A cultivated land patrol method and system based on geographical entity space identity coding

Abstract

The application discloses a kind of based on the cultivated land patrol method and system of geographical entity space identity coding, method includes the following steps: step S1: target cultivated land is assigned geographical entity space identity coding, cultivated land core information and the state identification of visual two-dimensional code containing are generated based on geographical entity space identity coding, geographical entity space identity coding and two-dimensional code information are stored in distributed database;Step S2: obtain unmanned aerial vehicle patrol image, and carry out pretreatment and feature extraction, detect cultivated land non-agricultural, non-food illegal behavior by intelligent identification model, generate the suspected illegal clues of associated unique code;Step S3: identification is carried out based on suspected illegal clues, and dynamically update two-dimensional code state according to identification result.

Description

Technical Field

[0001] This application relates to the field of farmland inspection technology, and in particular to a farmland inspection method and system based on geographic entity spatial identity coding. Background Technology

[0002] Regular and precise inspections and supervision of arable land are crucial for ensuring food security, but existing methods have significant shortcomings. Manual inspections are inefficient, satellite remote sensing lacks timeliness, and video surveillance is costly; furthermore, these methods lack coordination, failing to form a closed-loop supervision system. Simultaneously, the lack of a unified identification coding system makes it difficult to accurately link inspection data with specific land parcels, resulting in vague identification of violations. More importantly, the entire process from discovery to rectification lacks an effective tracking mechanism, leading to insufficient standardization and traceability in supervision.

[0003] Therefore, there is an urgent need to develop a smart field patrol technology solution with the unique geographical entity spatial identity code of arable land as the core, so as to achieve accurate identity association, full-area collaborative patrol, intelligent and accurate identification, and full-process traceability and disposal. Summary of the Invention

[0004] This application provides a method and system for farmland inspection based on geographic entity spatial identity coding to solve the above-mentioned problems.

[0005] On the one hand, this application provides a method for farmland inspection based on geographic entity spatial identity coding, the method comprising the following steps: Step S1: Assign a geographic entity spatial identity code to the target farmland, generate a visual QR code containing core farmland information and status identifier based on the geographic entity spatial identity code, and store the geographic entity spatial identity code and QR code information in a distributed database. Step S2: Acquire drone patrol images, perform preprocessing and feature extraction, detect illegal activities such as non-agricultural and non-grain use of arable land through intelligent recognition models, and generate suspected illegal clues with unique associated codes; Step S3: Identify suspected illegal clues and dynamically update the QR code status based on the identification results.

[0006] In one implementation of this application, in step S1, the proprietary identifier field of the geographic entity spatial identity code includes a root identifier code and a geographic entity-specific code, the standard field includes a location code, a classification code that distinguishes cultivated land types, and a sequence code, and the extended field is a hybrid code used to expand the functions or attributes of cultivated land.

[0007] In one implementation of this application, in step S2, the UAV patrol automatically creates differentiated patrol tasks with unique codes based on the farmland data associated with the codes and the patrol equipment status data. It plans the optimal flight route covering the farmland area corresponding to the code, allocates tasks through a distributed scheduling algorithm, and the patrol equipment automatically flies along the flight route and takes high-definition images. All patrol data is archived with the unique farmland code. The differentiated patrol tasks support two creation modes: batch creation by regional code or precise creation by the unique code of a single farmland. The task parameters include farmland location, patrol range, shooting accuracy, and patrol frequency matching the QR code status.

[0008] In one implementation of this application, the optimal route planning takes the center point of the cultivated land entity as the core, and uses obstacles and the endurance of the inspection equipment as constraints. The route is shortest and completely covers the cultivated land area through grid map construction, cost function calculation and curve fitting.

[0009] In one implementation of this application, in step S2, the intelligent recognition model adopts an architecture of lightweight backbone network + attention mechanism + multi-scale feature fusion. By removing redundant detection boxes, the recognition results are optimized to improve the feature extraction capability of illegal areas and the recognition accuracy of small target violations.

[0010] In one implementation of this application, in step 3, the violations include illegal construction, hardening of the ground, abandonment of land, digging ponds for fish farming, and planting of non-grain crops. Suspected violations are determined based on the identification confidence level, and clues are pushed out.

[0011] In one implementation of this application, in step S3, the QR code status update rule is as follows: when AI identifies and determines that there is a suspected violation, it is updated to yellow; when manual verification confirms the violation, it is updated to red; after rectification and acceptance, it is updated to blue; and when the violation is ruled out, it is restored to green.

[0012] In one implementation of this application, in step S4, the suspected illegal clues are sent to regulatory personnel via SMS push and platform real-time push, with alarm information including farmland code, QR code, violation category and location information.

[0013] In one implementation of this application, the method further includes: automatically increasing the patrol frequency, optimizing the shooting accuracy, and replanning the fine route for high-risk farmland and farmland with yellow / red QR codes associated with codes, thereby strengthening key supervision.

[0014] On the other hand, this application also provides a farmland inspection system based on geographic entity spatial identity coding, the system comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the aforementioned farmland inspection method based on geographic entity spatial identity coding.

[0015] This application provides a method and system for farmland inspection based on geographic entity spatial identity coding, which has the following beneficial effects: 1. Pioneering a "four-in-one" "on-the-spot" collaborative architecture, breaking through the inherent limitations of existing technologies: This invention integrates "farmland entity coding + drone "on-the-spot" inspection + AI "on-the-spot" identification + "on-the-spot" problem handling" to form a collaborative system. With the unique code of farmland as the core link, it achieves precise correlation of each link. Compared with the fragmentation of conventional farmland inspection and the singularity of ordinary inspection, it realizes the leap from "dispersed operation" to "integrated collaboration" and builds a core invention barrier.

[0016] 2. Full-process "code-based" automation significantly improves efficiency and accuracy: The drone "code-based" patrol subsystem automates the entire process, from task creation and route planning to execution and shooting, with all data linked and coded. Combined with the precise identification capabilities of AI "code-based" recognition, the accuracy of violation identification reaches over 95%, and the patrol efficiency is more than 50 times higher than that of manual patrols, solving the core problems of low efficiency and poor accuracy in conventional patrols.

[0017] 3. Dynamic linkage of multi-color QR codes to achieve full-process traceability: The status of cultivated land is dynamically identified by four colors of QR codes: green, yellow, red and blue. Combined with the unique code of cultivated land, the "code traces the whole process" is realized, which solves the defects of fragmented processing and difficulty in traceability of existing technologies and improves the standardization and traceability of cultivated land supervision.

[0018] 4. Multi-device compatibility + dynamic linkage scheduling, breaking through the bottleneck of ordinary inspection: Through standardized software protocol adaptation, multiple types of inspection equipment are networked. Combined with the status (QR code color) associated with cultivated land code, the inspection strategy is adjusted in linkage. It supports large-scale concurrent scheduling of hundreds of cultivated land plots, which not only improves the system scalability, but also strengthens the targeting of cultivated land supervision, and solves the defects of ordinary inspection that cannot be promoted on a large scale and lacks the accuracy of supervision. Attached Figure Description

[0019] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1A flowchart of a farmland inspection method based on geographic entity spatial identity coding provided in this application embodiment; Figure 2 This is a diagram illustrating the composition of a farmland inspection system based on geographic entity spatial identity coding, provided in an embodiment of this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0021] This application provides a method and system for farmland inspection based on geographic entity spatial identity coding. The technical solution proposed in this application will be described in detail below with reference to the accompanying drawings.

[0022] Figure 1 A flowchart illustrating a farmland inspection method based on geographic entity spatial identity coding, provided as an embodiment of this application. Figure 1 As shown, the method mainly includes the following steps: Step S1: Assign a geographic entity spatial identity code to the target farmland, generate a visual QR code containing core farmland information and status identifier based on the geographic entity spatial identity code, and store the geographic entity spatial identity code and QR code information in a distributed database. Step S2: Acquire drone patrol images, perform preprocessing and feature extraction, detect illegal activities such as non-agricultural and non-grain use of arable land through intelligent recognition models, and generate suspected illegal clues with unique associated codes; Step S3: Identify suspected illegal clues and dynamically update the QR code status based on the identification results.

[0023] Specifically, the spatial identity coding rules for geographic entities are as follows: Following the principles of uniqueness, practicality, compatibility, relative stability, and appropriate scalability, geographic entity coding rules are established to assign a unified spatial identity code to each piece of cultivated land. The geographic entity spatial identity coding adopts a three-segment coding paradigm: "property identifier domain + standard domain + extended domain," standardizing the coding structure and content of entity objects. Various feature codes in the entity spatial identity code are sequentially connected. For clear identification, hyphens are used between the proprietary identifier domain, standard domain, and extended domain. Within the extended domain, various feature codes are separated by the separator ".". Hyphens and separators can be omitted in application.

[0024] Furthermore, the drone patrol system uses farmland entities with unique geographic entity spatial identification codes as the sole patrol targets. It possesses fully automated capabilities across the entire process, including "code association, automatic creation of patrol tasks, automatic route planning, automatic task execution, and automatic data capture." All patrol data is bound to the unique farmland code, achieving "code follows the field, data is stored with the code." The core control logic flow (corresponding to the attached control flowchart, the steps are as follows): Step 1: Data Input Phase. The system simultaneously collects three types of data: ① Geographic entity coding database data (unique farmland code, location, risk level, historical status, QR code information); ② Inspection equipment status data (location / power / load); ③ Suspected violation results output by the AI ​​"code-based" recognition subsystem (associated codes); Step 2: Automatic Creation of "Code-Based" Patrol Tasks. Based on the distribution of farmland entities associated with codes and the requirements for patrol cycles, differentiated patrol tasks are automatically generated. Task information is forcibly associated with the unique code of the farmland, including core parameters such as code, farmland location, patrol range, shooting accuracy, and patrol frequency. It supports two modes: batch creation (batch association by regional code) and precise creation of single farmland plots (association by single code). Step 3: Automatic planning of patrol routes via QR code. Using the center point of the coded farmland entity as the core and the latitude and longitude boundaries as the boundaries, the algorithm takes obstacles (such as buildings and water bodies, obtained through a geographic information database) and the patrol equipment's range (calculated based on battery power and operating speed: range = remaining power × 1.2km / 10% power - reserved return range) as constraints. The objective function is to achieve the shortest route that completely covers the farmland entity range associated with the code. The algorithm steps are as follows: ① Initialize the starting point (current patrol position) and the ending point (center point of the farmland entity); ② Construct a grid map (grid size 10m × 10m), marking obstacle grids and the farmland entity range grids corresponding to the codes; ③ Calculate the cost function f(n) = g(n) + h(n) for each grid, where g(n) is the actual distance from the starting point to the current grid, and h(n) is the Manhattan distance from the current grid to the ending point; ④ Select the grid with the lowest cost as the next node, repeating until a route that completely covers the farmland entity is formed; ⑤ Smooth the route (using B-spline curve fitting) to reduce inflection points. Step 4: Automatic Task Execution Phase. A distributed load balancing algorithm (improved round-robin algorithm) is adopted. The core logic is as follows: ① Calculate the current load of each patrol unit (load rate = current number of tasks / maximum number of tasks, maximum number of tasks = number of patrol devices × 2); ② Assign automatically created "code-based" patrol tasks to the patrol unit with the lowest load rate. If multiple patrol units have the same load rate, assign the task to the patrol unit closest to the farmland entity corresponding to the code; ③ When the load rate of a single patrol device is ≥80%, trigger the task splitting logic to split the subsequent tasks of that device to patrol devices with a load rate <50%; ④ After receiving the task, the patrol device automatically executes it, flies along the planned route, and simultaneously starts the automatic shooting mode. All shooting actions are associated with the unique farmland code. Step 5: Automatic Data Capture and Status Feedback Stage. Standardized patrol instructions (including flight path coordinates, automatic shooting parameters, patrol frequency, and associated codes) are sent to the corresponding patrol units via the MQTT protocol. The patrol equipment automatically captures high-definition image data (1080P resolution, 30fps) according to the instructions, and transmits the shooting status and image thumbnails back in real time. All captured data is automatically tagged with a unique farmland code. The system receives the patrol equipment's operating status data (position, speed, and battery level) in real time. If there is a flight path deviation (e.g., the actual position deviates from the planned flight path by ≥5m), the system adjusts the operating attitude through a PID control algorithm (proportional coefficient Kp=0.8, integral coefficient Ki=0.2, derivative coefficient Kd=0.1) to ensure accurate coverage of the coded farmland entities and complete the shooting. The captured data is then transmitted to the storage subsystem for archiving after being associated with the codes.

[0025] Furthermore, by precisely linking the unique geographical entity spatial identity code of arable land with image data captured by drones during "on-the-spot" patrols, the system can intelligently identify violations of arable land conversion to non-agricultural or non-grain uses, generate suspected illegal clues with associated codes, and provide precise evidence for "on-the-spot" handling.

[0026] A deep learning model architecture of "lightweight backbone network + attention mechanism + multi-scale feature fusion" is adopted. The backbone network uses MobileNetV3-Large, replacing the fully connected layers of traditional CNNs with global average pooling layers to reduce the number of parameters (≤5M) and improve real-time inference speed. A CBAM attention module is added to the output layer of the backbone network to enhance the feature extraction capability of illegal areas (such as illegal construction, abandoned land, and fish farming ponds). FPN (Feature Pyramid Network) is used to achieve multi-scale feature fusion (fusion of 16×16, 32×32, and 64×64 scale features) to improve the recognition accuracy of small target violations (such as small buildings and scattered abandoned land). The specific process is as follows: Step 1: Image preprocessing. Spatial correction is performed on the associated coded image data captured by the drone; Step 2: Feature Extraction. Image features are extracted using the MobileNetV3-Large backbone network, outputting multi-scale feature maps; Step 3: Feature Fusion and Detection. Multi-scale features are fused using FPN and input into the detection head for bounding box regression and category prediction (categories include: normal farmland, illegal construction, hardened ground, abandoned land, fish farming ponds, and non-grain crop planting). Step 4: Post-processing and coding association. Redundant detection boxes are removed using NMS (Non-maximum suppression), with an NMS threshold of 0.5. The recognition results (including violation category, bounding box coordinates, and confidence level) are output and automatically associated with the unique farmland code corresponding to the input image. If the confidence level is ≥0.7, it is judged as a suspected violation, and a suspected violation clue with associated code is generated and pushed to the "Code-Based" problem handling subsystem.

[0027] Furthermore, using the unique geographical entity spatial identity code of cultivated land as a link, the status of cultivated land supervision is dynamically identified through four colors of QR codes: green, yellow, red, and blue. This enables full-process traceability from the discovery and identification of illegal and irregular activities on each piece of cultivated land to the handling and rectification, ensuring efficient and standardized handling.

[0028] QR code status definition: Green QR code indicates normal farmland, yellow QR code indicates suspected illegal farmland (AI recognition confidence level ≥ 0.7, manual verification not completed), red QR code indicates confirmed illegal farmland (manual verification confirmed, confidence level ≥ 0.9), and blue QR code indicates farmland that has completed rectification (illegal rectification eliminated, acceptance passed).

[0029] The "QR code" processing procedure and status are linked as follows: Step 1: Clue Reception and Alarm Triggering. The system receives suspected illegal clues with associated codes pushed by the AI ​​"Code-Based" recognition subsystem. It employs a single-trigger mechanism of "AI recognition confidence level + code association." When the confidence level is ≥0.7, a Level 1 alarm is automatically triggered. Simultaneously, the geographical entity coding engine updates the QR code status of the corresponding farmland from green to yellow. Alarm information includes: unique farmland code, yellow QR code, violation category, evidence image, latitude and longitude coordinates, and confidence level. Step 2: Alarm Push. Alarm information with associated codes is pushed to regulatory personnel via two methods: SMS gateway (using HTTP API to connect to China Mobile / China Unicom SMS platforms) and platform push (WebSocket real-time push). Step 3: Verification, Scheduling, and Status Update. The drone "code-based" inspection subsystem automatically generates precise verification tasks based on alarm-associated codes, scheduling the nearest inspection equipment to conduct secondary image verification of the target farmland entity. Supervisory personnel upload the verification results: ① Confirmed violation (confidence ≥ 0.9): Triggers a level-two alarm, updating the yellow QR code to red via the geographic entity coding engine; ② Removal of violation: Restores the yellow QR code to green via the geographic entity coding engine; All status changes are associated with and archived with codes. Step 4: Rectification Tracking and Status Linkage. For farmland with red QR codes, the system automatically generates rectification tasks with associated codes. The drone "code-based" inspection subsystem schedules inspection equipment weekly to track the rectification progress. After the illegal farmland entity completes rectification, the supervisors upload the acceptance materials. After the system approves the materials, the red QR code is updated to blue through the geographic entity coding engine. Step 5: Data Archiving and Full-Process Traceability. All alarm, verification, and rectification data are archived in association with the unique farmland code and the corresponding stage QR code status, forming a closed-loop data chain of "discovery (yellow) - identification (red) - rectification (blue) - archiving", which supports quick traceability of the entire supervision process of each piece of farmland through codes or QR codes.

[0030] The intelligent field inspection method of this invention achieves full automation of the entire process: "assigning codes to cultivated land entities - drone on-the-scan inspection - AI on-the-scan identification - on-the-scan handling of problems - full-process traceability". Each step revolves around the unique code of cultivated land. The specific steps are as follows: (1) Cultivated land entity coding. Based on the 2019 land parcel data, each cultivated land entity was coded to construct cultivated land entity data. Change survey attribute information was used as the basic entity attribute, and special survey results were used as the entity special attribute. Using the 2020 land parcel data as the update data source, and based on the change relationship between the 2019 and 2020 land parcels, each cultivated land and forest parcel in 2020 was updated and coded. That is, existing entity codes were inherited or new entities were recoded, thus completing the update of cultivated land entities. The same method was used to update cultivated land entities based on land parcel data from other years. At the same time, a visual QR code was generated based on the code. The initial state was set to green. The QR code data field contained the code and the core cultivated land information, completing the initial binding of "code-cultivated land-QR code". All code and QR code information were stored in the distributed database.

[0031] (2) Deployment of the "Code-Based" Patrol Network. Based on the distribution density of cultivated land, topographic features and signal coverage, the coverage area of ​​the patrol network is planned. A distributed collaborative scheduling architecture is adopted to overcome the shortcomings of ordinary patrols, such as monotony and unstable network. Communication adaptation and network access of patrol equipment are completed. Initialization operations such as parameter calibration of patrol equipment, task instruction adaptation and data transmission protocol matching are completed through software configuration to ensure that the equipment can accurately associate codes and execute tasks.

[0032] (3) Automatic creation and route planning of "QR code" patrol tasks. Based on farmland entity coding data and patrol equipment status data, the unique code of farmland is accurately bound to the patrol range; differentiated "QR code" patrol tasks (including code, shooting parameters, patrol frequency, etc.) are automatically created; routes covering the farmland entity range corresponding to the code are planned (obstacles are constrained and the range is extended), and daily automated patrol tasks containing route coordinates, automatic shooting parameters, and patrol frequency (1 time / week for green QR code plots, 1 time / day for yellow / red QR code plots) are generated; the tasks are distributed to each patrol equipment through a load balancing algorithm.

[0033] (4) UAV "Code-Based" Field Patrol and Automatic Shooting. The UAV "Code-Based" Patrol Subsystem issues automatic execution commands with associated codes. The patrol equipment performs field patrol operations according to the planned route and automatically starts shooting mode. During operation, the system receives the status data (location, speed, battery level) and shooting status data of the patrol equipment in real time. Combined with the farmland entity terrain data in the geographic information database, the system dynamically adjusts the patrol and shooting strategies. The patrol equipment fully covers the farmland entity range associated with the code according to the planned route, transmits live images back in real time, and automatically stores high-definition image data associated with the code. The operation trajectory and shooting location data are all transmitted to the edge node and the central node for storage with associated codes.

[0034] (5) AI "Code-Based" Identification and Preliminary Judgment of Suspected Violations. The data transmission subsystem pushes the preprocessed associated coded image data to the AI ​​"Code-Based" identification subsystem; it adopts a lightweight AI model and real-time inference technology to overcome the defects of low accuracy, poor real-time performance, and lack of precise association in conventional identification; it outputs identification results, including non-agricultural and non-grain farmland patches, and automatically associates them with the unique farmland code corresponding to the input image; when the confidence level is ≥0.7, the system automatically marks the plot as a suspected violation, generates suspected violation clues with associated codes, and pushes them to the problem "Code-Based" handling subsystem.

[0035] (6) Problem handling via QR code and linkage with QR code status. After receiving suspected illegal clues, the problem handling subsystem triggers a level one alarm and updates the QR code to yellow; pushes alarm information to regulatory personnel and dispatches drones for secondary verification; updates the QR code status (yellow to red / yellow to green) according to the verification results; tracks and rectifys farmland with red QR codes, and updates it to blue after rectification is completed; all process data are associated with codes and QR code status archives to form a closed-loop data chain. The multi-color QR code dynamic management technology overcomes the shortcomings of fragmented handling and difficulty in traceability in conventional handling.

[0036] (7) Enhanced inspection and full-process traceability of key cultivated land. The UAV "code-based" inspection subsystem automatically increases the inspection frequency to 2 times / day and the shooting accuracy to centimeter level based on the cultivated land risk level associated with the code (cultivated land in urban-rural fringe areas is high-risk) and the QR code status (yellow, red), and re-plans routes to cover more fine details; if clues from other sources are received, regulatory personnel can quickly locate the plot through the unique cultivated land code or QR code, and manually create a precise verification task to schedule inspection equipment for verification; it supports querying all inspection, identification, and disposal data of the corresponding cultivated land through the code or QR code to achieve full-process traceability. Dynamic scheduling and closed-loop management technology are adopted to overcome the shortcomings of conventional inspections that do not highlight key points and have no follow-up on rectification. The above describes a method for farmland inspection based on geographic entity spatial identity coding, as provided in this application. Based on the same inventive concept, this application also provides a farmland inspection system based on geographic entity spatial identity coding. Figure 2 A schematic diagram of a farmland inspection system based on geographic entity spatial identity coding provided in this application embodiment is shown below. Figure 2 As shown, the device mainly includes: at least one processor 201; and a memory 202 communicatively connected to the at least one processor; wherein the memory 202 stores instructions that can be executed by the at least one processor 201, and the instructions are executed by the at least one processor 201 to enable the at least one processor 201 to complete the aforementioned method for farmland inspection based on geographic entity spatial identity encoding.

[0037] In addition, this application embodiment also provides a non-volatile computer storage medium for farmland inspection based on geographic entity spatial identity coding, which stores computer-executable instructions. The computer-executable instructions are executed by a processor to implement the aforementioned farmland inspection method based on geographic entity spatial identity coding.

[0038] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0039] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0040] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0041] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0042] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the device embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0043] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0044] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for farmland inspection based on geographic entity spatial identity coding, characterized in that, The method includes the following steps: Step S1: Assign a geographic entity spatial identity code to the target farmland, generate a visual QR code containing core farmland information and status identifier based on the geographic entity spatial identity code, and store the geographic entity spatial identity code and QR code information in a distributed database. Step S2: Acquire drone patrol images, perform preprocessing and feature extraction, detect illegal activities such as non-agricultural and non-grain use of arable land through intelligent recognition models, and generate suspected illegal clues with unique associated codes; Step S3: Identify suspected illegal clues and dynamically update the QR code status based on the identification results.

2. The method for farmland inspection based on geographic entity spatial identity coding according to claim 1, characterized in that, In step S1, the proprietary identifier field of the geographic entity spatial identity code includes the root identifier code and the geographic entity special code, the standard field includes the location code, the classification code that distinguishes the type of cultivated land and the sequence code, and the extended field is a hybrid code used to expand the functions or attributes of cultivated land.

3. The method for farmland inspection based on geographic entity spatial identity coding according to claim 1, characterized in that, In step S2, the drone patrol automatically creates differentiated patrol tasks with unique codes based on the farmland data associated with the codes and the status data of the patrol equipment. It plans the optimal flight route covering the farmland area corresponding to the code, allocates tasks through a distributed scheduling algorithm, and the patrol equipment automatically flies along the flight route and takes high-definition images. All patrol data is archived with the unique farmland code. The differentiated patrol tasks support two creation modes: batch creation by regional code or precise creation by the unique code of a single farmland. The task parameters include farmland location, patrol range, shooting accuracy, and patrol frequency matching the QR code status.

4. The method for farmland inspection based on geographic entity spatial identity coding according to claim 3, characterized in that, The optimal route planning takes the center point of the cultivated land entity as the core, and uses obstacles and the endurance of the inspection equipment as constraints. Through grid map construction, cost function calculation and curve fitting, the route is made to achieve the shortest route and complete coverage of the cultivated land area.

5. The method for farmland inspection based on geographic entity spatial identity coding according to claim 1, characterized in that, In step S2, the intelligent recognition model adopts an architecture of lightweight backbone network + attention mechanism + multi-scale feature fusion. By removing redundant detection boxes, the recognition results are optimized to improve the feature extraction capability of illegal areas and the recognition accuracy of small target violations.

6. The method for farmland inspection based on geographic entity spatial identity coding according to claim 1, characterized in that, In step 3, the violations include illegal construction, hardening of ground, abandonment of land, digging ponds for fish farming, and planting of non-grain crops. Suspected violations are identified based on the confidence level of identification and clues are pushed out.

7. The method for farmland inspection based on geographic entity spatial identity coding according to claim 1, characterized in that, In step S3, the QR code status update rules are as follows: when AI identifies and determines that there is a suspected violation, it is updated to yellow; when manual verification confirms the violation, it is updated to red; after rectification and acceptance, it is updated to blue; and when the violation is ruled out, it is restored to green.

8. The method for farmland inspection based on geographic entity spatial identity coding according to claim 1, characterized in that, In step S4, the suspected illegal clues are sent to regulatory personnel via SMS push and platform real-time push, with alarm information including farmland code, QR code, violation category and location information.

9. A method for farmland inspection based on geographic entity spatial identity coding according to claim 1, characterized in that, The method also includes: automatically increasing the frequency of patrols, optimizing the shooting accuracy, and replanning detailed routes for high-risk farmland and farmland with yellow / red QR codes associated with codes, thereby strengthening key supervision.

10. A farmland inspection system based on geographic entity spatial identity coding, characterized in that, The system includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform a farmland inspection method based on geographic entity spatial identity coding as described in any one of claims 1-9.

Images