Agricultural machinery route planning system based on satellite remote sensing and unmanned aerial vehicle remote sensing

Abstract

The invention provides an agricultural machinery route planning system based on satellite remote sensing and unmanned aerial vehicle remote sensing, and belongs to the technical field of agricultural machinery route planning, and the system comprises an information acquisition unit which employs a satellite remote sensing or unmanned aerial vehicle remote sensing technology to acquire land parcel information of an agricultural land parcel, the land parcel information comprises a unique land parcel identifier, a position coordinate, an area size, a boundary shape rule degree and an agricultural terrain type of a corresponding agricultural land parcel, and the agricultural terrain type comprises a sentry land, a depression, a flat land and a slope land; the land parcel distribution unit is used for distributing agricultural land parcels to be operated for the agricultural machine terminal; the path planning unit is used for planning an operation path for the agricultural machine terminal according to the following modes: analyzing land parcel information by using an agricultural land parcel analysis method to obtain coordinate information, and generating an agricultural machine operation planning path according to the coordinate information in the agricultural land parcel information; the agricultural machinery operation path planning is automatically performed on a plurality of agricultural land parcels, the rationality and production efficiency of land operation are improved, and the disordered waste phenomenon of agricultural machinery is avoided.

Description

Technical Field

[0001] The present invention belongs to the technical field of agricultural machinery route planning, and in particular relates to an agricultural machinery route planning system based on satellite remote sensing and unmanned aerial vehicle remote sensing. Background Art

[0002] Smart agriculture is the advanced application of IoT technology in modern agriculture. It leverages sensors and software for remote monitoring and control of agricultural production through server, mobile, or computer platforms, making traditional agriculture more intelligent. Beyond precise sensing, control, and decision-making management, smart agriculture broadly encompasses agricultural e-commerce, food traceability and anti-counterfeiting, agricultural leisure tourism, and agricultural information services.

[0003] At present, power machinery is widely used in agriculture and has covered all aspects of farming. From the perspective of power machinery application, with the increase in planting density and agronomic requirements, the requirements for the path accuracy of agricultural machinery (i.e., a general term for power machinery in agriculture) are getting higher and higher. Through experiments, it was found that the optimal path accuracy when operated by humans can reach 10cm, while after applying RTK differential technology (Real Time Kinematic, carrier phase differential technology) and installing an electronic steering system, the path accuracy of agricultural machinery can reach 2.5cm. The accuracy is greatly improved, which effectively reduces the intensity of human work and also reduces the experience requirements of operators. As a result, more and more farmers are using automatic navigation to plan the paths of agricultural machinery operations.

[0004] Currently, many agricultural service providers have emerged in agricultural production, offering services across various farming operations, including tillage, sowing, and pesticide application. During these services, they coordinate the simultaneous operation of multiple agricultural machines. As efficiency improves in this process, problems with operational planning and management emerge. Machine operators are unfamiliar with the different plots of land, often relying on intuition or following farmers' preferences when selecting the order in which to operate them. This can result in delayed sowing in plots that should have been planted first and premature harvesting in plots that should have been harvested later, impacting productivity.

[0005] Due to crop growth characteristics and agricultural practices, different agricultural standards may be applied to the same crops sown on plots of land with different conditions. For example, sowing on hilly land should begin earlier than on low-lying land, and sowing on dry land can be delayed until after sowing on irrigated land. Large plots of land should be cultivated before smaller plots, and plots with regular boundaries should be worked on before those with irregular boundaries. Agricultural machinery should be operated on the basis of proximity, completing work on plots in its area before moving long distances to avoid reducing effective working time. Based on these agricultural and agronomic requirements, agricultural service operators need to plan their work plots appropriately when using agricultural machinery.

[0006] However, current automated navigation systems implement production navigation operations based on a user-sent path location, allowing for navigation within a single plot of land. To meet the needs of agricultural service operators for efficient operations on unfamiliar plots, it is necessary to automatically plan agricultural machinery operation paths across multiple plots, incorporating agronomic standards. This approach can guide agricultural service operators and improve the rationality and efficiency of their land operations. Summary of the Invention

[0007] In response to the problems of the prior art, the present invention provides an agricultural machinery route planning system based on satellite remote sensing and UAV remote sensing, comprising: An information acquisition unit, using satellite remote sensing or drone remote sensing technology to acquire plot information of the agricultural plot, wherein the plot information includes a unique plot identifier, location coordinates, area, regularity of boundary shape, and agricultural terrain type of the corresponding agricultural plot, wherein the agricultural terrain type includes hillock, depression, flat land, and sloping land; A plot allocation unit is used to allocate agricultural plots to be operated for agricultural machinery terminals; The path planning unit is used to plan the operation path for the agricultural machinery terminal in the following manner: the plot information is analyzed using the agricultural plot analysis method to obtain coordinate information, and the agricultural machinery operation planning route is generated based on the coordinate information in the agricultural plot information; the collision situation between the agricultural machinery and the obstacle is converted into the state representation of the neuron through convolution calculation and threshold setting, and the neurons in the target state and obstacle state are used as the excitatory and inhibitory inputs of the neural network; based on the definition of internal neural connections, the generated neural network activity space ensures that the neurons in the target state remain at the peak, so that the agricultural machinery can be globally attracted to the target state; at the same time, it ensures that the neurons in the obstacle state remain at the bottom to avoid collision; by sequentially selecting adjacent neurons with the maximum activity intensity, a collision-free path from the starting position to the target position is generated; in order to ensure that the collision-free path can meet the optimal path, the ant colony algorithm is used to search for the cultivated land operation path, and the path selection tendency and path pheromone are updated; the path selection tendency is the selection probability of the path, and the ant colony algorithm is used to select the next operation path for the cultivated land operation; A wireless communication unit, configured to push the operation task type and the planned agricultural machinery operation route to the agricultural machinery terminal, wherein the operation task type is land preparation operation, seeding operation, pesticide spraying operation or harvesting operation; The agricultural machinery terminal is wirelessly connected to the server for outputting and displaying the operation task type and the planned operation route of the agricultural machinery.

[0008] Furthermore, satellite remote sensing or UAV remote sensing technology is used to obtain plot information of agricultural plots. Satellite remote sensing technology uses a watershed network and remote sensing image derived from a digital terrain model (DEM). The farmland network derived from the digital terrain model (DEM) is used as a spatial constraint condition. The optical remote sensing image farmland index method is combined with adaptive threshold segmentation to realize the automated farmland terrain extraction of each image within the study period.

[0009] Furthermore, the UAV remote sensing technology includes a UAV image processing module, and the interior of the UAV image processing module is fixedly installed with a uniform color cropping module, a panoramic image stitching processing module and a low-altitude remote sensing image matching module. The output ends of the uniform color cropping module, the panoramic image stitching processing module and the low-altitude remote sensing image matching module are connected to the input end of the UAV image processing module.

[0010] Furthermore, the data acquisition module is compatible with and continues to use various types of single-beam echo sounding equipment and RTK positioning equipment, as well as industrial computers and data acquisition software, which are popular among various terrain surveying units. After the data acquisition software collects GPS positioning data and topograph data files, the system's newly added survey line data file return function will encrypt the collected data files and return them to the channel terrain data service of the designated IP and port through the network.

[0011] Furthermore, the image cropping software and system are used to crop the image information irrelevant to the edge areas; after the panoramic image stitching processing module generates the digital elevation model orthophoto map through aerial triangulation, the orientation elements are measured to enable the image matching to produce more discrete three-dimensional micro-points, and the orthophoto is obtained through human-computer interaction; after obtaining the orthophoto, the accuracy needs to be adjusted, and the detection control points on the ground are randomly selected. The detection results are listed in the orthophoto accuracy check table. At the same time, the detection points are numbered, and the coordinate difference and coordinate offset are recorded to effectively analyze the accuracy of the orthophoto.

[0012] Furthermore, the possibility of application after image processing is analyzed again based on the experimental results; the low-altitude remote sensing image matching module uses the correlation function relationship to find similar structures, and calculates the coordinate position of the image and the mapping transformation through mathematical definitions; during matching, the measurement of remote sensing image matching is also monitored, through functional relationships and correlation coefficients and covariance functions. From the perspective of matching strategy, the level of matching allocation treats the low-altitude remote sensing image as a unified and coordinated whole. In several information processing processes, after analyzing the global situation, the image of each layer is analyzed and fused in detail through structural information, and finally the disparity map is tested for consistency.

[0013] Compared with the prior art, the advantages and positive effects of the present invention are: The present invention provides an agricultural machinery route planning system based on satellite remote sensing and UAV remote sensing, which uses satellite remote sensing and UAV remote sensing to first collect the plot information of all agricultural plots, and then allocate the agricultural plots to be operated to the agricultural machinery terminals. Then, according to the plot information and combined with the agricultural and agronomic standards, the agricultural machinery operation paths are automatically planned for multiple agricultural plots. Finally, the operation task type and the agricultural machinery operation plan route are pushed to the operator of the agricultural machinery terminal, thereby guiding agricultural service operators to perform operations, improving the rationality and production efficiency of land operations, and avoiding the disorderly waste of agricultural machinery; it can provide agricultural machinery operators with convenience in operation sequence strategies when performing agricultural operations, maximize the use efficiency of agricultural machinery and meet the agricultural and agronomic needs of land crops, specifically for general agricultural machinery Path planning methods suffer from poor robustness, inaccurate path search, and insufficient obstacle avoidance capabilities. This paper proposes a method that uses different neural network spaces as the movement space for agricultural machinery. By continuously optimizing the connection weights between external inputs and neurons, the neural network-based path search is parallelized and adaptively adjusted to achieve accurate path planning and improve obstacle avoidance performance. Adaptive inertia weights are designed based on individual fitness values, global fitness values, and individual historical fitness values, leading to an adaptive parameter position movement strategy. An ant colony algorithm is then used to optimize the scheduling sequence of cultivated land operation paths, minimizing the total turning distance of the tractor during operation. Furthermore, a method for dynamically adjusting pheromone volatilization in an ant colony algorithm based on the sigmoid function is studied to address the problem of traditional ant colony algorithms being prone to falling into local optimality. Furthermore, considering the weak global search capabilities of traditional ant colonies, a harmonious search algorithm is used to determine the optimal parameter combination of the pheromone factor and heuristic factor in the ant colony algorithm. Compared with traditional ant colony algorithms and elite ant colony algorithms, the proposed improved harmonious ant colony algorithm significantly reduces the number of iterations, improves path search efficiency, and significantly improves algorithm performance in large-scale field scenarios of varying sizes. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0015] Figure 1 This is a structural diagram of the agricultural machinery route planning system based on satellite remote sensing and UAV remote sensing of the present invention. DETAILED DESCRIPTION

[0016] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is further described below in conjunction with the accompanying drawings and embodiments. It should be noted that the embodiments of the present application and the features therein can be combined with each other without conflict.

[0017] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways than those described herein. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0018] Example 1, as Figure 1 As shown, this application provides an agricultural machinery route planning system based on satellite remote sensing and UAV remote sensing, including: An information acquisition unit, using satellite remote sensing or drone remote sensing technology to acquire plot information of the agricultural plot, wherein the plot information includes a unique plot identifier, location coordinates, area, regularity of boundary shape, and agricultural terrain type of the corresponding agricultural plot, wherein the agricultural terrain type includes hillock, depression, flat land, and sloping land; A plot allocation unit is used to allocate agricultural plots to be operated for agricultural machinery terminals; The path planning unit is used to plan an operation path for the agricultural machinery terminal in the following manner: using an agricultural plot analysis method to analyze plot information to obtain coordinate information, and generating an agricultural machinery operation planning route based on the coordinate information in the agricultural plot information; A wireless communication unit, configured to push the operation task type and the planned agricultural machinery operation route to the agricultural machinery terminal, wherein the operation task type is land preparation operation, seeding operation, pesticide spraying operation or harvesting operation; The agricultural machinery terminal is wirelessly connected to the server for outputting and displaying the operation task type and the planned operation route of the agricultural machinery.

[0019] Satellite or drone remote sensing technology is used to obtain plot information for agricultural land. Satellite remote sensing technology uses a watershed network derived from a digital terrain model (DEM) and remote sensing imagery to automatically extract and map seasonal rivers during a specified time period. Using the DEM-derived farmland network as a spatial constraint, the optical remote sensing imagery farmland index method combined with adaptive threshold segmentation enables automated farmland topography extraction from each image within the study period. The frequency of farmland topography occurrence within the time period is calculated to overcome the monitoring difficulties caused by the dynamic and variable nature of seasonal farmland. Subsequently, object-based morphological operations are used to generate the centerline of the farmland topography mask and estimate the length, width, area, and other indicators of the farmland topography under various conditions.

[0020] The UAV remote sensing technology includes a UAV image processing module, and the interior of the UAV image processing module is fixedly installed with a uniform color cropping module, a panoramic image stitching processing module and a low-altitude remote sensing image matching module. The output ends of the uniform color cropping module, the panoramic image stitching processing module and the low-altitude remote sensing image matching module are connected to the input end of the UAV image processing module.

[0021] The data acquisition module is compatible with and continues to work with various single-beam bathymetry equipment and RTK positioning devices, as well as industrial computers and data acquisition software, commonly used by topographic surveying units. After the data acquisition software collects GPS positioning data and topographer data files (in survey line files), the system's newly added survey line data file return function encrypts the collected data files and transmits them back over the network to a channel topography data service with a designated IP and port. The color grading and trimming module applies color grading to the original image, paying attention to color contrast, grayscale, and texture variations to ensure a natural transition between images after grading.

[0022] At the same time, image cropping software and systems can be used to crop irrelevant image information from edge areas. After the panoramic image stitching processing module generates a digital elevation model orthophoto through aerial triangulation, it determines orientation elements to enable image matching to produce a large number of discrete three-dimensional micropoints, thereby obtaining an orthophoto through human-computer interaction. After obtaining the orthophoto, accuracy adjustment is required. Ground control points are randomly selected for inspection, and the inspection results are compiled into an orthophoto accuracy check table. The inspection points are numbered, and the coordinate differences and offsets are recorded to effectively analyze the orthophoto accuracy.

[0023] Furthermore, experimental results can be used to further analyze the potential applications of image processing. The low-altitude remote sensing image matching module utilizes correlation functions to identify similar structures between them, mathematically defining and calculating the image coordinate positions and mapping transformations. During matching, the remote sensing image matching metrics are monitored, calculated using functional relationships, correlation coefficients, and covariance functions. From a matching strategy perspective, the matching can be performed at the assigned level, treating the low-altitude remote sensing image as a unified and coordinated whole. After a global analysis is performed during several information processing steps, each layer of imagery is analyzed and fused in detail using structural information, ultimately performing a consistency check on the disparity map.

[0024] The neuron model senses external inputs and adjusts neuron activity based on the input strength and connection weights. Agricultural machinery implements path planning and obstacle avoidance based on environmental changes. The algorithm for planning the operation path of the agricultural machinery terminal is as follows: Where A is a parameter that controls the degree of response to input, B is a bias term, and D is an additional offset; xi is the activity of a neuron, t is the time step, zi and zj are the input influence terms of different neurons; [·]- is defined as [a]-=max{-a, 0} ; Ii is the response speed.

[0025] Path generation, through convolution calculation and threshold setting, converts the collision situation between agricultural machinery and obstacles into neuron state representation, and uses the neurons in the target state and obstacle state as the excitatory and inhibitory inputs of the neural network; based on the definition of internal neural connections, the generated neural network activity space ensures that the neurons in the target state remain at the peak, thereby being able to globally attract agricultural machinery towards the target state; at the same time, it ensures that the neurons in the obstacle state remain at the bottom to avoid collision; by sequentially selecting adjacent neurons with the maximum activity intensity, a collision-free path from the starting position to the target position is generated.

[0026] In order to ensure that the collision-free path can meet the optimal path, the ant colony algorithm is used to search for the farmland operation path, and the path selection tendency and path pheromone are updated. The path selection tendency is the probability of path selection. The ant colony algorithm is used to select the next operation line to be used for farmland operation, as shown in the following formula: Where: α is the pheromone heuristic factor; β is the expected heuristic factor; τij(t) is the pheromone concentration transferred from job row i to job row j when the ant searches for the path at time t; ηij is the heuristic function for the path to transfer from i to j; Ak is the set of nodes that the ant can move to in the next step.

[0027] As time t passes, after one iteration of the ant colony algorithm is completed, the pheromone on the path is volatilized and the pheromone update method on the tractor turning path (i, j) is as follows: Where: ρ is the pheromone volatilization factor; 1-ρ is the path pheromone residual factor; Δτij is the pheromone left by the ant in the iteration.

[0028] The ant colony algorithm is used to optimize the total turning path distance matrix of the tractor to form a tractor operation row scheduling sequence, and the pheromone concentration is updated. The selection of the tractor operation row scheduling sequence is repeated until the total turning distance of the tractor plowing operation is minimized. The optimal operation row scheduling sequence arrangement is obtained, and the tractor operation path optimization is completed: Where: Q is the pheromone action constant; Lk is the length of the k-th ant's walking path .

[0029] In the traditional ant colony algorithm, ant pheromones evaporate during the iteration process, resulting in the ants not updating their pheromones in time during the path search process, making it easy for the path search process to fall into local optimality, making it difficult to find the optimal path for random operations, and limiting the path optimization effect. Therefore, a pheromone dynamic adjustment strategy is introduced to improve the shortcomings of the traditional ant colony algorithm.

[0030] The update of ant colony pheromones affects the probability of tractor path selection, and a fixed pheromone volatility factor cannot optimize the algorithm performance. Based on this, this paper proposes a pheromone volatility update strategy to control the update of the ant colony pheromone volatility factor. By utilizing the characteristics of the sigmoid function, as the number of algorithm iterations increases, the pheromone volatility rate on the path gradually increases. The positive feedback effect of the information obtained by the algorithm is weakened, and the pheromone on the path with fewer search times increases, thereby improving the global search capability, as shown in the following formula: Where: Imax is the maximum number of iterations of the algorithm; I is the current iteration number of the algorithm; a and b are the parameters for adjusting the sigmoid function; ρnew is the updated pheromone volatility rate.

[0031] The introduction of parameters a and b allows for timely adjustment of the pheromone volatilization concentration range and volatilization speed. During the tractor's search for the shortest total turning distance operating path sequence, the sigmoid function is reconstructed using the iterative process of the ant colony algorithm. This aims to appropriately adjust and intervene in the pheromone volatilization intensity during the early, middle, and late stages of the ant colony algorithm, thereby preventing the algorithm from falling into a local optimum due to excessive differences in pheromone deposition during the search for different tractor operating paths.

[0032] The pheromone factor α reflects the relative importance of the information accumulated during ant movement in guiding the ant colony's path search. If its value is too large, the ants are more likely to choose previously traveled paths, reducing the randomness of the search; if it is too small, the search will prematurely fall into a local optimum. The heuristic function factor β represents the relative importance of heuristic information in guiding the ant colony's path search. This paper introduces a mechanism to adjust the pheromone factor α and the heuristic function factor β as the algorithm iterates, dynamically updating the heuristic factor, as shown in the following formula: Where: a, b, c are constants.

[0033] Because the harmony search algorithm works by mimicking the activities of an improvisator, it often finds the optimal input set for complex functions. The harmony search algorithm has a low time complexity. Based on its search mechanism, the individual perturbation strategy generates a new solution vector at each iteration, thereby increasing the diversity of harmonic combinations. To improve the diversity of the combinations of ant colony pheromone factors and heuristic function factors, the harmony search algorithm exploits the characteristic of continuously updating and replacing the combinations in the harmony memory with each new solution generated. With the objective function of minimizing the total turning distance of the unmanned tractor's plowing path, the optimal solution vector composed of pheromone factors and heuristic function factors is determined. The specific steps for improving the harmony ant colony algorithm are as follows.

[0034] 1) Initialize the main algorithm parameters: harmonic memory size (HMS), harmonic memory retention probability (HMCR), pitch adjustment rate (PAR), adjustment step size (BW) and maximum number of loops.

[0035] 2) The total turning distance of the unmanned tractor during plowing operation is selected as the objective function F(x).

[0036] 3) Construct a harmonic memory library. Obtain the heuristic factor and pheromone factor at each iteration through the iterative relationship and combine them into the solution vector X(x1, x2, x3…xImax) of the objective function F(x).

[0037] 4) Update the harmony memory. Based on the generated new solution vector, a new harmony vector is improvised from the HM using two methods: HMCR considerations and PAR combined with pitch adjustment. This is then compared with the solution harmony vector. If the resulting harmony is better than the worst harmony in the HM, the harmony vector Xi is updated.

[0038] 5) The solution obtained from each iteration of the harmony search is used to update the pheromone factor α and the heuristic factor β in the ant colony algorithm.

[0039] To address the poor robustness, inaccurate path search, and insufficient obstacle avoidance capabilities of conventional agricultural machinery path planning methods, this paper proposes a neural network space as the movement space for the agricultural machinery. By continuously optimizing the connection weights between external inputs and neurons, the neural network-based path search is parallelized and adaptively adjusted to achieve accurate path planning and improve obstacle avoidance performance. Adaptive inertia weights are designed based on individual fitness values, global fitness values, and individual historical fitness values, leading to an adaptive parameter position movement strategy. An ant colony algorithm is then used to optimize the scheduling sequence of cultivated land operation paths, minimizing the total turning distance of the tractor during operation. Furthermore, to address the problem of traditional ant colony algorithms prone to falling into local optimality, a method for dynamically adjusting pheromone volatilization in the ant colony algorithm based on the characteristics of the sigmoid function is studied. Furthermore, considering the weak global search capabilities of traditional ant colonies, a harmony search algorithm is used to determine the optimal parameter combination of the pheromone factor and heuristic factor in the ant colony algorithm. Compared with traditional ant colony algorithms and elite ant colony algorithms, the proposed improved harmony ant colony algorithm significantly reduces the number of iterations, improves the path search efficiency, and significantly improves algorithm performance in large-scale field scenarios of varying sizes.

[0040] The above description is merely a preferred embodiment of the present invention and does not limit the present invention in any other form. Any technician familiar with the present invention may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes for application in other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. An agricultural machinery route planning system based on satellite remote sensing and UAV remote sensing, characterized in that: include: An information acquisition unit, using satellite remote sensing or drone remote sensing technology to acquire plot information of the agricultural plot, wherein the plot information includes a unique plot identifier, location coordinates, area, regularity of boundary shape, and agricultural terrain type of the corresponding agricultural plot, wherein the agricultural terrain type includes hillock, depression, flat land, and sloping land; A plot allocation unit is used to allocate agricultural plots to be operated for agricultural machinery terminals; The path planning unit is used to plan the operation path for the agricultural machinery terminal in the following manner: the plot information is analyzed using the agricultural plot analysis method to obtain coordinate information, and the agricultural machinery operation planning route is generated based on the coordinate information in the agricultural plot information; the collision situation between the agricultural machinery and the obstacle is converted into the state representation of the neuron through convolution calculation and threshold setting, and the neurons in the target state and obstacle state are used as the excitatory and inhibitory inputs of the neural network; based on the definition of internal neural connections, the generated neural network activity space ensures that the neurons in the target state remain at the peak, so that the agricultural machinery can be globally attracted to the target state; at the same time, it ensures that the neurons in the obstacle state remain at the bottom to avoid collision; by sequentially selecting adjacent neurons with the maximum activity intensity, a collision-free path from the starting position to the target position is generated; in order to ensure that the collision-free path can meet the optimal path, the ant colony algorithm is used to search for the cultivated land operation path, and the path selection tendency and path pheromone are updated; the path selection tendency is the selection probability of the path, and the ant colony algorithm is used to select the next operation path for the cultivated land operation; A wireless communication unit, configured to push the operation task type and the planned agricultural machinery operation route to the agricultural machinery terminal, wherein the operation task type is land preparation operation, seeding operation, pesticide spraying operation or harvesting operation; The agricultural machinery terminal is wirelessly connected to the server for outputting and displaying the operation task type and the planned operation route of the agricultural machinery.

2. The agricultural machinery route planning system based on satellite remote sensing and UAV remote sensing according to claim 1, characterized in that: Satellite remote sensing or UAV remote sensing technology is used to obtain plot information of agricultural plots. Satellite remote sensing technology uses the watershed network and remote sensing image derived from the digital terrain model (DEM). The farmland network derived from the digital terrain model (DEM) is used as the spatial constraint condition. The optical remote sensing image farmland index method is combined with adaptive threshold segmentation to realize the automated farmland terrain extraction of each image within the study period.

3. The agricultural machinery route planning system based on satellite remote sensing and UAV remote sensing according to claim 1, characterized in that: UAV remote sensing technology includes a UAV image processing module, inside which a uniform color trimming module, a panoramic image stitching processing module and a low-altitude remote sensing image matching module are fixedly installed, and the output ends of the uniform color trimming module, the panoramic image stitching processing module and the low-altitude remote sensing image matching module are connected to the input end of the UAV image processing module.

4. The agricultural machinery route planning system based on satellite remote sensing and UAV remote sensing according to claim 3, characterized in that: The data acquisition module is compatible with and continues to support various types of single-beam echo sounding equipment and RTK positioning equipment, as well as industrial computers and data acquisition software popular among various terrain surveying units. After the data acquisition software collects GPS positioning data and topograph data files, the system's newly added survey line data file return function encrypts the collected data files and returns them to the channel terrain data service of the designated IP and port through the network.

5. The agricultural machinery route planning system based on satellite remote sensing and UAV remote sensing as claimed in claim 4, characterized in that: Image cropping software and systems are used to crop irrelevant image information from edge areas. After the panoramic image stitching processing module generates a digital elevation model orthophoto through aerial triangulation, the orientation elements are determined to enable image matching to produce more discrete three-dimensional micro-points, and orthophotos are obtained through human-computer interaction. After obtaining the orthophoto, the accuracy needs to be adjusted. Randomly select ground detection control points, list the detection results into an orthophoto accuracy check table, number the detection points, and record the coordinate difference and coordinate offset to effectively analyze the accuracy of the orthophoto.

6. The agricultural machinery route planning system based on satellite remote sensing and UAV remote sensing according to claim 5, characterized in that: Based on the experimental results, the possibility of application after image processing is analyzed again; the low-altitude remote sensing image matching module uses the correlation function relationship to find similar structures, and calculates the coordinate position of the image and the mapping transformation through mathematical definitions; during matching, the measurement of remote sensing image matching must also be monitored, through functional relationships and correlation coefficients and covariance functions. From the perspective of matching strategy, the level of matching allocation treats the low-altitude remote sensing image as a unified and coordinated whole. In several information processing processes, after analyzing the global situation, the image of each layer is analyzed and fused in detail through structural information, and finally the disparity map is tested for consistency.

Images