Facility agriculture land use system

Abstract

The invention discloses a facility agriculture land use system, and relates to the technical field of facility agriculture land use, and the system comprises a vector data processing unit which generates an initial pattern spot data set with a coincidence degree identifier; the field intelligent proof unit is used for generating a proof data packet with a unique file number; the third-level joint review unit is used for performing third-level review; the structured achievement management unit is used for establishing an associated index and a structured database; and the cross-module data pushing unit is used for outputting an audit completion voucher and realizing a data closed loop. According to the method, a systematic architecture of vector data processing, field intelligent proof, three-level joint review, structured achievement management and cross-module data pushing is constructed, pattern spot extraction, proof verification and trend analysis intelligentization are achieved through an AI algorithm, a typed template, layered review and an automatic pushing mechanism are used, and the system and the method are suitable for large-scale popularization and application. The problems that traditional facility agriculture land management is scattered in data, split in process and low in efficiency are solved, and a traceable integrated management closed loop is formed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of facility agricultural land, in particular to a facility agricultural land system. BACKGROUND

[0002] With the promotion of the rural revitalization strategy, facility agriculture, as an important form of efficient agriculture, continues to expand in scale. However, the current facility agricultural land management still faces many pain points in multiple links. In the data processing link, the traditional method relies on manual extraction of plot, and manual comparison of approved vector data and satellite images is required, which not only is low in efficiency, but also is prone to human error, leading to disconnection between plot range and approved range, affecting the accuracy of subsequent processes. In the field evidence collection stage, there is a lack of unified standard templates, and county-level personnel often have to repeat the rectification process due to incomplete shooting content or disconnection between photos and plot attributes, with an average of 2-3 times of rectification.

[0003] In the review process, county, city, and province levels are usually independently audited, information transmission relies on offline files, land classification standards are not unified, and provincial random checks have no fixed methods, which may lead to different identification results for the same type of land. The review period generally exceeds 15 working days. In the result management and data pushing link, data is scattered and stored in different systems, lacks structured association, and requires cross-platform queries for complete approval, evidence collection, and review records. Moreover, there is no automatic interface for data pushing, which requires manual export and import, not only consuming time but also easily losing data, making it difficult to form a management closed loop. These problems lead to low efficiency and difficulty in ensuring compliance in facility agricultural land review, and a systematic solution is urgently needed. SUMMARY

[0004] The purpose of the present application is to solve the above-mentioned problems by providing a facility agricultural land system.

[0005] The purpose of the present application can be achieved by the following technical solutions: The facility agricultural land system comprises a vector data processing unit, an intelligent field evidence collection unit, a three-level joint review unit, a structured result management unit, and a cross-module data pushing unit. The vector data processing unit is used to obtain approved vector data, integrate high-resolution satellite image data, extract changed plot through spatial superposition, calculate the coincidence degree of plot and approved range, generate initial plot data set with coincidence degree identifier, and transmit to the intelligent field evidence collection unit. The intelligent field evidence collection unit is used to receive the initial plot data set, match the evidence standard template according to the type of facility agricultural land, generate individualized evidence task list, guide personnel to collect photos inside and outside the plot, establish the association between photos and plot attributes, generate evidence data package with unique file number, and transmit to the three-level joint review unit. A three-level joint examination unit for receiving the evidence data package, conducting a three-level examination, and transmitting to the structured achievement management unit; A structured achievement management unit for receiving the examination conclusion file, integrating the approval data, the evidence data package, and the examination conclusion, establishing an associated index and a structured database, and transmitting to the cross-module data pushing unit; A cross-module data pushing unit for receiving the standardized achievement, screening qualified data and pushing to the cloud platform, calculating the approval range and the passing plot area ratio, and generating a visual chart, outputting an audit completion voucher, and realizing a data closed loop.

[0006] Preferably, the vector data processing unit comprises: A data acquisition subunit for extracting approval vector data from a provincial facility agricultural land supervision system through an API interface and synchronously obtaining high-resolution image data of a corresponding region from a natural resource satellite remote sensing platform; A plot extraction subunit for performing edge extraction on the image data based on a Canny edge detection algorithm, combining the spatial range of the approval vector data, identifying and extracting a change plot through an R-CNN target detection model, and generating an initial change plot vector; A coincidence degree calculation subunit for calculating the coincidence degree of the initial change plot and the approval vector range using a Jaccard index, marking the reference type equal to the approval range if the coincidence degree is ≥ 90%, marking the reference type equal to a provincial extraction segmentation identifier if the coincidence degree is < 90%, segmenting the plot part exceeding the approval range, supplementing plot center point coordinates and image shooting time attributes, and generating an initial plot data set.

[0007] Preferably, the vector data processing unit further comprises: A data standardization subunit for performing coordinate unification on the initial plot data set according to a geodetic coordinate system, verifying plot attribute integrity, generating a correction prompt for data with format errors or missing attributes, and re-including the corrected data in the initial plot data set; An invalid data filtering subunit for excluding construction land and non-facility agricultural land vectors in the initial plot data set through spatial overlay analysis, retaining only plot data that meets the land use type pre-judgment, and improving subsequent evidence and examination efficiency.

[0008] Preferably, the intelligent field evidence unit comprises: The evidence standard matching subunit is configured to receive an initial polygon data set, automatically match an evidence standard template according to a land use type in a polygon attribute, and take photos of a greenhouse and a big shed outside and a crop growth state inside for a crop planting type, take photos of a shed and a breeding pond outside and inside facilities for livestock and poultry and aquatic breeding, and take photos of a steel structure and a temperature control device for industrial planting. The task allocation subunit is configured to divide an evidence area based on a polygon spatial distribution, combine a county-level evidence person's qualification and geographical location, perform optimal matching of an evidence task and the person through a Hungarian algorithm, and generate an evidence task list containing a plot area range, a task time limit, and standard requirements. The field collection subunit is configured to enable an evidence person to receive a task list through a mobile terminal APP, scan a polygon attribute two-dimensional code to associate data, take photos according to a template requirement, and automatically add a shooting time, a GPS position watermark by the APP. The association and packaging subunit is configured to establish an association mapping of a photo and a polygon number, a land use type, and an evidence person, generate an evidence list in a JSON format, package the evidence list and a photo file into an evidence data package in a ZIP format, and attach a unique file number.

[0009] Preferably, the field intelligent evidence unit further includes: The compliance verification subunit is configured to automatically verify photos in the evidence data package in terms of a number of photos (≥2 for each polygon), a definition (determined by an edge gradient value, and a gradient value ≥50 is qualified), and content association (identified by a ResNet18 model whether a photo contains corresponding facility features, and an identification confidence ≥0.8 is qualified), and generate specific rectification prompts (such as “missing internal photos of a breeding pond”) for verification failure. The evidence tracing subunit is configured to add an operation log to the evidence data package that passes the verification, record a first collection time, a rectification time, and a verification result, form a traceable evidence file, and synchronously push the evidence file to a three-level joint review unit.

[0010] Preferably, the three-level joint review unit includes: The county-level preliminary review subunit is configured to receive the evidence data package, extract a polygon attribute and a photo, determine by comparison with a positive and negative list of facility agricultural land, determine a production auxiliary facility in the positive list as compliant, determine a leisure tourism and deep processing site in the negative list as non-compliant, and generate a county-level preliminary review report containing a compliance basis and a preliminary review result. The city-level review subunit is configured to receive the county-level preliminary review report, focus on verifying a polygon with a doubtful land class, supplement a review opinion by calling a photo detail and comparing a historical change in images in the past three years, mark a polygon that still has doubts for a provincial consultation, and generate a city-level review report to be pushed to a province. The provincial final review subunit is used for the provincial cultivated land red line regulation technical center to review the non-suspected plot, to change the plot information of the land class to facility agricultural land into the comprehensive monitoring and supervision module, and the provincial natural resource investigation center to randomly check 10%-20% of the proportion, to use the stratified sampling method to ensure that all types of land use types are covered; The consultation resolution subunit is used for the land class identification difference found by the random check to organize a consultation meeting of two departments, to record the difference point, the consultation basis and the final conclusion, and to generate a provincial final review report; The review conclusion generation subunit is used for integrating the three-level review opinions: generating a review passing conclusion and a land class change confirmation sheet for the plot passing the review, generating a rectification notice for the plot needing rectification, and forming a review conclusion file.

[0011] In runtime, the three-level joint review unit further includes: The standard updating subunit is used for receiving the national land change investigation technical procedure updating content every quarter, automatically updating the built-in positive and negative list, the change range refinement table, and generating an explanation document for the difference between the new and old standards for the review personnel to refer to; The review efficiency statistical subunit is used for statistically analyzing the review time, the passing rate and the rectification rate of each link, generating a visual efficiency analysis chart by using Matplotlib, serving as the basis for process optimization, and being simultaneously incorporated into the review conclusion file.

[0012] Preferably, the structured achievement management unit includes: The data integration subunit is used for receiving the review conclusion file, automatically associating the corresponding evidence data package and the initial plot data set, establishing an associated index according to the file number, the plot number and the review conclusion, using the PostgreSQL+PostGIS database to build a structured database, and supporting spatial query and attribute query; The task overview subunit is used for real-time reading of the structured database data, displaying the task progress, the pending tasks and the review passing rate statistics of each link, and generating a Web-side visual task board; The upper map storage subunit is used for the county-level user to call the review passing plot data in the structured database, to automatically match the national land space planning database coordinate system, to complete the upper map storage operation, and to generate a SHP format vector achievement.

[0013] Preferably, the structured achievement management unit further includes: The data preprocessing subunit is used for receiving the user batch query request, automatically processing the database data, including data deduplication, format conversion and attribute completion, and generating a preprocessing data package; The historical archive and decision support subunit is used for constructing a historical archive library according to years and land type types, supporting retrieval of historical data, analyzing land use change trends through an ARIMA time series model, and generating a decision analysis report of annual growth rate and land type structure proportion for reference of a management department.

[0014] Preferably, the cross-module data pushing unit comprises: The data screening subunit is used for screening the parcel data that passes the examination and is included in daily changes from the structured database, extracting parcel numbers, examination and approval ranges and land type identification result fields; The cross-module pushing subunit is used for pushing the screened data to the investigation and examination passed parcel list of the facility agricultural land project 2024 module of the cloud platform through a RESTful API interface, and attaching a provincial final examination report link for viewing by the farmland protection department; The proportion analysis subunit is used for calling the examination and approval range area and the passed parcel area in the initial parcel data set, obtaining the proportion, generating a column chart of proportion ranking of regions through ECharts, and displaying on the cloud platform visualization interface; The examination closed loop subunit is used for receiving the parcel data whose 100% examination and approval range is changed to facility agricultural land screened by the county level, pushing the parcel data to the provincial facility agricultural land supervision system, generating an examination completion voucher containing examination time, examination personnel and parcel list, and simultaneously returning the voucher to the structured achievement management unit for archiving, and completing the full-process data closed loop.

[0015] Compared with the prior art, the present application has the following beneficial effects: 1. The changed parcel is automatically extracted through Canny edge detection + R-CNN model, the coincidence degree is calculated by combining Jaccard index, the error of parcel extraction is reduced to within 5%, the efficiency is improved by 3 times compared with manual processing, the data standardization and invalid data filtering subunit ensures the uniformity of coordinates and only keeps the effective parcel, and the data preparation time in the subsequent evidence and examination link is reduced.

[0016] 2. The evidence template is matched according to the land use type, the shooting requirements are clear, the mobile terminal APP is used for automatically adding watermark and associating attributes, the compliance rate of evidence materials is improved, the compliance checking subunit intelligently verifies through edge gradient value and ResNet18 model, identifies problems in advance and generates rectification prompts, the number of evidence rectification is reduced, the county level manpower cost is saved, the positive and negative list and automatic updating mechanism are built in the three-level joint examination unit, the land type identification standard is ensured to be consistent, the provincial level adopts stratified sampling and spot check, efficiency and fairness are considered, the difference consultation mechanism quickly solves the identification controversy, the overall examination period is shortened, and the examination pass rate stability is improved.

[0017] 3. Structured database associates full link data through file number and plot number, shortens the time of querying the complete record of a certain plot, pushes data automatically through RESTful API by cross-module pushing unit, intuitively displays the audit result through proportion analysis and visual chart, automatically archives the audit completion voucher, forms a full-process closed loop, and provides data support for long-term supervision of facility agricultural land. BRIEF DESCRIPTION OF DRAWINGS

[0018] For the convenience of those skilled in the art to understand, the present application will be further described below in conjunction with the drawings.

[0019] Figure 1 The system framework diagram of the present application; Figure 2 The internal system framework diagram of the vector data processing unit in the present application; Figure 3 The internal system framework diagram of the intelligent field evidence unit in the present application; Figure 4 The internal system framework diagram of the three-level joint review unit in the present application; Figure 5 The internal system framework diagram of the structured achievement management unit in the present application; Figure 6 The internal system framework diagram of the cross-module data pushing unit in the present application; The drawings show: 1, vector data processing unit; 2, intelligent field evidence unit; 3, three-level joint review unit; 4, structured achievement management unit; 5, cross-module data pushing unit. DETAILED DESCRIPTION

[0020] In order for those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the present application.

[0021] In this paper, the phrase "embodiment" means that the specific features, structures or characteristics described in conjunction with the embodiment can be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily mean the same embodiment, nor is it an independent or alternative embodiment to other embodiments. Those skilled in the art explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0022] Please refer to Figure 1As shown, the facility agricultural land system includes: a vector data processing unit 1, an intelligent field evidence collection unit 2, a three-level joint review unit 3, a structured achievement management unit 4, and a cross-module data pushing unit 5. The vector data processing unit 1 is used to obtain approval vector data, integrate high-resolution satellite image data, extract change polygons through spatial superposition, calculate the coincidence degree of the polygons and the approval range, generate an initial polygon data set with a coincidence degree identifier, and transmit it to the intelligent field evidence collection unit 2. The intelligent field evidence collection unit 2 is used to receive the initial polygon data set, match the evidence standard template according to the type of facility agricultural land, generate a personalized evidence task list, guide personnel to collect photos inside and outside the polygons, establish an association between the photos and polygon attributes, generate an evidence data package with a unique file number, and transmit it to the three-level joint review unit 3. The three-level joint review unit 3 is used to receive the evidence data package, conduct three-level review, and transmit it to the structured achievement management unit 4. The structured achievement management unit 4 is used to receive the review conclusion file, integrate the approval data, evidence data package, and review conclusion, establish an association index and structured database, and transmit it to the cross-module data pushing unit 5. The cross-module data pushing unit 5 is used to receive standardized achievements, filter qualified data and push it to the cloud platform, calculate the area proportion of the approval range and the passed polygons, generate visual charts, output the audit completion certificate, and realize data closed loop.

[0023] Please refer to Figure 2 As shown, the vector data processing unit 1 includes: The data acquisition sub-unit is used to extract approval vector data from the provincial facility agricultural land supervision system through an API interface, and simultaneously obtain high-resolution image data of the corresponding region from the natural resource satellite remote sensing platform; The polygon extraction sub-unit is used to perform edge extraction on the image data based on the Canny edge detection algorithm, combine the spatial range of the approval vector data, identify and extract change polygons through an R-CNN target detection model, and generate initial change polygon vectors; The coincidence degree calculation sub-unit is used to calculate the coincidence degree of the initial change polygons and the approval vector range using the Jaccard index. If the coincidence degree is ≥ 90%, the reference type is marked as equal to the approval range. If the coincidence degree is < 90%, the part of the polygon that exceeds the approval range is segmented, the reference type is marked as equal to the provincial extraction plus segmentation identifier, the polygon center point coordinates and image shooting time attributes are supplemented, and the initial polygon data set is generated; For the first time, Canny edge detection is combined with R-CNN model for facility agricultural land use change polygon extraction, replacing traditional manual delineation, polygon extraction accuracy is improved, Jaccard index is introduced to quantify the overlap between polygon and approval range, and reference type is marked according to threshold difference, so as to avoid subsequent review errors caused by range deviation.

[0024] The vector data processing unit 1 further comprises: The data standardization subunit is used for unifying the coordinates of the initial polygon data set according to the geodetic coordinate system, checking the attribute integrity of the polygon, generating a correction prompt for data with format errors or missing attributes, and re-including the corrected data in the initial polygon data set; The invalid data filtering subunit is used for excluding non-facility agricultural land vectors in the initial polygon data set through spatial overlay analysis, and only retaining polygon data that meets the pre-judgment of land use type, so as to improve the efficiency of subsequent evidence and review; The data standardization subunit realizes the automation of coordinate unification and attribute checking, solves the problem of traditional data format confusion and coordinate inconsistency; the invalid data filtering link is added to actively exclude non-facility agricultural land vectors, reduce the invalid workload of subsequent evidence and review, and improve the efficiency.

[0025] Please refer to Figure 3 As shown in the figure, the intelligent evidence collection unit 2 comprises: The evidence standard matching subunit is used for receiving the initial polygon data set, automatically matching the evidence standard template according to the land use type in the polygon attribute, and including the panoramic photo outside the greenhouse and the photo of the internal crop growth state for crop planting, including the photo of the internal facilities outside the shed and the breeding pond for livestock and poultry and aquatic breeding, and including the photo of the steel structure and the temperature control equipment details for factory planting; The task allocation subunit is used for dividing the evidence area based on the spatial distribution of the polygon, combining the qualifications and geographical location of the county-level evidence collectors, and performing optimal matching of the evidence task and the personnel through the Hungarian algorithm to generate an evidence task list containing the area range, task time limit and standard requirement; The field collection subunit is used for the evidence collector to receive the task list through the mobile terminal APP, scan the polygon attribute two-dimensional code to associate the data, and shoot the photo according to the template requirement, and the APP automatically adds the shooting time, GPS position watermark; The association packaging subunit is used for establishing the association mapping of the photo and the polygon number, land use type, and evidence collector, generating a JSON format evidence list, and packaging the photo file into a ZIP format evidence data package with a unique file number; The evidence template is customized according to the land use type, the problems of traditional evidence standard are solved, the Hungarian algorithm is used to realize the optimal matching of tasks and personnel, the unreasonable personnel allocation is avoided to waste the efficiency, the two-dimensional code association and automatic watermarking are used to ensure that the evidence data and the plot attribute correspond one by one, and the data confusion is avoided.

[0026] The intelligent evidence collection unit 2 further includes: The compliance verification sub-unit is used for automatically verifying the photos in the evidence data package: the number of photos per plot is greater than or equal to 2, the clarity is determined by the edge gradient value, the gradient value is greater than or equal to 50, the content association is identified by the ResNet18 model whether the photo contains the corresponding facility features, and the identification confidence is greater than or equal to 0.8. If the verification fails, a rectification prompt is generated. The evidence tracing sub-unit is used for adding operation logs to the evidence data package that passes the verification, recording the first collection time, the rectification time and the verification result, forming a traceable evidence file, and synchronously pushing to the three-level joint review unit 3. The edge gradient value and the ResNet18 model are introduced to realize intelligent verification of evidence photos, replace traditional manual review, improve verification efficiency by 3 times, accurately locate problems and generate rectification prompts, add evidence tracing logs, realize traceable evidence in the whole process, and solve the problem that the evidence responsibility cannot be traced in subsequent review.

[0027] Please refer to Figure 4 The three-level joint review unit 3 includes: The county-level preliminary review sub-unit is used for receiving the evidence data package, extracting the plot attribute and the photo, determining by comparing with the positive and negative list of agricultural land, determining the production auxiliary facilities in the positive list as compliant, determining the leisure tourism and deep processing places in the negative list as non-compliant, and generating a county-level preliminary review report containing the compliance basis and the preliminary review result. The city-level review sub-unit is used for receiving the county-level preliminary review report, focusing on verifying the plot with doubtful land class, supplementing the review opinion by calling the photo details and comparing the image historical changes in the past three years, marking the plot that still has doubts for provincial consultation, and generating a city-level review report to push to the province; The provincial final review sub-unit is used for the provincial cultivated land red line regulation technical center to review the plots without doubts, change the land class of the plot with facility agricultural land to the comprehensive monitoring and supervision module, and the provincial natural resource investigation center randomly checks 10%-20% of the plots, and uses stratified sampling method to ensure that all types of land use are covered. The consultation resolution sub-unit is used for the land class identification difference found by the random check, organizes a consultation meeting of two departments, records the difference point, consultation basis and final conclusion, and generates a provincial final review report. The review conclusion generation subunit is configured to integrate the three-level review opinions, generate a review pass conclusion and a land class change confirmation sheet for the image patches that pass the review, generate a rectification notice for the image patches that need to be rectified, and form a review conclusion file; A three-level review system including county-level preliminary review, municipal-level review, and provincial-level final review and consultation is constructed, the review focuses of different levels are defined, the problems of unclear responsibility and inconsistent standards in traditional review are solved, the provincial level adopts stratified sampling for review to ensure that all types of land use are covered and sampling bias is avoided, a consultation mechanism is newly added to quickly solve land class identification differences and reduce review stagnation time.

[0028] The three-level joint review unit 3 further includes: The standard updating subunit is configured to receive the updated contents of the land change survey technical regulations every quarter, automatically update the built-in positive and negative lists and change range refinement table in the system, and generate an explanation document for the differences between the new and old standards for review personnel to refer to; The review efficiency statistical subunit is configured to statistically analyze the time length, pass rate, and rectification rate of each review link, generate visual efficiency analysis charts by using Matplotlib, and use the charts as the basis for process optimization and include the charts in the review conclusion file; The three-level joint review unit 3 further includes: The standard updating subunit is configured to receive the updated contents of the land change survey technical regulations every quarter, automatically update the built-in positive and negative lists and change range refinement table in the system, and generate an explanation document for the differences between the new and old standards for review personnel to refer to; The review efficiency statistical subunit is configured to statistically analyze the time length, pass rate, and rectification rate of each review link, generate visual efficiency analysis charts by using Matplotlib, and use the charts as the basis for process optimization and include the charts in the review conclusion file;

[0029] Please refer to Figure 5 As shown in the figure, the structured achievement management unit 4 includes: The data integration subunit is configured to receive the review conclusion file, automatically associate the corresponding evidence data package and initial image patch dataset, establish an associated index according to the file number, image patch number, and review conclusion, build a structured database by using PostgreSQL+PostGIS database, and support spatial query and attribute query; The task overview subunit is configured to read the data in the structured database in real time, display the task progress, pending tasks, and review pass rate statistics of each link, and generate a visual task board on the Web side; The figure importing subunit is configured to call the review pass image patch data in the structured database by the county-level user, automatically match the coordinate system of the land space planning database, complete the figure importing operation, and generate SHP format vector achievements; The structured database is built by using PostgreSQL+PostGIS database to realize the associated storage and spatial query of the approval, evidence, and review data, solve the problems of scattered data storage and difficult query in the traditional method, display the progress in real time through the visual task board, facilitate the work dynamics of personnel at different levels, and avoid coordinate deviation caused by manual conversion.

[0030] The structured achievement management unit 4 further comprises: A data preprocessing subunit is configured to receive a user batch query request, automatically process database data, including data deduplication, format conversion and attribute completion, and generate a preprocessed data package; A historical archive and decision support subunit is configured to construct a historical archive library according to years and land type, support retrieval of historical data, analyze land use change trend through an ARIMA time series model, and generate a decision analysis report of annual growth rate and land type structure proportion for reference by a management department; The new data preprocessing function automatically completes data deduplication, format conversion and attribute completion, improves data usability, introduces an ARIMA time series model to analyze land use change trend, provides decision support for the management department, and breaks through the limitation of traditional achievement management that only stores data and has no analysis capability.

[0031] Please refer to Figure 6 As shown in FIG. 5, the cross-module data pushing unit 5 comprises: A data screening subunit is configured to screen parcel data that has passed examination and is included in daily changes from the structured database, extract parcel number, examination and approval range and land type identification result fields; A cross-module pushing subunit is configured to push the screened data to the survey and examination passed parcel list of the facility agricultural land project 2024 module of the cloud platform through a RESTful API interface, and attach a provincial final examination report link for viewing by the farmland protection department; A proportion analysis subunit is configured to call the examination and approval range area and the passed parcel area in the initial parcel data set, obtain the proportion, generate a column chart of proportion ranking of each region through ECharts, and display on the cloud platform visualization interface; An examination closed loop subunit is configured to receive parcel data whose 100% examination and approval range has been changed to facility agricultural land screened by the county level, push the data to the provincial facility agricultural land supervision system, generate an examination completion voucher containing examination time, examination personnel and parcel list, and simultaneously return the voucher to the structured achievement management unit 4 for archiving, to complete the whole-process data closed loop; Cross-module data automatic pushing is realized through a RESTful API interface, replacing traditional manual export and import, avoiding data loss and delay; proportion analysis and visual display are newly added to intuitively present the examination passed situation of each region; an examination completion voucher is generated and returned for archiving to form a whole-process data closed loop, ensuring traceability and checkability of each link; Cross-module data automatic pushing is realized through a RESTful API interface, replacing traditional manual export and import, avoiding data loss and delay, proportion analysis and visual display are newly added to intuitively present the examination passed situation of each region; an examination completion voucher is generated and returned for archiving to form a whole-process data closed loop, ensuring traceability and checkability of each link.

[0032] In summary, the advantages of the present application are: Break through the pain point of "data isolation and link fragmentation" in traditional facility agriculture land management, build a "vector data processing-outside intelligent evidence-3-level joint review-structured achievement management-cross-module data pushing" full-link system architecture, realize the seamless connection from plot extraction to audit closed loop without human intervention, solve the problem of multi-department and multi-link information fault, and form a "data-process-achievement" integrated management system; For the first time, Canny edge detection algorithm and R-CNN target detection model are combined to replace traditional manual sketching, automatically extract facility agriculture land change plot, improve extraction accuracy, and improve processing efficiency compared with manual work. The Jaccard index is introduced to calculate the overlap between plot and approval range, and the benchmark type is marked according to the threshold difference of "≥90% / <90%", which avoids the subsequent audit error caused by range deviation. At the same time, through the data standardization subunit, coordinate unification and attribute verification automation are realized, solving the problem of traditional data format confusion; Add invalid data filtering subunit, eliminate non-facility agriculture land vector of construction land and unused land through spatial overlay analysis, reduce the invalid workload of subsequent evidence and review, and improve efficiency; According to the type of facility agriculture land, automatically match the exclusive evidence standard template, clarify the shooting content, solve the problem of traditional evidence standard ambiguity and incomplete content, and improve the compliance rate of evidence materials; Hungarian algorithm is used to realize the optimal matching of evidence task and personnel, and the mobile APP is used to scan the two-dimensional code to associate the plot attribute, automatically add the shooting time, GPS position watermark, ensure that "photo-plot-personnel" one-to-one correspondence, and prevent data confusion; Introduce edge gradient value and ResNet18 model to automatically check the number, clarity and content relevance of photos, accurately generate rectification prompts, add operation log to record the first collection, rectification time and result, realize the traceability of the whole process of evidence, and reduce the rectification times; Build a county-level preliminary review, city-level review and provincial-level final review system, clarify the key points of each level, use stratified sampling method to ensure the coverage of various types of land at the provincial level, avoid sampling bias, and shorten the review period.

[0033] The preferred embodiments of the application disclosed above are only to facilitate the elucidation of the application. The preferred embodiments do not describe all the details of the application and limit the application to the specific embodiments. Obviously, many modifications and variations can be made in light of the teachings above. The description is chosen and described in order to provide the best illustration of the application and its practical application to those skilled in the art and to enable those skilled in the art to best utilize the application. The application is limited only by the claims and their full scope and equivalents.

Claims

1. A facility agriculture land use system, characterized in that, include: Vector data processing unit (1), field intelligent evidence collection unit (2), three-level joint review unit (3), structured results management unit (4), cross-module data push unit (5). Vector data processing unit (1), which is used to acquire approval vector data, generate an initial map patch dataset with overlap degree marking, and transmit it to field intelligent evidence collection unit (2). The field intelligent evidence collection unit (2) is used to establish the association between photos and map patch attributes, generate evidence data packets with unique file numbers, and transmit them to the three-level joint review unit (3). The three-level joint review unit (3) is used to receive evidence data packets, conduct three-level review, and transmit them to the structured results management unit (4). The structured results management unit (4) is used to receive review conclusion documents, establish an associated index and structured database, and transmit them to the cross-module data push unit (5). Cross-module data push unit (5) is used to receive standardized results and output audit completion certificate.

2. The facility agriculture land use system according to claim 1, characterized in that, The vector data processing unit (1) includes: The data acquisition subunit is used to extract approval vector data from the provincial facility agriculture land supervision system through the API interface, and simultaneously acquire high-resolution image data of the corresponding area from the natural resources satellite remote sensing platform. The patch extraction subunit is used to extract edges from image data based on the Canny edge detection algorithm. Combined with the spatial range of the approval vector data, the R-CNN target detection model is used to identify and extract the changing patches and generate the initial changing patch vector. The overlap calculation subunit is used to calculate the overlap between the initial changed patch and the approval vector range using the Jaccard index, and generate the initial patch dataset.

3. A facility agriculture land use system according to claim 2, characterized in that, The vector data processing unit (1) further includes: The data standardization sub-unit is used to unify the coordinates of the initial patch dataset according to the geodetic coordinate system, verify the integrity of patch attributes, generate correction prompts for data with incorrect format or missing attributes, and re-incorporate the corrected data into the initial patch dataset. The invalid data filtering subunit is used to exclude construction land and non-facility agricultural land vectors of unused land in the initial map patch dataset through spatial overlay analysis, and only retain map patch data that conforms to the land use type prediction, thereby improving the efficiency of subsequent evidence collection and review.

4. A facility agriculture land use system according to claim 1, characterized in that, The field intelligent evidence collection unit (2) includes: The evidence standard matching subunit is used to receive the initial map patch dataset and automatically match the evidence standard template according to the land use type in the map patch attributes. The crop planting category should include panoramic photos of the greenhouse and shed exterior and photos of the crop growth status inside. The livestock and aquaculture category should include photos of the pens and breeding pond exterior and interior facilities. The factory planting category should include photos of the steel frame structure and temperature control equipment details. The task allocation subunit is used to divide the evidence collection area based on the spatial distribution of map patches. Combining the qualifications and geographical location of county-level evidence collection personnel, the Hungarian algorithm is used to perform the optimal matching of evidence collection tasks and personnel, and generate an evidence collection task list containing the area scope, task time limit, and standard requirements. The on-site data collection subunit is used by evidence collectors to receive a task list via a mobile app, scan the QR code of the map patch attribute to associate data, take photos according to the template requirements, and the app automatically adds the shooting time and GPS location watermark. The associated packaging subunit is used to establish an association mapping between photos and land parcel numbers, land use types, and evidence providers, generate a JSON-formatted evidence list, and package it with the photo files into a ZIP-formatted evidence data package with a unique file number.

5. A facility agriculture land use system according to claim 4, characterized in that, The field intelligent evidence collection unit (2) also includes: The compliance verification subunit is used to automatically verify the photos in the evidence data package. If the verification fails, a rectification prompt is generated. The evidence traceability subunit is used to add operation logs to the verified evidence data packets, record the first collection time, rectification time and verification results, form a traceable evidence file, and push it to the three-level joint review unit (3).

6. A facility agriculture land use system according to claim 1, characterized in that, The three-level joint review unit (3) includes: The county-level preliminary review subunit is used to receive evidence data packets, extract map patch attributes and photos, compare them with the positive and negative lists of facility agricultural land, determine whether the production auxiliary facilities in the positive list are compliant, and determine whether the leisure tourism and deep processing sites in the negative list are non-compliant, and generate a county-level preliminary review report containing compliance basis and preliminary review results. The municipal-level review subunit is used to receive the county-level preliminary review report, focus on verifying the map patches with questionable land categories, supplement the review opinions by retrieving photo details and comparing historical changes in images, and require provincial consultation for the marking of map patches that are still questionable, and generate a municipal-level review report to be sent to the provincial level. The provincial final review sub-unit is used by the provincial farmland red line control technology center to review the map plots without any doubts, and to write the map plot information that has changed the land use to facility agricultural land into the comprehensive monitoring and supervision module. The provincial natural resources survey center conducts random checks according to a certain proportion, and uses stratified sampling to ensure that all types of land use are covered. The consultation and resolution subunit is used to organize a consultation meeting between the two departments to address the discrepancies in land category identification discovered during the spot check, record the points of disagreement, the basis for the consultation, and the final conclusion, and generate a provincial final review report. The review conclusion generation sub-unit is used to integrate the three levels of review opinions: for map patches that have passed the review, a review approval conclusion and land category change confirmation form are generated; for map patches that need to be rectified, a rectification notice is generated, thus forming a review conclusion document.

7. A facility agriculture land use system according to claim 6, characterized in that, The three-level joint review unit (3) also includes: The standard update sub-unit is used to receive the updated content of the technical regulations for land change surveys every quarter, automatically update the built-in positive and negative lists and the detailed change scope table, and generate explanatory documents on the differences between the old and new standards for reviewers to refer to. The review efficiency statistics sub-unit is used to calculate the review time, pass rate, and rectification rate of each stage. It generates visual efficiency analysis charts using Matplotlib, which serve as the basis for process optimization and are simultaneously incorporated into the review conclusion document.

8. A facility agriculture land use system according to claim 1, characterized in that, The structured results management unit (4) includes: The data integration subunit is used to receive review conclusion documents, automatically associate the corresponding evidence data packets and initial map patch datasets, establish association indexes according to file number, map patch number and review conclusion, and build a structured database using PostgreSQL+PostGIS database; The Task Overview sub-unit is used to read structured database data in real time, display the progress of tasks at each stage, pending tasks, and review pass rate statistics, and generate a web-based visual task dashboard. The above image entry sub-unit is used by county-level users to access approved map data from the structured database, automatically match the coordinate system of the land and space planning database, complete the image entry operation, and generate vector results in SHP format.

9. A facility agriculture land use system according to claim 8, characterized in that, The structured results management unit (4) also includes: The data preprocessing subunit is used to receive batch query requests from users, automatically process database data, including data deduplication, format conversion, and attribute completion, and generate preprocessed data packages. The Historical Archives and Decision Support subunit is used to build historical archives by year and land type, support the retrieval of historical data, analyze land use change trends through ARIMA time series model, and generate decision analysis reports on annual growth rate and land type structure ratio for management departments to refer to.

10. A facility agriculture land use system according to claim 1, characterized in that, The cross-module data push unit (5) includes: The data filtering subunit is used to filter map patch data that have passed review and been included in daily changes from the structured database, and extract the map patch number, approval scope and land category identification result fields; The cross-module push sub-unit is used to push the filtered data to the list of surveyed and approved plots via a RESTful API interface, with a link to the provincial final review report for viewing by the farmland protection department. The percentage analysis subunit is used to retrieve the area of ​​the approval range and the area of ​​the passed-through patches from the initial patch dataset, obtain the percentage, generate a bar chart of the percentage ranking for each region using ECharts, and display it on the cloud platform visualization interface. The closed-loop audit sub-unit is used to receive map data of the approved scope changed to facility agricultural land selected by the county level, push it to the provincial facility agricultural land supervision system, generate an audit completion certificate containing audit time, audit personnel and map list, and send the certificate back to the structured results management unit (4) for archiving, thus completing the data closed loop of the whole process.

Images