Method for pruning plants through intelligent unmanned aerial vehicle

Abstract

The invention belongs to the technical field of unmanned aerial vehicles, and particularly relates to a method for pruning plants through an intelligent unmanned aerial vehicle, and the method comprises the following steps: starting the intelligent unmanned aerial vehicle to take off and navigate to a working area, and enabling a camera on the intelligent unmanned aerial vehicle to recognize a target plant, the intelligent unmanned aerial vehicle adjusts the position of the intelligent unmanned aerial vehicle to the height suitable for pruning according to the recognition result, the shearing device is started, the plant is guided into the range of the shearing device through the front guide-in device, and pruning is completed. Through combination of the intelligent unmanned aerial vehicle and the plant pruning device, the intelligent unmanned aerial vehicle can efficiently prune plants, the flying capability of the intelligent unmanned aerial vehicle is utilized, the effects of no terrain influence and no plant height influence can be achieved, plant pruning can be achieved, the top ends of the plants can be pruned, and the growth height of the plants can be limited; by adjusting the flying angle, the purpose of pruning plants at different angles is achieved, the labor intensity is relieved, and safety is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of unmanned aerial vehicles, and in particular relates to a method for pruning plants by an intelligent unmanned aerial vehicle. Background Art

[0002] At present, plant pruning is basically done by manpower using scissors or electric oil-driven cutting machines. When the plants are high, auxiliary tools such as long ladders are needed. There is a certain risk of working at height. The original plant pruning method basically relies on manpower. The use of mechanical tools and long ladders is inefficient, and there is a certain risk of working at height. When the plants are dense, there is a possibility of snakes and insects injuring people. Some wheeled pruning machines are restricted by the terrain. If the plants are dense, the additional damage to the plants caused by the wheeled machines is also relatively large.

[0003] Existing plant pruning methods mainly rely on manpower or wheeled machinery, which have the problems of low efficiency, high cost, lack of safety and adaptability. Summary of the invention

[0004] The purpose of the present invention is to provide a method for pruning plants using an intelligent unmanned aerial vehicle, which reduces the intensity of manual labor and increases safety, so as to solve the problems of low efficiency, high cost, insufficient safety and adaptability raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention adopts the following technical solution: a method for pruning plants using an intelligent unmanned aerial vehicle, the method comprising the following steps:

[0006] The intelligent unmanned aerial vehicle is turned on, takes off and navigates to the working area. The target plants are identified through the camera on the intelligent unmanned aerial vehicle. The intelligent unmanned aerial vehicle adjusts its position to a suitable height for pruning based on the identification results. The shearing device is then started and the plants are guided into the range of the shearing device through the front introduction device to complete the pruning.

[0007] Preferably, the intelligent unmanned aerial vehicle includes a shearing device installed at the bottom thereof, the shearing device is driven by a high-speed motor to drive the shearing component to move, and the maximum size diameter of the shearing component matches the small opening width of the front introduction device.

[0008] Preferably, the front introduction device is installed on the landing gear below the intelligent unmanned aerial vehicle;

[0009] The front introduction device includes two tubes in an eight-shaped shape, with the large end of the eight-shaped shape facing the forward direction of the intelligent unmanned aerial vehicle, and the small end is connected to the shearing device. A camera is installed on one of the tubes.

[0010] Preferably, the intelligent unmanned aerial vehicle also includes an airflow evacuation device, which is installed between the rotor and the shearing device of the intelligent unmanned aerial vehicle and is in the shape of a closed trumpet, with the small mouth of the trumpet facing upward close to the rotor and the large mouth facing downward, wrapping the landing gear and the shearing device.

[0011] Preferably, before performing the pruning task, the intelligent unmanned aerial vehicle collects image data of the target plant through a camera and pre-processes the image data, including adjusting the image size, regularization and data enhancement, to meet subsequent plant identification needs.

[0012] Preferably, the intelligent unmanned aerial vehicle also performs preprocessing of text data, including cleaning text data, word segmentation, stemming and removal of stop words, and converting the text data into a processable format to support the formulation of plant pruning strategies.

[0013] Preferably, the intelligent unmanned aerial vehicle extracts image features using a pre-trained visual model through a multimodal architecture, extracts text features using a ChatGLM4 model, and fuses feature vectors of the image and text through an attention mechanism.

[0014] Preferably, the intelligent unmanned aerial vehicle classifies the fused features through a fully connected layer or a multi-layer perceptron, outputs plant categories, and thus determines the objects to be pruned.

[0015] Preferably, the shearing component is configured as a shearing blade.

[0016] Preferably, the shearing member is configured as a metal rope.

[0017] Technical effects and advantages of the present invention: Compared with the prior art, the method for pruning plants by an intelligent unmanned aerial vehicle proposed by the present invention has the following advantages:

[0018] The present invention combines an intelligent unmanned aerial vehicle with a plant pruning device, so that the intelligent unmanned aerial vehicle can efficiently prune plants. By utilizing the flight capability of the intelligent unmanned aerial vehicle, plant pruning can be achieved without being affected by the terrain or the height of the plants. The tops of plants can be trimmed and the growth height of plants can be limited. The purpose of pruning plants at different angles can be achieved by adjusting the flight angle, thereby reducing the intensity of manual labor and increasing safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A flow chart of a method for pruning plants by an intelligent unmanned aerial vehicle of the present invention;

[0020] Figure 2 It is a structural schematic diagram of the intelligent unmanned aerial vehicle of the present invention;

[0021] Figure 3It is a schematic diagram of the visual model of the present invention.

[0022] In the figure: 1. guide plate; 2. shearing component; 3. camera; 4. landing gear; 5. tube body; 6. intelligent unmanned aerial vehicle. DETAILED DESCRIPTION

[0023] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. The specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0024] The present invention provides a method for pruning plants using an intelligent unmanned aerial vehicle, which allows the intelligent unmanned aerial vehicle to have more uses. By combining the intelligent unmanned aerial vehicle with a plant pruning device, the intelligent unmanned aerial vehicle can efficiently prune plants, avoiding the shortcomings of manpower and wheeled machinery. By utilizing the flight capability of the intelligent unmanned aerial vehicle, the top of the plant and the side of some angles can be pruned without being affected by the terrain or the height of the plant.

[0025] For example: when corn is artificially produced, the pollination of the male and female flowers of the seeds is the key to corn artificial production. Corn is hermaphrodite, and through artificial screening, the male and female flowers of different varieties of corn are distinguished and retained, and the seed production requirement of specific varieties of corn is reached. At this time, it is necessary to remove the top male flowers of some corns, and the industry is called emasculation operation, and the original emasculation operation mode is completed by hand or wheeled machinery, and there are limitations, such as using the method of the present invention, the work tasks can be completed quickly, efficiently and economically, and the target of corn emasculation is reached. In the same way, the present invention can also quickly prune other plants that need to be pruned.

[0026] For the convenience of description, this embodiment takes pruning corn as an example for explanation:

[0027] like Figure 1-3 As shown, a method for pruning plants by an intelligent unmanned aerial vehicle proposed in the present invention comprises the following steps:

[0028] The intelligent unmanned aerial vehicle 6 is turned on and takes off and navigates to the working area. The target plant is identified by the camera on the intelligent unmanned aerial vehicle. The intelligent unmanned aerial vehicle adjusts its position to a suitable height for pruning according to the identification result. Then the shearing device is started and the plant is guided into the shearing device range by the front introduction device to complete the pruning. The details are as follows:

[0029] The above-mentioned intelligent unmanned aerial vehicle includes a shearing device installed at the bottom thereof, and the shearing device is driven by a high-speed motor to drive the shearing component 2 to move, and the maximum size diameter of the shearing component matches the small opening width of the front introduction device.

[0030] In another embodiment, the shearing component 2 may be a shearing blade or a metal rope.

[0031] Solution 1: Shearing component 2 is a shearing blade:

[0032] The shearing device is driven by a high-speed motor to drive the shear blade to move. The maximum diameter of the shear blade matches the width of the small opening of the front introduction device to ensure that the plants will not be missed. The shear blade is installed at the bottom of the intelligent unmanned aerial vehicle, the rotating blade is parallel to the ground, and the shearing direction faces the small opening of the eight-shaped introduction tube.

[0033] Specific implementation process: Take-off and navigation: The intelligent unmanned aerial vehicle takes off from the ground and navigates to the work area through the GPS Beidou satellite navigation system. Target identification: After arriving at the work area, the camera on the intelligent unmanned aerial vehicle identifies the target plant. The camera uses an AI algorithm to identify the type and location of the plant to ensure accurate pruning. Adjust the height: The intelligent unmanned aerial vehicle adjusts its position to a suitable height for pruning based on the recognition results. Shearing device starts: The front introduction device guides the plant into the range of the shearing device, and the high-speed motor starts to drive the shear blade to rotate. Pruning plants: The plant is guided by the figure eight introduction device and contacts the rotating shear blade to complete the pruning.

[0034] The high-speed rotation of the shear blade ensures efficient pruning and can complete the pruning of a large number of plants in a short time. The recognition ability of the AI ​​camera enables the intelligent unmanned aerial vehicle to accurately identify the target plant and avoid mis-pruning. The design of the front introduction device ensures that the plant will not deviate from the range of the shear device under the influence of airflow, reducing plant damage.

[0035] Solution 2: The shearing component 2 is a metal rope or a metal blade:

[0036] The maximum diameter of the metal rope matches the small opening width of the front introduction device to ensure that the plants will not be missed. The metal rope is installed at the bottom of the intelligent unmanned aerial vehicle, and the installation direction can be adjusted as needed, either upward or downward.

[0037] Specific implementation process: Take-off and navigation: The intelligent unmanned aerial vehicle takes off from the ground and navigates to the work area through the GPS Beidou satellite navigation system. Target identification: After arriving at the work area, the camera on the intelligent unmanned aerial vehicle identifies the target plant. The camera uses an AI algorithm to identify the type and location of the plant to ensure accurate pruning. Adjust the height: The intelligent unmanned aerial vehicle adjusts its position to a suitable height for pruning based on the recognition results. Shearing device starts: The front introduction device guides the plant into the range of the shearing device, and the high-speed motor starts, driving the metal rope to move at high speed. Pruning plants: The plants are guided by the figure eight introduction device and come into contact with the high-speed moving metal rope or blade to complete the pruning.

[0038] The flexibility of the metal rope makes it suitable for different types of plants, especially those with thin branches that are easy to break. The high-speed movement of the metal rope ensures the precision of pruning and can work effectively in complex environments. The metal blade is suitable for plants with thicker branches and more lignification. The design of the front introduction device ensures that the plants will not deviate from the range of the shearing device under the influence of airflow, reducing plant damage. The camera at the front of the front introduction device helps guide the flight direction of the unmanned aerial vehicle and put the plants to be cut into the front introduction device.

[0039] Through the above two solutions, the shearing device of the intelligent unmanned aerial vehicle can complete the plant pruning task efficiently and accurately, significantly improving the safety and efficiency of pruning operations while reducing labor costs and plant damage.

[0040] Specifically, the front introduction device is installed on the landing gear 4 below the intelligent unmanned aerial vehicle; the front introduction device includes two eight-shaped tubes 5, and the large ends of the eight-shaped tubes face the forward direction of the intelligent unmanned aerial vehicle, and the small ends are connected to the shearing device, and a camera 3 is installed on one of the tubes.

[0041] Specifically, before performing the pruning task, the intelligent unmanned aerial vehicle collects image data of the target plants through a camera and pre-processes the image data, including adjusting the image size, regularization, and data enhancement to meet subsequent plant identification needs.

[0042] The intelligent UAV also performs pre-processing of text data, including cleaning text data, word segmentation, stemming, and removing stop words, converting the text data into a processable format to support the formulation of plant pruning strategies.

[0043] The intelligent UAV uses a pre-trained visual model through a multimodal architecture to extract image features, uses the ChatGLM4 model to extract text features, and fuses the feature vectors of the image and text through an attention mechanism. The intelligent UAV classifies the fused features through a fully connected layer or a multi-layer perceptron, outputs the plant category, and thus determines the object to be pruned.

[0044] The front introduction device is made of lightweight rigid materials, such as carbon fiber tubes, and consists of two tubes in an eight-shaped shape. The small end of the eight-shaped tube is connected to the bottom landing gear of the intelligent unmanned aerial vehicle, and the other end of the large end is suspended horizontally. One of the tubes is connected to a camera for AI to identify plant species and distances, so as to control the intelligent unmanned aerial vehicle to accurately prune plants. When the intelligent unmanned aerial vehicle is flying, the pruned plants move from the middle of the two tubes to the shearing device. The two tubes prevent the plants from leaving the range of the shearing device under the influence of the airflow of the intelligent unmanned aerial vehicle.

[0045] The AI ​​software algorithm collects a large number of pictures and videos of target plants from different angles before executing the repair task, uses the open source large language model (LLM) of the base, inputs pictures and videos to fine-tune the model, and finally uses deep learning to enable the intelligent unmanned aerial vehicle to have the ability to identify target plants from multiple angles. The steps to implement the algorithm are:

[0046] (1) Construct a network model; (2) Obtain the training set and data set required by the model; (3) Use the convolutional layer of a deep convolutional neural network to detect targets; (4) Extract some features; (5) Establish connections between multiple layers of feature maps; (6) Use the anchor and regional production network structure to generate candidate boxes; (7) Pre-train the YOLO network; (8) Formal training of the model network; (9) Take a picture of corn to be detasseled and input it into the model to identify the male flowers and obtain the correct detasseling category.

[0047] The specific method is:

[0048] (1) Data collection:

[0049] Image data: Contains pictures of various plant types, saved in different categories.

[0050] Text data: text descriptions corresponding to the image, such as plant names and corresponding drone pruning methods.

[0051] (2) Data preprocessing:

[0052] Image preprocessing:

[0053] Resize images: For uniform processing, you can use a library such as Pillow or OpenCV to resize images to a fixed size.

[0054] Regularization: Scaling pixel values ​​to the range [0,1] or standardizing them to have a mean of 0 and a variance of 1.

[0055] Data enhancement: such as horizontal flipping, rotation, cropping, etc., to increase the diversity of training samples.

[0056] (3) Text preprocessing:

[0057] Clean text data, remove punctuation, etc.

[0058] Tokenization, stemming, and stop word removal.

[0059] Convert the text into a format that the model can process and input it into a word embedding model (such as BERT, GPT) or directly into the ChatGLM4 model.

[0060] (4) Label encoding:

[0061] Convert plant class labels into numerical values ​​(e.g., using Label Encoder or One-Hot Encoding).

[0062] Multimodal large language model (based on chatg lm4) architecture design

[0063] (5) Visual branch:

[0064] Use a pre-trained vision model (such as CLIP or Visual Transformer) to extract image features.

[0065] (6) Text branch:

[0066] The ChatGLM4 model is used to extract text features from text descriptions.

[0067] (7) Feature Fusion:

[0068] The feature vectors of image and text are fused using the attention mechanism.

[0069] (8) Classification layer:

[0070] Finally, the fused features are classified through a fully connected layer or a multi-layer perceptron (MLP) to output the plant category.

[0071] (9) Model training:

[0072] Initialization parameters: If it is transfer learning, the parameters of the pre-trained model can be used as initialization parameters.

[0073] Define loss function: The cross entropy loss function is usually used for classification tasks.

[0074] Select optimizer: such as Adam, SGD, etc.

[0075] Start training: Iteratively train the model on the training set, evaluate the performance on the validation set, and adjust the hyperparameters.

[0076] Multimodal alignment: During training, one can try to align the representations of images and text, for example using contrast loss to maximize the similarity of similar image and text features.

[0077] (10) Model evaluation:

[0078] Use the validation set for fine-tuning (e.g., adjusting hyperparameters to prevent overfitting).

[0079] Evaluate the performance of the final model on the test set to ensure generalization ability.

[0080] (11) Deployment Model:

[0081] Model export: Export the trained model to a deployable format.

[0082] Integrate into applications: Integrate the model into mobile applications, web services, or IoT devices.

[0083] Furthermore, the intelligent unmanned aerial vehicle also includes an airflow evacuation device, which includes a guide plate 1 installed between the rotor of the intelligent unmanned aerial vehicle and the shearing device, and has a closed trumpet shape. The small mouth of the trumpet is upward and close to the rotor, and the large mouth is downward, wrapping the landing gear and the shearing device.

[0084] The airflow evacuation device is installed between the intelligent unmanned aerial vehicle rotor and the shearing device to isolate the airflow caused by the rotation of the intelligent unmanned aerial vehicle rotor. If the airflow is not isolated, the airflow will blow away the leaves of the plant, and the change in the position of the plant leaves may make it impossible for the shearing device to cut the plant leaves smoothly. The airflow evacuation device can be composed of a lightweight plate, or a frame and a cover cloth similar to an umbrella.

[0085] The specific method is as follows:

[0086] The intelligent UAV takes off and navigates to the work area through visual or GPS Beidou satellite navigation. After reaching the pruning target, the intelligent UAV camera with edge computing capabilities identifies the target plant;

[0087] According to the pre-set pruning plan, the intelligent unmanned aerial vehicle is controlled to descend to a suitable pruning height, the shearing device starts working, the high-speed motor drives the blade to rotate, and the front guide device points to the plant to be sheared;

[0088] Under the guidance of the camera, the intelligent unmanned aerial vehicle flies towards the plants to be cut. The cut plants come into contact with the rotating blades of the shearing device through the introduction device, thereby achieving the purpose of pruning.

[0089] Since the airflow of the intelligent unmanned aerial vehicle will affect the location of plant cultivation, there is an airflow diversion and dispersion device under the rotor of the intelligent unmanned aerial vehicle to guide the aircraft's airflow in other directions to accurately trim the target plants.

[0090] The shearing device is installed under the intelligent unmanned aerial vehicle and forms a firm physical connection with the bottom of the fuselage through locking screws and other parts. The shearing device is driven by a high-speed electric motor to drive the shearing blade or metal rope. The blade or metal rope can be installed upward or downward. The power supply can be provided by the intelligent unmanned aerial vehicle or it can be self-contained.

[0091] The front introduction device is installed near the landing gear under the intelligent unmanned aerial vehicle, in the forward direction of the nose of the intelligent unmanned aerial vehicle, and forms a firm physical connection with the fuselage landing gear through locking screws and other parts.

[0092] The airflow evacuation device is installed under the rotor of the intelligent unmanned aerial vehicle, surrounding the fuselage in a closed trumpet shape, composed of lightweight plates, and forms a firm physical connection with the fuselage and landing gear through locking screws and other parts. The small end of the trumpet is upward and close to the rotor, and the large end of the trumpet is downward, wrapping the landing gear and shearing device. In this way, the rotor airflow of the intelligent unmanned aerial vehicle will be guided to the surroundings along the slope of the trumpet, and will not form turbulence to blow away the plants and affect the work of the shearing device.

[0093] The present invention combines an intelligent unmanned aerial vehicle with a plant pruning device, so that the intelligent unmanned aerial vehicle can efficiently prune plants. By utilizing the flight capability of the intelligent unmanned aerial vehicle, plant pruning can be achieved without being affected by the terrain or the height of the plants. The tops of plants can be trimmed and the growth height of plants can be limited. The purpose of pruning plants at different angles can be achieved by adjusting the flight angle, thereby reducing the intensity of manual labor and increasing safety.

[0094] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for pruning plants using an intelligent unmanned aerial vehicle, characterized in that: The method comprises the following steps: The intelligent unmanned aerial vehicle is turned on, takes off and navigates to the working area. The target plants are identified through the camera on the intelligent unmanned aerial vehicle. The intelligent unmanned aerial vehicle adjusts its position to a suitable height for pruning based on the identification results. The shearing device is then started and the plants are guided into the range of the shearing device through the front introduction device to complete the pruning.

2. The method for pruning plants using an intelligent unmanned aerial vehicle according to claim 1, characterized in that: The intelligent unmanned aerial vehicle comprises a shearing device installed at the bottom thereof. The shearing device is driven by a high-speed motor to drive the shearing component to move. The maximum diameter of the shearing component matches the small opening width of the front introduction device.

3. The method for pruning plants using an intelligent unmanned aerial vehicle according to claim 2, characterized in that: The front introduction device is installed on the landing gear below the intelligent unmanned aerial vehicle; The front introduction device includes two tubes in an eight-shaped shape, with the large end of the eight-shaped shape facing the forward direction of the intelligent unmanned aerial vehicle, and the small end is connected to the shearing device. A camera is installed on one of the tubes.

4. The method for pruning plants using an intelligent unmanned aerial vehicle according to claim 3, characterized in that: The intelligent unmanned aerial vehicle also includes an airflow evacuation device, which is installed between the rotor and the shearing device of the intelligent unmanned aerial vehicle and is in the shape of a closed trumpet, with the small mouth of the trumpet facing upward and close to the rotor, and the large mouth facing downward, wrapping the landing gear and the shearing device.

5. The method for pruning plants using an intelligent unmanned aerial vehicle according to claim 4, characterized in that: Before performing the pruning task, the intelligent unmanned aerial vehicle collects image data of the target plant through a camera and pre-processes the image data, including adjusting the image size, regularization and data enhancement, to meet subsequent plant recognition needs.

6. The method for pruning plants using an intelligent unmanned aerial vehicle according to claim 5, characterized in that: The intelligent unmanned aerial vehicle also performs preprocessing of text data, including cleaning text data, word segmentation, stemming and removing stop words, and converting the text data into a processable format to support the formulation of plant pruning strategies.

7. The method for pruning plants using an intelligent unmanned aerial vehicle according to claim 6, characterized in that: The intelligent unmanned aerial vehicle uses a pre-trained visual model through a multimodal architecture to extract image features, uses the ChatGLM4 model to extract text features, and fuses the feature vectors of the image and text through an attention mechanism.

8. The method for pruning plants using an intelligent unmanned aerial vehicle according to claim 7, characterized in that: The intelligent unmanned aerial vehicle classifies the fused features through a fully connected layer or a multi-layer perceptron, outputs plant categories, and thus determines the objects to be pruned.

9. The method for pruning plants using an intelligent unmanned aerial vehicle according to claim 2, characterized in that: The shearing component is configured as a shearing blade.

10. The method for pruning plants using an intelligent unmanned aerial vehicle according to claim 2, characterized in that: The shearing member is configured as a metal rope or a metal blade.

Images