Farmland checking method and system based on unmanned aerial vehicle, electronic equipment and storage medium

Abstract

The invention belongs to the technical field of unmanned aerial vehicle planning and image processing, and particularly relates to a farmland checking method and system based on an unmanned aerial vehicle, electronic equipment and a storage medium. The method comprises the following steps: designing a flight path covering the whole field based on farmland area and topographic characteristic data; based on the flight path, an unmanned aerial vehicle is adopted for shooting, and farmland images are obtained; preprocessing the farmland image to obtain processed data; inputting the processed data into a constructed recognition model, and performing crop category recognition to obtain a recognition result; and based on the identification result, generating a verification report, and completing verification. According to the invention, the flight path covering the whole field is designed based on the farmland area and topographic characteristic data, and it is ensured that the unmanned aerial vehicle can efficiently and comprehensively acquire farmland images. According to the method, the accuracy of crop category identification is improved, a detailed check report can be quickly generated, and a scientific decision basis is provided for an agricultural management department.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) planning and image processing, and in particular to a farmland verification method, system, electronic device and storage medium based on an unmanned aerial vehicle (UAV). Background Art

[0002] When using drones for farmland inspection, how to ensure that the acquired farmland image data has sufficient clarity and resolution is a key technical issue. Due to the limitations of the flight altitude and speed of drones, as well as the complexity of the farmland environment, how to obtain high-quality image data in a limited time faces many challenges. First, it is necessary to reasonably plan the flight route and altitude of the drone according to the size of the farmland and the terrain characteristics to ensure the integrity and consistency of the image data. Secondly, it is necessary to select appropriate sensors and lenses to meet the needs of different farmland inspection indicators, such as multispectral sensors for crop chlorophyll content analysis, infrared sensors for soil heat distribution analysis, etc. Furthermore, it is also necessary to consider the influence of external environmental factors, such as light intensity, cloud and fog occlusion, etc., which will directly affect the quality of image data. Finally, in the data processing and analysis stage, appropriate image processing algorithms and analysis models need to be adopted, such as image stitching, geometric correction, radiation correction and other preprocessing operations, as well as machine learning-based crop identification and growth status assessment models to improve the accuracy and efficiency of farmland inspection. Summary of the invention

[0003] In order to solve the above technical problems, the present invention provides a farmland inspection method based on a drone, the method comprising:

[0004] Based on the farmland area and terrain feature data, a flight path covering the entire field is designed; the steps include: based on the farmland area and terrain feature data, obtaining key terrain feature parameters; optimizing the key terrain feature parameters, and using a path planning algorithm to design a flight path covering the entire field;

[0005] Based on the flight path, a drone is used to shoot and obtain images of the farmland; when the real-time monitored light intensity and cloud and fog obstruction meet the preset conditions, the drone is controlled to fly along the flight path to obtain images of the farmland; otherwise, the flight altitude, speed and heading of the drone are adjusted until the preset conditions are met;

[0006] Preprocessing the farmland images to obtain processed data; stitching the original images using an image stitching algorithm to generate a preliminary stitching image; using an image fusion algorithm to perform smooth transition processing on the stitching area to eliminate stitching traces and obtain a preliminary fused image;

[0007] The processed data is input into a constructed recognition model to perform crop category recognition to obtain a recognition result; the recognition model is constructed using a CNN network;

[0008] Based on the identification results, a verification report is generated to complete the verification.

[0009] Preferably, the key terrain parameters include: the highest point, the lowest point, and the average slope; based on the key terrain parameters, the flight parameters are optimized using a genetic algorithm, and the steps include:

[0010] Parameter encoding: encoding the flight parameters to form an individual gene sequence;

[0011] Initialize the population: randomly generate a number of individuals to form an initial population;

[0012] Fitness evaluation: Based on the flight parameters of each individual, simulate the flight of the drone and obtain the farmland image data, and then calculate the fitness of each individual based on the clarity, resolution, and coverage completeness of the image data;

[0013] Selection operation: Roulette selection method is used to calculate the probability of each individual being selected according to the fitness score;

[0014] Crossover operation: Randomly select two parent individuals, perform single-point crossover at the middle position of the gene sequence, and generate two offspring individuals;

[0015] Mutation operation: Gene mutation is performed on the newly generated offspring individuals with a preset probability;

[0016] Iterative optimization: Repeat fitness evaluation, selection, crossover, and mutation, and iterate continuously until the termination condition is met.

[0017] Preferably, the A-star algorithm is used for path planning, and its expression is as follows:

[0018]

[0019] in, Represents the actual cost from the starting point to the current node n; represents the heuristic estimated cost from the current node n to the target node; Represents the total evaluation cost, which is used to determine the priority of the node.

[0020] Preferably, the acquired farmland images are stitched together using an image stitching algorithm to generate a preliminary stitched image:

[0021]

[0022] Where H represents the homography transformation matrix; xi and x j Respectively represent the matching feature points in the two images; Represents the square of the Euclidean distance.

[0023] Preferably, the recognition model constructed includes:

[0024] Convolutional layer 1: The number of filters is 32, the filter size is: 3×3; the activation function uses ReLU; the step size is 1; the input layer, convolutional layer 1, pooling layer 1, convolutional layer 2, pooling layer 2, fully connected layer and output layer, its structure is as follows:

[0025] Pooling layer 1: pooling window size is 2×2; stride is 2;

[0026] Convolutional layer 2: The number of filters is 64; the filter size is 3×3; the activation function is ReLU; the step size is 1;

[0027] Pooling layer 2: pooling window size is 2×2; step size is 2;

[0028] Fully connected layer: the number of neurons is 128; the activation function is ReLU;

[0029] Output layer: Number of neurons: N (in this embodiment, N is the number of crop categories); activation function is Softmax.

[0030] The present invention also provides a farmland inspection system based on a drone, the system is used to implement the above method, and includes: a design module, a shooting module, a processing module, an identification module and an inspection module;

[0031] The design module is used to design a flight path covering the entire field based on the farmland area and terrain characteristics data; the steps include: obtaining key terrain characteristic parameters based on the farmland area and terrain characteristics data; optimizing the key terrain characteristic parameters, and using a path planning algorithm to design a flight path covering the entire field;

[0032] The shooting module is used to shoot with a drone based on the flight path to obtain images of the farmland; when the real-time monitored light intensity and cloud and fog occlusion meet the preset conditions, the drone is controlled to fly along the flight path to obtain images of the farmland; otherwise, the flight altitude, speed and heading of the drone are adjusted until the preset conditions are met;

[0033] The processing module is used to pre-process the farmland image to obtain processed data; stitch the original image through an image stitching algorithm to generate a preliminary stitching image; and use an image fusion algorithm to perform smooth transition processing on the stitching area to eliminate stitching traces and obtain a preliminary fused image;

[0034] The recognition module is used to input the processed data into a constructed recognition model to perform crop category recognition and obtain a recognition result; the recognition model is constructed using a CNN network;

[0035] The verification module is used to generate a verification report based on the identification result to complete the verification.

[0036] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the above method is implemented when the processor executes the program.

[0037] A computer-readable storage medium stores a computer program, and when the computer program is executed, the above method is implemented.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] The present invention ensures that the drone can efficiently and comprehensively acquire images of farmland by designing a flight path that covers the entire field based on the farmland area and terrain characteristics data. The present invention not only improves the accuracy of crop category recognition, but also can quickly generate detailed verification reports, providing a basis for scientific decision-making for agricultural management departments. This not only helps to improve the level of refined management of agricultural production, but also optimizes resource allocation, improves yield and quality, while reducing environmental impact and promoting the sustainable development of agriculture. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0041] Figure 1 A schematic diagram of a method flow of an embodiment of the present invention;

[0042] Figure 2 Schematic diagram of the structure of an electronic device according to an embodiment of the present invention.

[0043] 1010, processor; 1020, memory; 1030, input / output interface; 1040, communication interface; 1050, bus. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present disclosure should be understood by people with ordinary skills in the field to which the present disclosure belongs. The "first", "second" and similar words used in the embodiments of the present disclosure do not indicate any order, quantity or importance, but are only used to distinguish different components. "Including" or "comprising" and similar words mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0046] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0047] Embodiment 1

[0048] As can be seen from the background technology, due to the limitations of the flight altitude and speed of drones, as well as the complexity of the farmland environment, there are many challenges in obtaining high-quality image data within a limited time.

[0049] The embodiment of the present invention provides a farmland inspection method based on a drone, such as Figure 1 As shown, the steps include:

[0050] S1. Based on the farmland area and terrain characteristics data, design a flight path covering the entire field.

[0051] Obtain farmland area data and terrain feature data, and input the data into the UAV flight control system. According to the farmland area data and terrain feature data, analyze the terrain and extract the key terrain feature parameters that affect the flight of the UAV, such as the highest point, the lowest point, the average slope, etc. Based on the key terrain feature parameters, optimize the flight altitude, speed, route interval and other parameters through machine learning algorithms to obtain the optimal flight parameter combination suitable for the terrain of the area. According to the optimized flight parameters, the path planning algorithm is used to automatically generate a flight path covering the area. The flight path is required to completely cover all locations in the area and the flight distance is the shortest. The flight paths of each area are spliced ​​to obtain a complete flight path covering the entire farmland. The generated flight path is sent to the UAV flight control system, and the flight mission is performed through the autonomous flight mode to complete the coverage of the entire farmland.

[0052] Specifically, this embodiment uses a genetic algorithm to optimize flight parameters, and the steps include:

[0053] Parameter coding: Encode parameters such as flight altitude, speed, and route interval to form an individual's genetic sequence.

[0054] Initialize the population: Randomly generate a certain number (such as 100) of individuals to form an initial population.

[0055] Fitness evaluation: Based on the flight parameters of each individual, simulate UAV flight and obtain farmland image data. Then, calculate the fitness of each individual based on indicators such as clarity, resolution, and coverage completeness of the image data.

[0056] Selection operation: Roulette selection method is used to calculate the probability of each individual being selected based on the fitness score. Individuals with higher probabilities are more likely to be selected. For example, individuals with higher fitness scores have a greater probability of being selected, thus having more opportunities to reproduce.

[0057] Crossover operation: Randomly select two parent individuals, perform a single-point crossover at the middle position of the gene sequence, and generate two offspring individuals.

[0058] Mutation operation: Perform gene mutation on the newly generated offspring individuals with a certain probability (such as 0.01).

[0059] Iterative optimization: Repeat the fitness evaluation, selection, crossover, mutation and other steps, and iterate continuously until the termination condition is met. For example, after 100 iterations, the flight parameters represented by the individual with the highest fitness score are the optimized flight parameters.

[0060] The A-star algorithm is used for path planning, and its expression is as follows:

[0061]

[0062] in, Represents the actual cost from the starting point to the current node n; represents the heuristic estimated cost from the current node n to the target node; Represents the total evaluation cost, which is used to determine the priority of the node.

[0063] S2. Based on the flight path, use drones to shoot and obtain farmland images.

[0064] If the real-time monitored light intensity and cloud obstruction meet the preset conditions, the UAV is controlled to fly along the flight path to obtain farmland images. Otherwise, the flight altitude, speed and heading of the UAV are adjusted until the preset conditions are met to obtain high-quality farmland images.

[0065] The sensor obtains the light intensity and cloud cover data of the drone's flight area in real time, and compares the obtained data with the preset conditions to determine whether the preset conditions are met. If the preset conditions are met, the drone is controlled to fly according to the predetermined flight path, and the high-definition camera on the drone obtains farmland images, and the obtained image data is transmitted to the ground control center. If the preset conditions are not met, the fuzzy control algorithm is used according to the obtained light intensity and cloud cover data to dynamically adjust the drone's flight altitude, speed and heading until the preset conditions are met.

[0066] S3. Preprocess the farmland images to obtain processed data.

[0067] Obtain the original image data of the field collected by the drone, and use the image stitching algorithm to stitch the original images to generate a preliminary stitched image. Specifically, extract the feature points in the image to be stitched and pair them. Use the matched feature point pairs to calculate the homography transformation matrix H. The homography transformation matrix is ​​a 3x3 matrix that describes the geometric transformation relationship between the two images. The expression is as follows:

[0068]

[0069] Among them, x i and x j Respectively represent the matching feature points in the two images; Represents the square of the Euclidean distance.

[0070] For the preliminary stitching images, the image fusion algorithm is used to perform smooth transition processing on the stitching area, eliminate the stitching traces, and obtain the preliminary fused image. According to the preset image quality evaluation index, the quality of the preliminary fused image is evaluated. If the quality requirements are met, the image is output as a complete image covering the entire field; if the quality requirements are not met, the stitching process is returned and re-executed.

[0071] The above image fusion algorithm adopts a multi-resolution pyramid image fusion algorithm, and the steps include:

[0072] Construct a Gaussian pyramid: Construct a Gaussian pyramid for each of the two input images, and the image resolution of each layer gradually decreases.

[0073] Construct a Laplacian pyramid: Subtract the downsampled and upsampled image from each layer of the Gaussian pyramid to obtain a Laplacian pyramid.

[0074] Fusing Laplacian Pyramids: The Laplacian pyramids of two images are fused at each level, usually using a simple weighted average or a more complex fusion strategy.

[0075] Reconstruct the image: Start from the bottom layer of the fused Laplacian pyramid, gradually upsample and accumulate, and finally get the fused image.

[0076] S4. Input the processed data into the constructed recognition model to perform crop category recognition and obtain the recognition result.

[0077] This implementation uses the CNN network results to build a recognition model, including: input layer, convolution layer 1, pooling layer 1, convolution layer 2, pooling layer 2, fully connected layer and output layer. Its specific structure is as follows:

[0078] Convolutional layer 1: The number of filters is 32, the filter size is 3×3; the activation function uses ReLU; the step size is 1;

[0079] Pooling layer 1: pooling window size is 2×2; stride is 2;

[0080] Convolutional layer 2: The number of filters is 64; the filter size is 3×3; the activation function is ReLU; the step size is 1;

[0081] Pooling layer 2: pooling window size is 2×2; step size is 2;

[0082] Fully connected layer: the number of neurons is 128; the activation function is ReLU;

[0083] Output layer: Number of neurons: N (in this embodiment, N is the number of crop categories); activation function is Softmax.

[0084] The processed data is input into the above model to perform crop identification and obtain the identification results.

[0085] S5. Based on the identification results, generate a verification report and complete the verification.

[0086] Based on these analysis results, agricultural experts can generate detailed farmland verification reports. The report not only contains a quantitative description of crop distribution and growth conditions, but also provides targeted management suggestions. For example, for areas with low DVI values, farmers are advised to carry out pest control or supplementary fertilization in a timely manner. For areas with good growth, inputs can be appropriately reduced to improve resource utilization efficiency. This farmland verification report provides a basis for scientific decision-making for agricultural management departments. Management departments can adjust the planting structure based on the report, such as increasing the planting area of ​​high-yield crops, or optimize resource allocation, such as adjusting the irrigation system to improve water resource utilization efficiency. In this way, refined management of agricultural production can be achieved, yield and quality can be improved, while reducing environmental impact and promoting sustainable agricultural development.

[0087] The technical solution of the present invention ensures that the drone can efficiently and comprehensively obtain farmland images by designing a flight path covering the entire field based on the farmland area and terrain characteristics data. It not only improves the accuracy of crop category recognition, but also can quickly generate detailed verification reports, providing a basis for scientific decision-making for agricultural management departments. This not only helps to improve the level of refined management of agricultural production, but also optimizes resource allocation, improves yield and quality, while reducing environmental impact and promoting the sustainable development of agriculture.

[0088] It should be noted that the method of the embodiment of the present disclosure can be performed by a single device, such as a computer or a server. The method of the present embodiment can also be applied in a distributed scenario and completed by multiple devices cooperating with each other. In the case of such a distributed scenario, one of the multiple devices can only perform one or more steps in the method of the embodiment of the present disclosure, and the multiple devices will interact with each other to complete the described method.

[0089] It should be noted that some embodiments of the present disclosure are described above. Other embodiments are within the scope of the attached claims. In some cases, it should be understood that the size of the serial number of each step in the above embodiment does not mean 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 on the implementation process of the embodiment of the present invention. The actions or steps recorded in the claims can be performed in an order different from that in the above embodiment and still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0090] Embodiment 2

[0091] Based on the same inventive concept, corresponding to any of the above-mentioned embodiment methods, the present disclosure also provides a drone-based farmland inspection system, including: a design module, a shooting module, a processing module, an identification module and a inspection module.

[0092] The following will explain in detail how the present invention solves technical problems in real life in conjunction with this embodiment.

[0093] First, the design module designs a flight path covering the entire field based on the farmland area and terrain characteristics data.

[0094] Obtain farmland area data and terrain feature data, and input the data into the UAV flight control system. According to the farmland area data and terrain feature data, analyze the terrain and extract the key terrain feature parameters that affect the flight of the UAV, such as the highest point, the lowest point, the average slope, etc. Based on the key terrain feature parameters, optimize the flight altitude, speed, route interval and other parameters through machine learning algorithms to obtain the optimal flight parameter combination suitable for the terrain of the area. According to the optimized flight parameters, the path planning algorithm is used to automatically generate a flight path covering the area. The flight path is required to completely cover all locations in the area and the flight distance is the shortest. The flight paths of each area are spliced ​​to obtain a complete flight path covering the entire farmland. The generated flight path is sent to the UAV flight control system, and the flight mission is performed through the autonomous flight mode to complete the coverage of the entire farmland.

[0095] Specifically, this embodiment uses a genetic algorithm to optimize flight parameters, and the steps include:

[0096] Parameter coding: Encode parameters such as flight altitude, speed, and route interval to form an individual's genetic sequence.

[0097] Initialize the population: Randomly generate a certain number (such as 100) of individuals to form an initial population.

[0098] Fitness evaluation: Based on the flight parameters of each individual, simulate UAV flight and obtain farmland image data. Then, calculate the fitness of each individual based on indicators such as clarity, resolution, and coverage completeness of the image data.

[0099] Selection operation: Roulette selection method is used to calculate the probability of each individual being selected based on the fitness score. Individuals with higher probabilities are more likely to be selected. For example, individuals with higher fitness scores have a greater probability of being selected, thus having more opportunities to reproduce.

[0100] Crossover operation: Randomly select two parent individuals, perform a single-point crossover at the middle position of the gene sequence, and generate two offspring individuals.

[0101] Mutation operation: Perform gene mutation on the newly generated offspring individuals with a certain probability (such as 0.01).

[0102] Iterative optimization: Repeat the fitness evaluation, selection, crossover, mutation and other steps, and iterate continuously until the termination condition is met. For example, after 100 iterations, the flight parameters represented by the individual with the highest fitness score are the optimized flight parameters.

[0103] The A-star algorithm is used for path planning, and its expression is as follows:

[0104]

[0105] in, Represents the actual cost from the starting point to the current node n; represents the heuristic estimated cost from the current node n to the target node; Represents the total evaluation cost, which is used to determine the priority of the node.

[0106] The shooting module uses drones to shoot and obtain farmland images based on the flight path.

[0107] If the real-time monitored light intensity and cloud obstruction meet the preset conditions, the UAV is controlled to fly along the flight path to obtain farmland images. Otherwise, the flight altitude, speed and heading of the UAV are adjusted until the preset conditions are met to obtain high-quality farmland images.

[0108] The sensor obtains the light intensity and cloud cover data of the drone's flight area in real time, and compares the obtained data with the preset conditions to determine whether the preset conditions are met. If the preset conditions are met, the drone is controlled to fly according to the predetermined flight path, and the high-definition camera on the drone obtains farmland images, and the obtained image data is transmitted to the ground control center. If the preset conditions are not met, the fuzzy control algorithm is used according to the obtained light intensity and cloud cover data to dynamically adjust the drone's flight altitude, speed and heading until the preset conditions are met.

[0109] The processing module pre-processes the farmland images to obtain processed data.

[0110] Obtain the original image data of the field collected by the drone, and use the image stitching algorithm to stitch the original images to generate a preliminary stitched image. Specifically, extract the feature points in the image to be stitched and pair them. Use the matched feature point pairs to calculate the homography transformation matrix H. The homography transformation matrix is ​​a 3x3 matrix that describes the geometric transformation relationship between the two images. The expression is as follows:

[0111]

[0112] Among them, x i and x j Respectively represent the matching feature points in the two images; Represents the square of the Euclidean distance.

[0113] For the preliminary stitching images, the image fusion algorithm is used to perform smooth transition processing on the stitching area, eliminate the stitching traces, and obtain the preliminary fused image. According to the preset image quality evaluation index, the quality of the preliminary fused image is evaluated. If the quality requirements are met, the image is output as a complete image covering the entire field; if the quality requirements are not met, the stitching process is returned and re-executed.

[0114] The above image fusion algorithm adopts a multi-resolution pyramid image fusion algorithm, and the steps include:

[0115] Construct a Gaussian pyramid: Construct a Gaussian pyramid for each of the two input images, and the image resolution of each layer gradually decreases.

[0116] Construct a Laplacian pyramid: Subtract the downsampled and upsampled image from each layer of the Gaussian pyramid to obtain a Laplacian pyramid.

[0117] Fusing Laplacian Pyramids: The Laplacian pyramids of two images are fused at each level, usually using a simple weighted average or a more complex fusion strategy.

[0118] Reconstruct the image: Start from the bottom layer of the fused Laplacian pyramid, gradually upsample and accumulate, and finally get the fused image.

[0119] The recognition module inputs the processed data into the constructed recognition model to perform crop category recognition and obtain recognition results.

[0120] This implementation uses the CNN network results to build a recognition model, including: input layer, convolution layer 1, pooling layer 1, convolution layer 2, pooling layer 2, fully connected layer and output layer. Its specific structure is as follows:

[0121] Convolutional layer 1: The number of filters is 32, the filter size is 3×3; the activation function uses ReLU; the step size is 1;

[0122] Pooling layer 1: pooling window size is 2×2; stride is 2;

[0123] Convolutional layer 2: The number of filters is 64; the filter size is 3×3; the activation function is ReLU; the step size is 1;

[0124] Pooling layer 2: pooling window size is 2×2; step size is 2;

[0125] Fully connected layer: the number of neurons is 128; the activation function is ReLU;

[0126] Output layer: Number of neurons: N (in this embodiment, N is the number of crop categories); activation function is Softmax.

[0127] The verification module generates a verification report based on the identification results and completes the verification.

[0128] Based on these analysis results, agricultural experts can generate detailed farmland verification reports. The report not only contains a quantitative description of crop distribution and growth conditions, but also provides targeted management suggestions. For example, for areas with low DVI values, farmers are advised to carry out pest control or supplementary fertilization in a timely manner. For areas with good growth, inputs can be appropriately reduced to improve resource utilization efficiency. This farmland verification report provides a basis for scientific decision-making for agricultural management departments. Management departments can adjust the planting structure based on the report, such as increasing the planting area of ​​high-yield crops, or optimize resource allocation, such as adjusting the irrigation system to improve water resource utilization efficiency. In this way, refined management of agricultural production can be achieved, yield and quality can be improved, while reducing environmental impact and promoting sustainable agricultural development.

[0129] The technical solution of the present invention ensures that the drone can efficiently and comprehensively obtain farmland images by designing a flight path covering the entire field based on the farmland area and terrain characteristics data. It not only improves the accuracy of crop category recognition, but also can quickly generate detailed verification reports, providing a basis for scientific decision-making for agricultural management departments. This not only helps to improve the level of refined management of agricultural production, but also optimizes resource allocation, improves yield and quality, while reducing environmental impact and promoting the sustainable development of agriculture.

[0130] It should be noted that the method of the embodiment of the present disclosure can be performed by a single device, such as a computer or a server. The method of the present embodiment can also be applied in a distributed scenario and completed by multiple devices cooperating with each other. In the case of such a distributed scenario, one of the multiple devices can only perform one or more steps in the method of the embodiment of the present disclosure, and the multiple devices will interact with each other to complete the described method.

[0131] The system of the above embodiment is used to implement the corresponding drone-based farmland verification method in any of the above embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0132] It should be noted that the above-mentioned drone-based farmland inspection system is embodied in the form of functional units. The term "module" here can be implemented in the form of software and / or hardware, and is not specifically limited to this.

[0133] For example, a "module" may be a software program, a hardware circuit, or a combination of the two that implements the above functions. The hardware circuit may include an application specific integrated circuit (ASIC), an electronic circuit, a processor (such as a shared processor, a dedicated processor, or a group processor, etc.) and a memory for executing one or more software or firmware programs, a combined logic circuit, and / or other suitable components that support the described functions.

[0134] Embodiment 3

[0135] Based on the same inventive concept, corresponding to any of the above-mentioned embodiments, the present disclosure also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the drone-based farmland inspection method described in any of the above embodiments is implemented.

[0136] Figure 2 A more specific schematic diagram of the hardware structure of an electronic device provided in this embodiment is shown, and the device may include: a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040 are connected to each other through the bus 1050 in the device.

[0137] The processor 1010 can be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0138] The memory 1020 may be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 may store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program codes are stored in the memory 1020 and are called and executed by the processor 1010.

[0139] The input / output interface 1030 is used to connect the input / output module to realize information input and output. The input / output module can be configured in the device as a component (not shown in the figure), or it can be externally connected to the device to provide corresponding functions. The input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, etc.

[0140] The communication interface 1040 is used to connect a communication module (not shown in the figure) to realize communication interaction between the device and other devices. The communication module can realize communication through a wired mode (such as USB (Universal Serial Bus), network cable, etc.), or through a wireless mode (such as mobile network, WIFI (Wireless Fidelity), Bluetooth, etc.).

[0141] The bus 1050 includes a path that transmits information between the various components of the device (eg, the processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040).

[0142] It should be noted that, although the above device only shows the processor 1010, the memory 1020, the input / output interface 1030, the communication interface 1040 and the bus 1050, in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include the components necessary for implementing the embodiments of the present specification, and does not necessarily include all the components shown in the figure.

[0143] The system of the above embodiment is used to implement the corresponding drone-based farmland verification method in any of the above embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0144] Embodiment 4

[0145] Based on the same inventive concept, corresponding to any of the above-mentioned embodiment methods, the present disclosure also provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the drone-based farmland verification method as described in any of the above embodiments.

[0146] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device.

[0147] The computer instructions stored in the storage medium of the above embodiment are used to enable the computer to execute the drone-based farmland verification method as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0148] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present disclosure (including the claims) is limited to these examples. Based on the concept of the present disclosure, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the embodiments of the present disclosure as described above, which are not provided in detail for the sake of simplicity.

[0149] In addition, to simplify the description and discussion, and in order not to make the embodiments of the present disclosure difficult to understand, the known power / ground connections to the integrated circuit (IC) chips and other components may or may not be shown in the provided figures. In addition, the device can be shown in the form of a block diagram to avoid making the embodiments of the present disclosure difficult to understand, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform on which the embodiments of the present disclosure will be implemented (that is, these details should be fully within the scope of understanding of those skilled in the art). Where specific details (e.g., circuits) are set forth to describe exemplary embodiments of the present disclosure, it is apparent to those skilled in the art that the embodiments of the present disclosure can be implemented without these specific details or with changes in these specific details. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0150] Although the present disclosure has been described in conjunction with specific embodiments of the present disclosure, many replacements, modifications and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.

[0151] Therefore, the units of each example described in the embodiments of the present application can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present application.

[0152] The embodiments of the present disclosure are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present disclosure should be included in the scope of protection of the present disclosure.

Claims

1. A method for farmland inspection based on drones, characterized in that: The method comprises: Based on the farmland area and terrain feature data, a flight path covering the entire field is designed; the steps include: based on the farmland area and terrain feature data, obtaining key terrain feature parameters; optimizing the key terrain feature parameters, and using a path planning algorithm to design a flight path covering the entire field; Based on the flight path, a drone is used to shoot and obtain images of the farmland; when the real-time monitored light intensity and cloud and fog obstruction meet the preset conditions, the drone is controlled to fly along the flight path to obtain images of the farmland; otherwise, the flight altitude, speed and heading of the drone are adjusted until the preset conditions are met; Preprocessing the farmland images to obtain processed data; stitching the original images using an image stitching algorithm to generate a preliminary stitching image; using an image fusion algorithm to perform smooth transition processing on the stitching area to eliminate stitching traces and obtain a preliminary fused image; The processed data is input into a constructed recognition model to perform crop category recognition to obtain a recognition result; the recognition model is constructed using a CNN network; Based on the identification results, a verification report is generated to complete the verification.

2. The method for farmland inspection based on drone according to claim 1, characterized in that: The key terrain parameters include: the highest point, the lowest point, and the average slope; based on the key terrain parameters, the flight parameters are optimized using a genetic algorithm, and the steps include: Parameter encoding: encoding the flight parameters to form an individual gene sequence; Initialize the population: randomly generate a number of individuals to form an initial population; Fitness evaluation: Based on the flight parameters of each individual, simulate the flight of the drone and obtain the farmland image data, and then calculate the fitness of each individual based on the clarity, resolution, and coverage completeness of the image data; Selection operation: Roulette selection method is used to calculate the probability of each individual being selected according to the fitness score; Crossover operation: Randomly select two parent individuals, perform single-point crossover at the middle position of the gene sequence, and generate two offspring individuals; Mutation operation: Gene mutation is performed on the newly generated offspring individuals with a preset probability; Iterative optimization: Repeat fitness evaluation, selection, crossover, and mutation, and iterate continuously until the termination condition is met.

3. The method for farmland inspection based on drone according to claim 1, characterized in that: The A-star algorithm is used for path planning, and its expression is as follows: in, Represents the actual cost from the starting point to the current node n; represents the heuristic estimated cost from the current node n to the target node; Represents the total evaluation cost, which is used to determine the priority of the node.

4. The farmland inspection method based on drone according to claim 1, characterized in that: The acquired farmland images are stitched together using an image stitching algorithm to generate a preliminary stitched image: Where H represents the homography transformation matrix; x i and x j Respectively represent the matching feature points in the two images; Represents the square of the Euclidean distance.

5. The method for farmland inspection based on drone according to claim 1, characterized in that: The recognition model constructed includes: Convolutional layer 1: The number of filters is 32, the filter size is: 3×3; the activation function uses ReLU; the step size is 1; the input layer, convolutional layer 1, pooling layer 1, convolutional layer 2, pooling layer 2, fully connected layer and output layer, its structure is as follows: Pooling layer 1: pooling window size is 2×2; stride is 2; Convolutional layer 2: The number of filters is 64; the filter size is 3×3; the activation function is ReLU; the step size is 1; Pooling layer 2: pooling window size is 2×2; step size is 2; Fully connected layer: the number of neurons is 128; the activation function is ReLU; Output layer: Number of neurons: N, N is the number of crop categories; activation function is Softmax.

6. A farmland inspection system based on drones, the system being used to implement the method described in any one of claims 1 to 5, characterized in that: include: Design module, shooting module, processing module, recognition module and verification module; The design module is used to design a flight path covering the entire field based on the farmland area and terrain characteristics data; the steps include: obtaining key terrain characteristic parameters based on the farmland area and terrain characteristics data; optimizing the key terrain characteristic parameters, and using a path planning algorithm to design a flight path covering the entire field; The shooting module is used to shoot with a drone based on the flight path to obtain images of the farmland; when the real-time monitored light intensity and cloud and fog occlusion meet the preset conditions, the drone is controlled to fly along the flight path to obtain images of the farmland; otherwise, the flight altitude, speed and heading of the drone are adjusted until the preset conditions are met; The processing module is used to pre-process the farmland image to obtain processed data; stitch the original image through an image stitching algorithm to generate a preliminary stitching image; and use an image fusion algorithm to perform smooth transition processing on the stitching area to eliminate stitching traces and obtain a preliminary fused image; The recognition module is used to input the processed data into a constructed recognition model to perform crop category recognition and obtain a recognition result; the recognition model is constructed using a CNN network; The verification module is used to generate a verification report based on the identification result to complete the verification.

7. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed, the method according to any one of claims 1 to 5 is implemented.

Images