] In a preferred embodiment, in step 1, the drone uses a real-time terrain perception module to identify obstacles in the farmland and adjust the flight path. The obstacle identification formula is:

[0083]

[0084] Among them, D(x,y) represents the distance from the current position of the drone to the obstacle position, d min For a safe distance.

[0085] In this embodiment, the drone is equipped with a real-time terrain perception module, which includes sensors such as laser radar, ultrasonic sensors and cameras, for detecting obstacles in the farmland. The drone can obtain information about the surrounding environment on its flight path in real time and detect and identify obstacles.

[0086] Obstacle identification: The real-time terrain perception module of the drone can obtain the coordinate data of its current position (x, y) and calculate the distance to the surrounding obstacles based on the feedback from the sensor. Obstacle identification is performed using the following formula:

[0087]

[0088] Where D(x,y) represents the distance from the current position of the drone to the obstacle, (x0,y0) is the coordinate of the obstacle, and d min is the preset safety distance. If the distance between the drone and the obstacle is less than the safety distance d min, then the location is considered an obstacle and the drone will perform an avoidance operation. During the flight, when the drone identifies an obstacle, it will adjust the flight path in real time based on the relationship between the obstacle's location and the drone's current flight path. Based on the location of the obstacle, the path planning algorithm is used to calculate a new flight path to ensure safety during the flight. If the obstacle is above the flight path, the flight altitude will be adjusted appropriately to avoid the obstacle. To ensure a smooth flight, the adjusted path will be smoothed to avoid sharp turns or excessive flight attitude changes. Throughout the flight, the real-time terrain perception module continuously monitors the environment around the drone. Whenever a new obstacle is detected or the location of an existing obstacle changes, the system automatically recalculates the flight path and adjusts the flight direction or altitude in time to ensure that the drone can avoid all obstacles.

[0089] The obstacle recognition and avoidance strategy in this embodiment ensures that the drone can fly safely in a complex farmland environment, avoiding obstacles such as trees, buildings, and field equipment, thereby achieving accurate farmland area measurement tasks. This solution can improve the safety of flight missions and avoid flight accidents or data collection failures caused by obstacle interference.

[0090] In order to improve the accuracy of GPS coordinate data during UAV farmland measurement, the present invention provides a preferred embodiment. In step 2, the boundary point sampling algorithm uses the Kalman filter algorithm to process the GPS coordinate data. The formula is:

[0091] x k =x k-1 +K k (z k -Hx k-1 );

[0092] Among them, x k represents the current estimated value, z k is the observed value, K k is the Kalman gain, and H is the measurement matrix.

[0093] In this embodiment, the Kalman filter algorithm is used to process the boundary point data. Specifically, in step 2, the boundary point coordinate data collected by the GPS module carried by the drone will be pre-processed to reduce the error caused by GPS noise.

[0094] The update formula of Kalman filter is as follows:

[0095] x k =x k-1 +K k (z k -Hx k-1 );

[0096] in:

[0097] x k Indicates the estimated value at the current moment, that is, the corrected value of the current GPS coordinates;

[0098] z k is the observation value at the current moment, that is, the GPS data currently received by the drone;

[0099] K k is the Kalman gain, which determines the weighted ratio between the observed value and the predicted value;

[0100] H is the measurement matrix, which represents the observation model.

[0101] In this method, the Kalman filter combines the previous estimate and the current observation, and optimizes the accuracy of the GPS data by calculating the Kalman gain, thereby reducing the impact of errors on the measurement results. The boundary point data processed by this algorithm will be used for subsequent polygon generation and area calculation.

[0102] In an optional embodiment, when a drone flies to an area to measure farmland, the received GPS data may contain certain noise points, resulting in inaccurate coordinates of boundary points. To correct these errors, the Kalman filter algorithm updates the estimated value of each coordinate point in real time to reduce the error and adjust the Kalman gain K k The value of z k For the estimated value x k The degree of impact is gradually optimized.

[0103] For example, the following observations are collected by the GPS module during the flight of the drone:

[0104] z1=(x1,y1),z2=(x2,y2),···,z k (x k ,y k ) and updates the coordinate value x of each boundary point through the Kalman filter k To obtain farmland boundary data.

[0105] The algorithm dynamically adjusts the Kalman gain K according to the difference between the observed value and the predicted value. k , realize the update of boundary point coordinates and improve the measurement results.

[0106] In order to reduce redundant vertices in the farmland boundary polygon, the present invention provides a preferred embodiment, in which the closed polygon figure generated in step 3 is reduced in redundant vertices by a polygon vertex optimization algorithm, and the optimization rule is:

[0107]

[0108] Among them, θ i-1,i,i+1 represents the angle between the three vertices, ∈ is the optimization threshold, when cosθ i-1,i,i+1 >∈, remove vertex P i .

[0109] Specifically, the polygon vertex coordinates of the farmland boundary are P1, P2, ..., P n , each vertex P i =(x i ,y i ) represents a coordinate point. To optimize these vertices, we need to check every three consecutive vertices P i-1 ,P i ,P i+1 The angle θ between i-1,i,i+1 , use the cosine theorem to calculate the angle, and use the following formula to determine whether to remove vertex P i :

[0110]

[0111] in:

[0112] P i =(x i ,y i ) is the current vertex coordinate;

[0113] ||P i-1 -P i || and ||P i+1 -P i || respectively represent the i To its previous vertex P i-1 and the next vertex P i+1 The modulus of the vector, which means from the current vertex P i Distances to adjacent vertices;

[0114] (P i-1 -P i )·(P i+1 -P i ) is the product of the lengths of two vectors and the cosine of the angle between them.

[0115] If the calculated cosine value of the angle is greater than the preset optimization threshold ∈, that is, cosθ i-1,i,i+1 >∈, then vertex P i It contributes little to the shape of the polygon and can be removed.

[0116] More specifically, for each set of three consecutive vertices P i-1 ,P i ,P i+1, calculate the angle cosine value. Determine whether the angle cosine value is greater than the threshold ∈, if so, remove the middle vertex P i Repeat the above steps until all redundant vertices are removed and an optimized polygon is obtained.

[0117] In this way, the present invention has the following advantages:

[0118] 1. The polygons of farmland boundaries will be simplified to reduce redundant points.

[0119] 2. After reducing redundant vertices, the generated polygons are simpler and can more accurately reflect the actual boundaries of the farmland.

[0120] In a preferred embodiment, in step 4, based on the closed polygon generated in step 3, when the boundary shape is relatively irregular, the Gaussian area formula and curve fitting technology are used to process the irregular boundary, and the equation of the fitting curve used is: y = a n x n +a n-1 x n-1 +···+a1x+a0;

[0121] Among them, [a0,a1,...,a n ] are fitting coefficients, which are determined by the least squares method.

[0122] In this embodiment, based on the farmland boundary point data collected by the drone, a closed polygon is first generated through step three. On this basis, when the boundary shape of the farmland is relatively irregular, the Gaussian area formula and curve fitting technology are used to process the irregular boundary. The specific method is as follows:

[0123] Boundary fitting: For the irregular part of the boundary data, firstly process it in segments. Use the Gaussian area formula to divide the irregular boundary into several small segments, and then use the curve fitting method to process each segment. The equation of the fitting curve is:

[0124] y=a n x n +a n-1 x n-1 +···+a1x+a0;

[0125] Among them, [a0,a1,...,a n ] is the fitting coefficient, and the fitting coefficient is optimized by the least squares method to ensure the best match between the fitting curve and the actual boundary data.

[0126] Specifically, the fitting coefficients are calculated by the least squares method to minimize the error between the fitting curve and the actual boundary points. The goal of the least squares method is to adjust the coefficients [a0, a1, ..., an ], so that the fitting curve fits the boundary point data and reduces the error.

[0127] In step 4, the area of ​​the polygon is calculated using the Gaussian area formula. If the farmland boundary is relatively irregular, a curve fitting technique can provide a more accurate area calculation method after generating a closed polygon, so this embodiment supplements and improves step 4.

[0128] In step 2, the Kalman filter algorithm is used to process the GPS data, which helps to improve the accuracy and stability of the boundary point data. The curve fitting technology in this embodiment further optimizes and refines the representation of the farmland boundary based on the high-precision boundary point data, making the final area calculation more accurate.

[0129] This embodiment of the present invention has the following advantages:

[0130] 1. Through piecewise approximation and curve fitting, it can handle irregular boundaries and reduce the calculation errors caused by irregular boundaries.

[0131] 2. The fitting curve transforms the irregular boundary into a simple mathematical expression, reducing the complexity of subsequent area calculation.

[0132] 3. It can handle complex farmland boundary shapes, especially when encountering complex bends or curves, the fitting method can still maintain high accuracy and stability.

[0133] 4. Through least squares optimization, the fitting results can minimize noise and errors.

[0134] 5. The fitting curve of each segment will be combined with the fitting results of the adjacent segments to finally generate a smooth fitting curve that can effectively describe the boundary shape of the entire farmland. This curve can better represent the actual contour of the boundary and improve the accuracy of area calculation.

[0135] In step 5, the results received by the terminal device include original coordinate data, boundary graphics and area calculation results.

[0136] In this embodiment, based on the farmland boundary data collected by the drone, the data received by the terminal device includes original coordinate data, boundary graphics and calculated area values. After the drone completes the measurement, the measurement data is transmitted to the terminal device in real time through the wireless communication module.

[0137] The transmitted data includes: the coordinates of the farmland boundary points (x1, y1), (x2, y2), ..., (x n ,y n). Through the polygon generation algorithm, the terminal device generates and stores the boundary graphics of the farmland, which are expressed as closed polygons. The area of ​​the farmland is calculated by the area calculation algorithm in step (iv). The received data can be used for subsequent analysis. The operator can perform visual analysis on the data as needed, view the boundary graphics of the farmland, the calculated area value, or perform further statistics and processing. All received data (including original coordinates, boundary graphics and area values) will be stored in the database of the terminal device. The operator can query and trace back historical measurement data at any time.

[0138] In this embodiment, the present invention has the following advantages:

[0139] 1. Can store relevant data (original coordinates, boundary graphics and area values) for subsequent use and management.

[0140] 2. The data can be used for subsequent analysis to support further calculations and optimization.

[0141] In some optional embodiments, the visualized farmland area report generated in step six may support export in multiple formats, including PDF and GeoJSON file formats.

[0142] In this embodiment, the farmland area report generated by the terminal device can support exporting to multiple formats, including geographic information formats such as PDF and GeoJSON. When the operator views the measurement results through the terminal device, the system will automatically generate a farmland area report.

[0143] Specifically, the report includes the following: Based on the calculated area value, the report will show the total area of ​​the farmland; the report will include a visual graphic of the farmland boundary to show the specific outline of the farmland; the report will include the original coordinate data for the operator to verify and review.

[0144] Operators can choose to export reports to different formats for further processing and sharing; operators can export farmland boundary graphics and related data to GeoJSON format, which is a standard geographic information format that is compatible with other geographic information system (GIS) software. The GeoJSON format allows operators to further process farmland data, perform spatial analysis, or integrate data into more complex map applications. Exported reports and data files can be shared via wireless networks, USB devices, etc.

[0145] In step (1), during the flight of the drone, the boundary points are image-verified by combining optical imaging technology with GPS coordinates. The image verification matching formula is:

[0146]

[0147] Among them, S is the matching degree, I1 and I2 are the pixel matrices of the two images.

[0148] In this embodiment, in order to verify the accuracy of the boundary points, the collected images are compared with the stored reference images, and an image matching algorithm is used to ensure the accuracy of the boundary point data. Boundary point collection and image verification, when the drone is flying, the camera carried will collect image data of the farmland boundary in real time, and record the coordinates of each collection point through the GPS module.

[0149] The following image verification matching formula is used to calculate the image matching degree:

[0150] Match Among them, I1 and I2 represent the pixel matrices of the two images respectively. The similarity of the two images is determined by calculating the matching degree S. If the matching degree reaches the set threshold, the position of the boundary point is considered to be accurate.

[0151] Numerator ∑I1·I2: represents the product of the corresponding pixels of the two images I1 and I2, reflecting the consistency of the images at the pixel points.

[0152] Denominator It normalizes the pixel values ​​of the two images to ensure that the result is in the range of [-1, 1][-1, 1][-1, 1], eliminating the impact of image brightness or contrast differences.

[0153] Result interpretation:

[0154] When S=1: it means the two images are exactly the same.

[0155] When S=0: it means that the two images are completely unrelated.

[0156] When S=-1: it means that the two images are negatively correlated.

[0157] In a specific embodiment, if the image verification result shows a low matching degree, the system will prompt to re-collect the image data of the point, or adjust the collection path to ensure the accuracy of the boundary point.

[0158] The image verification in the present invention has the following advantages:

[0159] 1. Image verification through optical imaging technology combined with GPS coordinates can better improve the accuracy of boundary point data.

[0160] 2. Automatically perform image matching, reduce human intervention, and improve data processing efficiency.

[0161] Step 1 to Step 4 After the drone completes a measurement, the drone can re-plan the path to perform multiple flight measurements and superimpose multiple measurement results to calculate the final value. The superposition formula is:

[0162]

[0163] Among them, A i is the area measured for the i-th time, and m is the number of measurements.

[0164] In this embodiment, after the first measurement is completed, the drone will re-plan the flight path according to the measurement data and perform multiple flight measurements. After each flight, the system will recalculate and store the farmland area A measured each time. i .

[0165] The final area value of multiple measurements is calculated by superposition formula:

[0166]

[0167] A i is the farmland area value obtained by each independent measurement. Each measurement may be affected by some factors (such as flight path, GPS accuracy, environmental factors, etc.), so the results of each measurement will be different; m is the number of measurements; since each measurement will have a certain error, in order to obtain a more accurate farmland area, the final area A final It is obtained by combining multiple measurement results A1, A2, ..., A m Find the average.

[0168] In the present invention, multiple measurements have the following advantages:

[0169] 1. Through multiple measurements, accidental errors can be effectively reduced and the accuracy of area measurement can be improved.

[0170] 2. The error of each measurement will be optimized through average calculation to obtain more accurate farmland area data.

[0171] Preferably, in step six, the farmland area report includes the total farmland area and corresponding boundary graphics.

[0172] In this embodiment,

[0173] Furthermore, after the drone completes the flight mission and transmits the measurement results, the data received by the terminal device includes the original boundary point coordinates, boundary graphic data, and area calculation results. The terminal device processes the received data and first converts the boundary point coordinates into corresponding boundary graphics. Based on the farmland area calculated in step four, the terminal device will calculate the total area of ​​the farmland based on the boundary graphic data of the closed polygon and display the result in the report. The total area is the actual farmland area in the measurement area, and the unit can be set to square meters, mu or hectares as needed. Using the boundary point coordinates and graphics generation algorithm, the terminal device draws the boundary graphics of the farmland. The graphics can be displayed in different ways,

[0174] For example:

[0175] A two-dimensional map of the farmland is displayed on the screen of the terminal device, with the boundaries indicated by clear lines;

[0176] Generate high-resolution images of farmland boundaries and support zooming in and out so that users can view detailed information of different areas.

[0177] If necessary, a three-dimensional map view can be generated to show the undulations and boundaries of the farmland through changes in color and shading.

[0178] The final farmland area report will include the following:

[0179] Total farmland area: Displays the total area of ​​the measured area, and the unit is automatically converted according to user needs (such as square meters, mu, acres, etc.).

[0180] Boundary graphics: The external outline of the farmland is displayed through the farmland boundary graphics, which may include marked coordinates, area distribution and other information.

[0181] Additional information: If necessary, you can also add additional information, such as the accuracy of the measurement, the measurement time, the algorithm parameters used, etc.

[0182] In summary, the present invention provides a method for measuring farmland area based on drones, which realizes efficient and accurate measurement of farmland area by combining GPS positioning, image acquisition, boundary point processing, polygon generation algorithm and area calculation technology. The method supports multiple measurements and data fusion. The drone has the ability of dynamic path adjustment and obstacle avoidance during flight, and is suitable for a variety of complex terrain scenes. The management and application of measurement data are facilitated by real-time data transmission and generation of visual reports by terminal devices.

[0183] The working principle of the farmland area measurement method based on drone described in the present invention is: drone flight and data collection, the drone flies according to the preset path, equipped with a high-precision GPS module and a high-definition camera, collects the coordinate data and boundary images of the farmland boundary points in real time, and verifies the collected boundary points by combining the GPS coordinates through optical imaging technology to ensure the accuracy of the boundary point data. During the flight, the drone will adjust the flight altitude and angle according to the real-time terrain perception data to ensure complete coverage of the farmland area. Boundary point data preprocessing, the collected boundary point data is preprocessed by the boundary point sampling algorithm, including removing abnormal points and supplementing areas with insufficient sampling density. The abnormal points are identified and deleted by setting thresholds, and the supplementary sampling is achieved by dynamically adjusting the sampling frequency. Generate closed polygon graphics. The polygon generation algorithm is used to generate closed polygons for the preprocessed boundary point data. This step ensures that the boundary points form a complete closed loop and accurately represent the boundary shape of the farmland. Calculate the area of ​​the graphics, and use the Gaussian area formula to calculate the area of ​​the closed polygon. The area formula is based on the coordinates of the polygon vertices and is applicable to polygons of any shape. For irregular boundaries, the calculation results are further optimized by curve fitting technology to improve the area calculation accuracy. Data transmission and storage: The real-time measured area data and boundary point coordinate data are transmitted to the terminal device through the wireless communication module, and the data is stored in the database for subsequent analysis and management. Generate a visual report: The terminal device generates a visual report of the farmland area based on the received data. The report includes the total area of ​​the farmland, boundary graphics and original coordinate data. Multiple measurements and data fusion: In order to improve the measurement accuracy, the drone can perform multiple measurements at different times or different flight angles, and use data fusion technology to perform weighted averaging of multiple measurement results to finally generate more accurate measurement results. Obstacle identification and avoidance strategy: The drone detects obstacles in the farmland through a real-time terrain perception module (including lidar, ultrasonic sensor and camera) during flight. When an obstacle conflicts with the flight path, the drone adjusts the flight direction or altitude in real time through the path planning algorithm to ensure flight safety.

[0184] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Therefore, any modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the technical solution of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A method for measuring farmland area based on drone, characterized in that: The following steps are involved: Step 1: Drone flight and data collection: The drone flies along a preset path and collects coordinate data of farmland boundary points through the GPS module and camera on board; Step 2: preprocessing the boundary point data: preprocessing the collected boundary point data using a boundary point sampling algorithm, wherein the preprocessing of the boundary point data includes removing abnormal points and supplementing areas with insufficient sampling density; Step 3, generating a closed polygonal figure: processing the boundary point data to generate a closed polygonal figure; Step 4, calculate the area: calculate the area of ​​the closed polygon using the Gaussian area formula. The area calculation formula is: Among them, (x i ,y i ) represents the coordinates of the boundary point, A is the area, n is the total number of vertices of the farmland boundary, i represents the vertex index from 1 to n, is the addition term from vertex 1 to n-1, (x n y1-x1y n ) is the calculation item of the last point, the nth point and the first point; Step 5: Data transmission: The measurement results and boundary point coordinate data are transmitted to the terminal device through the wireless communication module; Step 6: Generate a visualization report: Generate a visualization report of the farmland area in the terminal device.

2. The method for measuring farmland area based on drone according to claim 1, characterized in that: In step 1, the drone is equipped with a real-time terrain perception module to identify obstacles in the farmland and adjust the flight path. The terrain perception module includes a laser radar, an ultrasonic sensor, and a camera. The formula for identifying obstacles is: Among them, D(x,y) represents the distance from the current position of the drone to the obstacle position, (x0,y0) represents the coordinate position of the obstacle, and d min is the safety distance. When the distance from the current position of the drone D(x,y) to the obstacle position (x0,y0) is less than the preset safety distance d min is considered an obstacle.

3. The method for measuring farmland area based on drone according to claim 1, characterized in that: In step 2, the boundary point sampling algorithm uses the Kalman filter algorithm to process the GPS coordinate data, and the formula is: x k =x k-1 +K k (z k -Hx k-1 ); Among them, x k represents the current estimated value, z k is the observed value, K k is the Kalman gain, and H is the measurement matrix.

4. The method for measuring farmland area based on drone according to claim 1, characterized in that: In step 3, the generated closed polygon graphics are optimized by polygon vertex optimization algorithm to reduce redundant vertices. The optimization rules are: Among them, θ i-1,i,i+1 represents the angle between the three vertices, ∈ is the optimization threshold, when cosθ i-1,i,i+1 >∈, remove vertex P i .

5. The method for measuring farmland area based on drone according to claim 1, characterized in that: In step 4, based on the closed polygon generated in step 3, when the closed polygon is irregular, the Gaussian area formula and curve fitting technology are used to process the irregular boundary. The equation of the fitting curve used is: y=a n x n +a n-1 x n-1 +···+a1x+a0; Among them, y represents the ordinate value on the fitting curve, x represents the abscissa value on the fitting curve, n represents the highest power of the fitting curve, [a0, a1, ..., a n ] are fitting coefficients, which are determined by the least squares method.

6. The method for measuring farmland area based on drone according to claim 1, characterized in that: In step five, the measurement results received by the terminal device include original coordinate data, boundary graphics and area calculation results.

7. The method for measuring farmland area based on drone according to claim 1, characterized in that: In the process of the drone's flight in step 1, the boundary points are image-verified by combining optical imaging technology with GPS coordinates. The image verification matching formula is: Among them, S is the matching degree, I1 and I2 are the pixel matrices of the two images.

8. The method for measuring farmland area based on drone according to claim 1, characterized in that: In steps 1 to 4, after the drone completes a measurement, the drone replans the path to perform multiple flight measurements, and superimposes the multiple measurement results to calculate the final value. The superposition formula is: Among them, A i is the area measured for the i-th time, and m is the number of measurements.

9. The method for measuring farmland area based on drone according to claim 1, characterized in that: In step six, the farmland area report includes the total farmland area and corresponding boundary graphics.

Images