A method for verifying farmland by an unmanned aerial vehicle

Abstract

The present invention relates to the technical field of farmland verification, and in particular to a method for verifying farmland by an unmanned aerial vehicle, which includes: obtaining a navigation route; flying to the starting point of the navigation route and navigating along the navigation route from the starting point; during the navigation along the navigation route, collecting video frame images of the farmland area to be inspected in real time; splicing the collected multiple video frame images into an area image of the farmland area to be inspected; after the area image is spliced, analyzing the area image to obtain the crop types and crop areas in the area image; after the area image is spliced, returning from the end position to the starting point along the navigation route in a second mode and collecting the farmland distances in the vertical direction from each crop area in real time; judging the growth conditions of the crops in each crop area according to the farmland distances of each crop area and the initial farmland distances of each crop area. The present invention has the effect of understanding the growth conditions of the crops in the farmland.

Description

Technical Field

[0001] The present invention relates to the technical field of farmland verification, and in particular to a method for verifying farmland by using an unmanned aerial vehicle (UAV). Background Art

[0002] Farmland refers to the land for agricultural production; cultivated fields. In order to promote the rational utilization and planning of farmland, the Ministry of Land and Resources often needs to verify farmland to judge the basic situation of the farmland. A UAV is a flying device controlled by a remote control or other terminals. Cameras, microphones and other devices can be installed on the UAV to collect images and sounds of the places passed by during flight.

[0003] With the development of UAVs, they have gradually been applied to farmland, but generally they only stay in the simple stage of shooting and sampling, and it is impossible to understand the growth situation of crops in the farmland. Summary of the Invention

[0004] In order to understand the growth situation of crops in farmland, the present invention provides a method for verifying farmland by using a UAV.

[0005] The first invention object of the present invention is achieved through the following technical solutions:

[0006] A method for verifying farmland by using a UAV includes:

[0007] Obtaining a navigation route, where the navigation route is horizontal;

[0008] Flying to the starting point of the navigation route and navigating along the navigation route from the starting point;

[0009] During the navigation along the navigation route, video frame images of the farmland area to be inspected are collected in real time;

[0010] Stitching the collected multiple video frame images into an area image of the farmland area to be inspected;

[0011] After the area image is stitched, analyzing the area image to obtain the crop types and crop areas in the area image;

[0012] After the area image is stitched, returning from the end position to the starting point along the navigation route in a second mode, and during the return process, the farmland distances in the vertical direction from the UAV to each crop area are collected in real time by a radar arranged at the bottom of the UAV. Among them, the UAV navigating along the navigation route from the starting point is in a first mode, and the speed at any point between the end point and the starting point in the second mode is the same as the speed at any point between the end point and the starting point in the first mode;

[0013] Judge the crop growth conditions of each crop area based on the farmland distance of each crop area and the initial farmland distance of each crop area.

[0014] By adopting the above technical solution, after the drone obtains a horizontal navigation route, it flies to the starting point of the navigation route and sails along the navigation route from the starting point. During the navigation along the navigation route, it collects video frame images of the farmland area to be inspected in real time, stitches the collected multiple video frame images into an area image of the farmland area to be inspected. After the area image stitching is completed, analyze the area image to obtain the crop types and crop areas in the area image. And after the area image stitching is completed, return from the end position to the starting point along the navigation route in the second mode, and during the return process, use the radar set at the bottom of the drone to collect the farmland distance in the vertical direction from the drone to each crop area in real time. Among them, the drone sails along the navigation route from the starting point in the first mode, and the speed at any point between the end point and the starting point in the second mode is the same as the speed at any point between the end point and the starting point in the first mode. Then judge the crop growth conditions of each crop area according to the farmland distance of each crop area and the initial farmland distance of each crop area, so as to judge the crop growth conditions of each crop area through the crop growth height of each crop area.

[0015] In a preferred example of the present invention, it can be further configured as follows: the navigation route comes from the terminal. The terminal is built-in with a preset map system, and a three-dimensional farmland map model is pre-stored in the map system. The model coordinates of the three-dimensional farmland map model are corresponding and associated with the real farmland position coordinates. The terminal provides a selection function for the verification personnel to input instructions on the terminal to select some or all areas in the three-dimensional farmland map model; the verification personnel input instructions to select and determine an area on the three-dimensional farmland map model, the terminal generates corresponding information of the area to be inspected, generates a navigation route according to the information of the area to be inspected, and sends the navigation route to the drone.

[0016] By adopting the above technical solution, after the verification personnel select an area in the three-dimensional farmland map model in the terminal, the terminal can generate a navigation route and send it to the drone.

[0017] In a preferred example of the present invention, it can be further configured as follows: generating a navigation route according to the information of the area to be inspected includes the following steps:

[0018] Input the information of the area to be inspected into a pre-trained model, and perform inference through the model to obtain a navigation route;

[0019] The model is trained in the following way:

[0020] Perform annotation processing on each inspection area information sample in the inspection area information sample training set to mark the navigation route in each inspection area information sample. Each inspection area information sample includes inspection area information, and the navigation route is associated with all or part of the information in the inspection area information sample; and train the neural network through the inspection area information sample training set that has undergone annotation processing to obtain a model;

[0021] Among them, the inspection area information sample and the corresponding marked navigation route are obtained by the verification personnel through experiments according to the actual situation. The verification personnel conduct a large number of experiments and select the navigation routes corresponding to different inspection area information that meet the conditions.

[0022] By adopting the above technical solution, the method of obtaining the navigation route through model inference is more accurate, and as the number of inspection area information samples increases, the inference result is more accurate.

[0023] In a preferred example of the present invention, it can be further configured that: the analysis of the area image includes:

[0024] Compare the area image with each pre-shot crop image to obtain the crop types in the area image. Take the area formed by the crop image cluster as the crop area to obtain the crop types and crop areas in the area image.

[0025] By adopting the above technical solution, the crop types and crop areas in the stitched area image can be obtained.

[0026] In a preferred example of the present invention, it can be further configured that: during the return process, the radar set at the bottom of the drone is used to collect the farmland distances in the vertical direction from the drone to each crop area, including:

[0027] Reverse the frames of all video frame images at the original speed from the last frame to the first frame. Take several points with specific pixel coordinates in each video frame image, map them in the stitched area image, and display the positions in the area image through the mapped coordinates, so as to obtain the crop areas passed by the drone currently. And associate the real-time collected distances with each crop area to obtain the farmland distances in the vertical direction from the drone to each crop area.

[0028] By adopting the above technical solution, through the method of reversing the frames at the original speed, during the process of the drone returning to the starting point according to the navigation route in the second mode, the positions in the corresponding area image can be obtained in real time, so as to associate the real-time collected distances with each crop area to obtain the farmland distances in the vertical direction from the drone to each crop area.

[0029] In a preferred example, the present invention can be further configured such that the crop height in the crop area is the difference between the initial farmland distance and the farmland distance in the crop area.

[0030] In a preferred example, the present invention can be further configured such that the crop growth conditions may include the crop planting time, the remaining ripening time, and the current growth cycle.

[0031] By adopting the above technical solution, it is possible to comprehensively know the crop production situation.

[0032] In summary, the present invention includes at least one of the following beneficial technical effects:

[0033] 1. After the drone obtains a horizontal flight route, it flies to the starting point of the flight route and sails along the flight route from the starting point. During the process of sailing along the flight route, it real-time collects video frame images of the farmland area to be inspected, stitches the collected multiple video frame images into an area image of the farmland area to be inspected. After the area image stitching is completed, the area image is analyzed to obtain the crop types and crop areas in the area image. And after the area image stitching is completed, it returns from the end position to the starting point along the flight route in the second mode, and during the return process, it real-time collects the farmland distance in the vertical direction from each crop area through the radar set at the bottom of the drone. Among them, the drone sails along the flight route from the starting point in the first mode, and the speed at any point between the end point and the starting point in the second mode is the same as the speed at any point between the end point and the starting point in the first mode. Then, according to the farmland distance of each crop area and the initial farmland distance of each crop area, the crop growth conditions of each crop area are judged, so that the crop growth conditions of each crop area can be judged through the crop growth height of each crop area;

[0034] 2. The method of obtaining the flight route through model inference is more accurate, and as the information samples of the area to be inspected increase, the inference result is more accurate;

[0035] 3. By the method of reversing frames at the original speed, it is possible to obtain the position in the corresponding area image in real time during the process of the drone returning to the starting point along the flight route in the second mode, so as to correspond the real-time collected distance with each crop area to obtain the farmland distance in the vertical direction from the drone to each crop area. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 is a flowchart of a method for drone farmland verification in an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0037] To enable those of ordinary skill in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings.

[0038] It should be noted that the terms "first", "second", etc. in the present invention are used to distinguish similar objects and do not necessarily have to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure.

[0039] In addition, the term "and / or" herein is merely a description of the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " herein, unless otherwise specified, generally represents an "or" relationship between the associated objects before and after.

[0040] The present invention discloses a method for verifying farmland by an unmanned aerial vehicle, in combination with Figure 1 , specifically including the following steps:

[0041] S10. Obtain a navigation route;

[0042] The navigation route is horizontal. The navigation route comes from a terminal, and the terminal can be a smart handheld terminal or a server. The unmanned aerial vehicle is communicatively connected to the terminal. The terminal is built-in with a preset map system, and a three-dimensional farmland map model is pre-stored in the map system. The model coordinates of the three-dimensional farmland map model are corresponding and associated with the real farmland position coordinates, and a selection function is provided for the verification personnel to input instructions on the terminal to select part or all of the areas in the three-dimensional farmland map model. For example, the verification personnel hold the terminal and select and determine a certain area on the three-dimensional farmland map model. The terminal generates corresponding area information to be inspected, and generates a navigation route according to the area information to be inspected, and sends the navigation route to the unmanned aerial vehicle;

[0043] Among them, generating a navigation route according to the area information to be inspected includes the following steps:

[0044] Input the area information to be inspected into a pre-trained model, and perform inference through the model to obtain a navigation route;

[0045] The model is trained in the following manner:

[0046] Perform annotation processing on each inspection area information sample in the inspection area information sample training set to mark the navigation route in each inspection area information sample. Each inspection area information sample includes inspection area information, and the navigation route is associated with all or part of the information in the inspection area information sample; and train the neural network with the marked inspection area information sample training set to obtain a model;

[0047] Among them, the inspection area information sample and the corresponding marked navigation route are obtained by the verification personnel through experiments according to the actual situation, that is, the verification personnel conduct a large number of experiments to select the navigation routes corresponding to different inspection area information that meet the conditions.

[0048] In another embodiment, generating a navigation route according to the inspection area information is implemented in the form of a preset database: that is, matching the corresponding navigation route in the preset database according to the inspection area information. Multiple different inspection area information and the navigation routes corresponding to each inspection area information are stored in the preset database in advance for subsequent calling when the same inspection area information is matched.

[0049] S12: Fly to the starting point of the navigation route and navigate from the starting point according to the navigation route;

[0050] The starting point is a three-dimensional coordinate. After the UAV obtains the navigation route, it autonomously flies from the current position to the starting point of the navigation route. The flight speed when navigating according to the navigation route can be uniform or variable, and the determination of the flight speed or the determination of its range can be determined according to the resolution of the camera installed for shooting the farmland.

[0051] S14: During the navigation according to the navigation route, collect video frame images of the farmland area to be inspected in real time;

[0052] Specifically, use a camera installed at the bottom of the UAV to shoot the farmland area. The optical axis of the camera is set vertically when the UAV is flying horizontally. The camera can use a high-definition aerial photography UAV brushless folding anti-shake camera;

[0053] S16: Stitch the collected multiple video frame images into an area image of the farmland area to be inspected;

[0054] Specifically, the stitching of the video frame images can start immediately after the video frame images are obtained, can start during the shooting of the video frame images, or can perform image stitching on the collected multiple video frame images when the UAV navigates to the end point of the navigation route.

[0055] S18: After the area image stitching is completed, analyze the area image to obtain the crop types and crop areas in the area image;

[0056] After stitching multiple video frame images into a regional image, the regional image is analyzed. Specifically, the analysis can be performed by comparing the regional image with each pre-taken crop image, so as to obtain the crop types in the regional image. The area formed by the crop image cluster is used as the crop area, so that the crop types and crop areas in the regional image can be analyzed and obtained;

[0057] S20. After the regional image stitching is completed, return from the end position to the starting point according to the navigation route in the second mode, and during the return process, use the radar set at the bottom of the UAV to collect the farmland distances from the UAV to each crop area in the vertical direction in real time;

[0058] Continuing with the above example, the UAV sails from the starting point according to the navigation route in the first mode, and the speed at any point between the end point and the starting point in the second mode is the same as the speed at any point between the end point and the starting point in the first mode;

[0059] Specifically, during the process of the UAV returning to the starting point according to the navigation route, since the speed at any point between the end point and the starting point is the same as the speed of the aforementioned UAV at any point between the end point and the starting point, and the time taken is the same. When the UAV returns through the navigation route, all video frame images are played in reverse at the original speed from the last frame to the first frame. Several points with specific pixel coordinates are taken in each video frame image and mapped onto the stitched regional image. The positions in the regional image are displayed through the mapped coordinates, so as to obtain the crop areas passed by the UAV at present, so that the distances collected in real time can be corresponding to each crop area, and thus the farmland distances from the UAV to each crop area in the vertical direction are generated;

[0060] S22. Judge the crop growth conditions of each crop area according to the farmland distances of each crop area and the initial farmland distances of each crop area.

[0061] The initial farmland distance of each crop area is the distance from each crop area to the navigation route in the vertical direction, which can be obtained by flying the UAV on the navigation route when there are no crops planted in the farmland; through the farmland distance of the crop area and the initial farmland distance of the crop area, the crop height of each crop area can be calculated, that is, the difference between the initial farmland distance of the crop area and the farmland distance of the crop area; thus, the corresponding crop growth conditions are matched in the pre-set database. The pre-set database can store the crop growth conditions corresponding to the height of each crop. The crop growth conditions can include the crop planting time, the remaining ripening time, the current growth cycle, etc.

[0062] It should be understood that the sequence numbers of the steps in the above embodiments do not indicate the order of execution, and the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0063] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for verifying farmland by an unmanned aerial vehicle, characterized in that, Including: Obtain a navigation route, and the navigation route is horizontal; Fly to the starting point of the navigation route and navigate along the navigation route from the starting point; During the navigation along the navigation route, video frame images of the farmland area to be inspected are collected in real time; Stitch multiple collected video frame images into an area image of the farmland area to be inspected; After the area image is stitched, analyze the area image to obtain the crop types and crop areas in the area image; After the area image is stitched, return from the end position to the starting point along the navigation route in the second mode, and during the return process, the farmland distances in the vertical direction from the drone to each crop area are collected in real time by a radar arranged at the bottom of the drone. Among them, the drone navigates along the navigation route from the starting point in the first mode, and the speed at any position between the end point and the starting point in the second mode is the same as the speed at any position between the starting point and the end point in the first mode; Judge the crop growth conditions of each crop area according to the farmland distances of each crop area and the initial farmland distances of each crop area; Among them, the step of collecting the farmland distances in the vertical direction from the drone to each crop area in real time by a radar arranged at the bottom of the drone during the return process includes: Play all the video frame images in reverse at the original speed from the last frame to the first frame. Take several points with specific pixel coordinates in each video frame image, map them in the stitched area image, and display the positions in the area image through the mapped coordinates, so as to obtain the crop areas passed by the drone currently, and correspond the real-time collected distances to each crop area, so as to obtain the farmland distances in the vertical direction from the drone to each crop area.

2. The method for verifying an unmanned aerial vehicle in a farmland according to claim 1, wherein The navigation route comes from a terminal. A preset map system is built in the terminal. A three-dimensional farmland map model is pre-stored in the map system. The model coordinates of the three-dimensional farmland map model are corresponding and associated with the real farmland position coordinates. The terminal provides a selection function for the verification personnel to input instructions on the terminal to select some or all areas in the three-dimensional farmland map model; the verification personnel input instructions to select and determine an area on the three-dimensional farmland map model. The terminal generates corresponding area information to be inspected, generates a navigation route according to the area information to be inspected, and sends the navigation route to the drone.

3. The method for verifying farmland by an unmanned aerial vehicle according to claim 2, wherein The step of generating a navigation route according to the area information to be inspected includes: Input the area information to be inspected into a pre-trained model, and perform inference through the model to obtain a navigation route; The model is trained in the following way: Perform annotation processing on each area information sample to be inspected in the training set of area information samples to be inspected, so as to mark the navigation route in each area information sample to be inspected. Each area information sample to be inspected includes area information to be inspected, and the navigation route is associated with all or part of the information in the area information sample to be inspected; and train a neural network through the training set of area information samples to be inspected that have undergone annotation processing to obtain the model; Among them, the information samples of the areas to be inspected and the corresponding marked navigation routes are obtained through tests by the verification personnel according to the actual situation, that is, the verification personnel conduct a large number of tests to select the navigation routes corresponding to different information of the areas to be inspected that meet the conditions.

4. The method for verifying farmland by drone according to claim 2, characterized in that Generating a navigation route according to the information of the area to be inspected includes the following steps: Match the corresponding navigation route in the preset database according to the information of the area to be inspected.

5. The method for verifying farmland by using a drone according to claim 1, wherein The analysis of the area image includes: Compare the area image with each pre-shot crop image to obtain the crop types in the area image, and use the area formed by the crop image cluster as the crop area to obtain the crop types and crop areas in the area image.

6. The method for verifying farmland by using a drone according to claim 1, wherein, The crop height of the crop area is the difference between the initial farmland distance and the farmland distance of the crop area.

7. The method for verifying farmland by an unmanned aerial vehicle according to claim 1, characterized in that The crop growth situation includes the crop planting time, the remaining ripening time, and the current growth cycle.

Images