Supplementary cultivated land identification method based on fusion of spatio-temporal data and remote sensing image

Abstract

The invention discloses a supplementary cultivated land affirmation method based on spatio-temporal data and remote sensing image fusion, and relates to the field of remote sensing, and the technical scheme of the invention comprises the following steps: S1, the supplementary cultivated land is planned to be accurately pre-judged and anchored in the early stage of newly added cultivated land; s2, refined determination and calculation of newly added cultivated land and supplemented cultivated land; and S3, intelligent judgment and grading verification of the result quality. By fusing multi-temporal high-resolution remote sensing images and various vector spatio-temporal data, fuzzy experience judgment is converted into accurate mathematical and spatial rules, and standardization, transparency and intelligence of supplementary cultivated land determination are achieved.

Description

A supplementary method for identifying arable land based on the fusion of spatiotemporal data and remote sensing imagery Technical Field

[0001] This invention relates to the field of remote sensing, and in particular to a method for identifying supplementary arable land based on the fusion of spatiotemporal data and remote sensing images. Background Technology

[0002] Supplementing arable land is the core mechanism for implementing the balance between arable land occupation and replenishment. It refers to the process of developing unused land and abandoned construction land into arable land through land consolidation. It is necessary to ensure the balance of quantity and pay more attention to quality improvement and ecological protection. At present, the identification of supplementary arable land mainly relies on single land change survey data or phased remote sensing interpretation after project implementation. These methods generally have the following problems and shortcomings: (1) The calculation logic is vague and lacks process traceability: Existing methods mostly focus on the consistency comparison of the final result between the map and the data. They lack a refined and verifiable quantitative description of the dynamic transformation process of supplementary arable land from "plan" to "realization". The area calculation is often based on the overall map patch deducting a fixed coefficient. It fails to accurately separate and deduct non-cultivated areas such as field ridges and unused land, resulting in an opaque identification logic and the calculation results are easily questioned.

[0003] (2) Limited technical means, unable to verify the authenticity of the project: Existing technologies usually rely on static, point-in-time data, which makes it difficult to effectively verify the fundamental premise that "the land was non-arable before implementation and became arable after implementation". The lack of intelligent comparison of land use changes and project traces before and after implementation using high-resolution, multi-temporal remote sensing images makes it impossible to technically eliminate the risk of "misattribution" or "digital land creation".

[0004] (3) Significant audit risks and weak internal control: In the existing process, the spatial and quantitative logical relationships between the planning scope, implementation scope, identification scope, and engineering scope lack mandatory and automated verification mechanisms. Data at each stage (vector range, area ledger) is prone to human error or logical contradictions, which will expose high-risk problems such as inconsistent data and broken supporting material chains in subsequent natural resource supervision and audit. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method for identifying supplementary arable land based on the fusion of spatiotemporal data and remote sensing imagery. By fusing multi-temporal high-resolution remote sensing images with various vector spatiotemporal data, fuzzy empirical judgments are transformed into precise mathematical and spatial rules, achieving standardization, transparency, and intelligence in the identification of supplementary arable land.

[0006] The technical solution adopted by this invention to solve its technical problem is: a method for identifying supplementary arable land based on the fusion of spatiotemporal data and remote sensing images, including the following steps: S1 Precise prediction and anchoring of newly added arable land in the early planning stage; S2 Refined identification and calculation of newly added arable land and supplementary arable land; S3 Intelligent judgment and graded verification of the quality of the results.

[0007] Preferably, in step S1, based on the latest annual land change survey database, all non-arable land parcels within the project area are extracted as a potential base; high-resolution orthophotos with resolutions better than 0.2 meters are integrated with the field images collected before project initiation, and manual or semi-automatic fine interpretation is performed to further accurately delineate the specific plots of land within the potential base area for planned supplementary arable land; this process must meet the core constraint condition: S(P)≤S(N), where S(P) represents the total area of ​​planned supplementary arable land, and S(N) represents the total area of ​​non-arable land in the corresponding area of ​​the land change survey within the project area; this step anchors the supplementary arable land task from macro-indicators to specific spatial coordinates (x, y, y). i ,y i The plan baseline is formed by recording the geometric shape and its area S(P).

[0008] Preferably, step S2 includes the following: (1) accurate extraction of the scope after implementation; (2) intelligent verification of the rationality of land use changes; (3) precise calculation model of net newly added cultivated land area; and (4) multi-level data consistency.

[0009] Preferably, regarding the precise extraction of the scope after implementation: after the project is completed, high-definition orthophotos with a resolution better than 0.2 meters are obtained on-site; based on the planned scope set P determined in step S1, and combined with the actual completed land boundary, the scope of the final supplementary cultivated land to be identified is delineated, denoted as set A; within this scope, field ridges with a width greater than 1 meter are accurately identified and delineated, denoted as set K, and K⊂A.

[0010] Preferably, the intelligent verification of the rationality of land use changes is performed by overlaying and analyzing pre-implementation imagery (Img). pre Post-implementation image Img post For each plot in set A, an automated or interactive comparison is performed; the following verification rule is constructed: ∀P∈A, Class(Imgpre(P))∈{non-arable land} ∧ Class(Imgpost(P))∈{arable land}. This rule ensures that the source of each identified arable land plot is legal and the project is genuine and valid; preferably, regarding the precise calculation model for the net newly added arable land area: let A be the i-th supplementary arable land plot participating in the identification. i The area of ​​the corresponding patch is S(A) i Within this image patch: Refined subtraction of field ridge area: For field ridges wider than 1 meter, based on their image texture features and continuous length lij It can be divided into m line segment units; the deducted area of ​​each unit takes into account not only its average width w ij It also introduces an effectiveness coefficient α based on vegetation coverage. ij (0≤α ij ≤1), used to characterize whether the field ridge is necessary for agricultural production; completely bare or hardened field ridges α≈1, field ridges partially covered with crops α<1); then the total area of ​​field ridges to be deducted within the i-th plot is: The total area of ​​the field ridges in the project, after deducting the area, is: Non-arable land area deduction: For areas U within the scope of supplementary arable land parcels that are still non-arable land after the calculation of newly added arable land areas, the relationship is U⊂P∩A; in, Let be the area of ​​the i-th non-cultivated map patch.

[0011] ③ Comprehensive formula for net increase in cultivated land area: Therefore, the total net increase in cultivated land area S Net The calculations are based on the following models: Preferably, regarding multi-level data consistency: ① Range inclusion constraint (hard constraint): Where R represents the project's boundary line, and L represents the scope of the land leveling work to be carried out by the project. To ensure project coverage, the planned area should, in principle, be covered by the project implementation area by at least 1. Error minimization model for area consistency check: The ledger record value S record (K), S record (U) Compare with the spatially calculated value, and define the permissible relative error threshold between the ledger recorded area and the spatially calculated area as ϵ; Consistency of field ridge area: Minimize the objective function F. K : Constraints: F K ≤ϵ;where K i It falls within the identified map patch A i Within the field ridges; consistency of non-cultivated land area: minimizing the objective function F U : Constraints: F U ≤ϵ; where P i To be with A i The corresponding part of the original plan.

[0012] Advantages of this invention: By integrating multi-temporal high-resolution remote sensing images with various vector spatiotemporal data, this invention transforms fuzzy empirical judgments into precise mathematical and spatial rules, thereby achieving standardization, transparency, and intelligence in the identification of supplementary arable land. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only one of the drawings of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 is a flowchart of an embodiment of the present invention. Detailed Implementation

[0015] To enhance understanding of the present invention, it will be further described in detail below with reference to the accompanying drawings and embodiments. These embodiments are only for explaining the present invention and do not constitute a limitation on the scope of protection of the present invention. Embodiments

[0016] As shown in Figure 1, the present invention aims to solve the problems in the prior art. Its core lies in building a technical system that integrates "planning-implementation-identification-verification" and provides a closed-loop management of the entire chain.

[0017] This method integrates multi-temporal high-resolution remote sensing images with various vector spatiotemporal data to transform fuzzy empirical judgments into precise mathematical and spatial rules, achieving standardization, transparency, and intelligentization in the identification of supplementary arable land. The technical principles and workflow are as follows: S1: Precise Prediction and Anchoring of Newly Added Arable Land Plans. Based on the latest annual land change survey database, all non-arable land parcels within the project area are extracted as a potential baseline. High-resolution orthophotos (better than 0.2 meters) collected before project initiation are integrated, and manual or semi-automatic fine interpretation is performed. Within the potential baseline, the specific plot range of the planned supplementary arable land (denoted as set P) is further precisely delineated. This process must satisfy the core constraint: S(P) ≤ S(N), where S(P) represents the total area of ​​planned supplementary arable land, and S(N) represents the total area of ​​non-arable land surveyed within the corresponding area of ​​the project area. This step anchors the supplementary arable land task from macro-indicators to specific spatial coordinates (xi, yi) and geometric shapes, and records its area S(P), forming a traceable "planning baseline."

[0018] S2: Refined Identification and Calculation of Newly Added Cultivated Land (1) Precise Extraction of Scope After Implementation: After the project is completed, high-definition orthophotos with a resolution better than 0.2 meters are obtained on-site. Based on the planned scope set P determined in Step 1, and combined with the actual completed land boundary, the scope of the final identified supplementary cultivated land is delineated (denoted as set A). At the same time, within this scope, field ridges with a width greater than 1 meter are accurately identified and delineated (denoted as set K, and K⊂A), as well as areas that are within the planned scope but are still confirmed as not being valid cultivated land after implementation (such as scattered unused land, small structures occupying land, road occupancy, etc., denoted as set U, and U⊂P).

[0019] (2) Intelligent verification of the rationality of land use change: by overlaying and analyzing the pre-implementation image Img pre Post-implementation image Img post For each plot in set A, perform automated or interactive comparison. Construct the following verification rule: ∀p∈A,Class(Img pre (p))∈{non-arable land} ∧ Class(Img post (p))∈{arable land} This rule ensures that each piece of land identified as arable land has a legitimate origin (formerly non-arable land) and that the project is genuine and valid (has been transformed into arable land).

[0020] (3) Precise Calculation Model for Net Newly Added Cultivated Land Area The calculation of the net newly added cultivated land area not only considers the addition and subtraction of areas, but also introduces weighting factors based on image features and spatial topology to ensure the scientific validity and robustness of the calculation results. The model is described as follows: Let A be the i-th supplementary cultivated land patch participating in the identification. i The area of ​​the corresponding patch is S(A) i Within this image patch: ① Refined subtraction of field ridge area: For field ridges wider than 1 meter, based on their image texture features and continuous length l ij This can be divided into m line segment units. The area deducted from each unit takes into account not only its average width w. ij (Through image measurements), an effectiveness coefficient α based on vegetation cover is also introduced. ij (0≤α ij ≤1), used to characterize whether the field ridge is necessary for agricultural production (completely bare or hardened field ridges α≈1, field ridges partially covered with crops α<1). Then the total area of ​​field ridges to be deducted within the i-th plot is: The total area of ​​the field ridges in the project, after deducting the area, is: ② Deduction of non-arable land areas: For areas U within the scope of the supplementary arable land area that are still non-arable land after the calculation of newly added arable land areas, the relationship is U⊂P∩A. in, Let be the area of ​​the i-th non-cultivated map patch.

[0021] ③ Comprehensive formula for net increase in cultivated land area: Therefore, the total net increase in cultivated land area S Net The calculations are based on the following models: (4) Multi-level data consistency transforms consistency checks from discrete rule judgments into a comprehensive constraint optimization and error minimization problem.

[0022] ① Scope inclusion constraint (hard constraint): Where R represents the project boundary, L represents the land leveling area to be implemented, and "θ" represents the project coverage, which in principle requires that the planned area be covered by the project implementation area by at least 1.

[0023] ② Error minimization model for area consistency check (soft constraint): The ledger record value S... record( K), S record (U) is compared with the spatially calculated value, and the permissible relative error threshold between the ledger recorded area and the spatially calculated area is defined as ϵ.

[0024] Consistency of field ridge area: Minimize the objective function F K : Constraints: F K ≤ϵ. Where K i It falls within the identified map patch A i The inner ridge section.

[0025] Non-arable land area consistency: Minimize the objective function F U : Constraints: F U ≤ϵ. Where P i To be with A i The corresponding part of the original plan.

[0026] Step 3: Intelligent Assessment and Grading Verification of Results Quality, Automated Rule Screening and Risk Classification in Internal Processing: The system uses the planned scope set P, the identified scope set A, the land use change verification results, and the net increase in cultivated land area generated in Steps 1 and 2 as the basis for the assessment. , field ridges deducted from total area Non-arable land deducted from total area and area consistency error F K F U The system loads a pre-defined set of quantitative judgment rules for batch comparison and analysis. The core quantitative indicators of the judgment rule set are defined as follows: ① Range compliance indicator I R : ② Land Category Compliance Rate R C : ③Area Consistency Index I A : ④ Logical rationality index I L : Based on the values ​​of the above indicators, the results are automatically divided into three categories: Category A (low-risk results): All core rules are met. Specifically, this includes: simultaneously meeting I... R =1, R C ≥T C (e.g. T) C =0.98), I A =1, I L= 1. Such results indicate that all core quality control points meet the requirements.

[0027] Category B (results with medium - risk doubts): The core framework is compliant, but there are abnormalities or uncertainties in some non - key rules. It mainly includes: It needs to meet I R = 1, but R C , I A , I L At least one of them does not meet the Category A standard and does not trigger the Category C condition. The specific manifestations are: T C_low ≤R C <T C (The land type is ambiguous, the remote - sensing image features are in the critical state between cultivated land and non - cultivated land, and the automatic interpretation confidence is low), or at least one of F K , F U satisfies ϵ < F ≤ δ*ϵ (for example, δ = 1.5, the error slightly exceeds the threshold but is explainable), or there are other explainable non - principle abnormalities.

[0028] Category C (high - risk unqualified results): Violating one or more principle rules. It mainly includes: Any of the following conditions is judged as Category C: I R = 0 (range violation), R C <T C_low (For example, T C_low = 0.90, there are major doubts about the source of the land type), I L = 0 (logical error or S Net <0), or there is a serious and unexplainable contradiction between the ledger of key area data (such as deducted land types) and the spatial calculation results (F K >δ*ϵ or F U >δ*ϵ).

[0029] (2) Hierarchical disposal and manual intervention for review For Category A (low - risk results): Directly enter the "random sampling field verification" process.

[0030] For Category B (results with medium - risk doubts): Discriminate based on high - definition images, auxiliary materials and experience: If it can be confirmed that the abnormality is within the acceptable range or can be eliminated through in - house correction, it will be classified into Category A; if it is confirmed that there are substantial problems, it will be downgraded to Category C; if it is still impossible to make a definite conclusion in - house after research, the relevant map patches will be marked as "set Q of map patches that must be checked in the field".

[0031] For Category C (high - risk unqualified results): Issue a review result with a detailed problem list and evidence screenshots, determine it as unqualified, and return the entire result, requiring the identifying unit to rectify it within a time limit and re - apply for identification.

[0032] (3) Differentiated sampling and closed-loop judgment for field verification: Field verification adopts a differentiated sampling strategy based on risk level to balance verification efficiency and control effect.

[0033] ① Let N be the total number of map patches identified in the project. total .

[0034] ② Let the set of "must-check map features in the field, Q" contain N map features. Q .

[0035] ③ Randomly select N number of map spots from the Class A result map spots. rand , must satisfy N rand =max([ℎ×(N total -N Q )],1), ℎ is the field sampling ratio. For example, ℎ=0.05 means that at least 5% of the Class A (low-risk results) portion will be sampled.

[0036] ④ Total number of field verification sites N field for:

[0037] ①The inspectors checked N field Each map patch was inspected on-site. Let N be the number of non-compliant map patches found during the inspection. fail .

[0038] ②If N fail If the result is 0, the result is considered to have passed the final verification.

[0039] If N fail If the value is >0, calculate the field verification failure rate ρ:

[0040] ③ If ρ > λ (for example, λ = 0.20), then the overall result of the project is deemed unqualified and should be returned for redo.

[0041] If ρ≤λ, the nature of the unqualified plots will be further analyzed. If it does not fall under fundamental issues such as "violation of scope" or "illegal land use source," targeted rectification of the unqualified plots and similar plots will be required, and they will be resubmitted for review. If it involves fundamental issues, it can still be determined as unqualified as a whole or trigger a more stringent comprehensive review.

[0042] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0043] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for identifying supplementary arable land based on the fusion of spatiotemporal data and remote sensing imagery, characterized in that, The process includes the following steps: S1 Precise prediction and anchoring of newly added arable land in the early planning stage; S2 Refined identification and calculation of newly added arable land and supplementary arable land; S3 Intelligent judgment and graded verification of the results quality.

2. The method for identifying supplementary cultivated land based on the fusion of spatiotemporal data and remote sensing imagery as described in claim 1, characterized in that: Step S1 involves extracting all non-arable land parcels within the project area as a potential baseline based on the latest annual land change survey database. This is then integrated with high-resolution orthophotos (better than 0.2 meters) collected before project initiation, and subjected to manual or semi-automatic fine interpretation. Within the potential baseline, the specific land parcels planned for supplementary arable land are further precisely delineated. This process must satisfy the core constraint: S(P) ≤ S(N), where S(P) represents the total planned supplementary arable land area, and S(N) represents the total area of ​​non-arable land surveyed within the corresponding area of ​​the project area. This step anchors the supplementary arable land task from macro-level indicators to specific spatial coordinates (x, y, y). i ,y i The plan baseline is formed by recording the geometric shape and its area S(P).

3. The method for identifying supplementary cultivated land based on the fusion of spatiotemporal data and remote sensing imagery as described in claim 1, characterized in that: Step S2 includes the following: (1) accurate extraction of the scope after implementation; (2) intelligent verification of the rationality of land type changes; (3) precise calculation model of net newly added cultivated land area; and (4) multi-level data consistency.

4. The method for identifying supplementary cultivated land based on the fusion of spatiotemporal data and remote sensing imagery according to claim 1, characterized in that: Regarding the precise extraction of the scope after implementation: After the project is completed, high-definition orthophotos with a resolution better than 0.2 meters are obtained on site; based on the planned scope set P determined in step S1, and combined with the actual completed land boundary, the scope of the final supplementary cultivated land to be identified is delineated, denoted as set A; within this scope, field ridges with a width greater than 1 meter are accurately identified and delineated, denoted as set K, and K⊂A.

5. The method for identifying supplementary cultivated land based on the fusion of spatiotemporal data and remote sensing imagery according to claim 1, characterized in that: Intelligent verification of the rationality of land use changes: This is achieved through overlay analysis of pre-implementation imagery (Img). pre Post-implementation image Img post For each plot in set A, perform automated or interactive comparison; construct the following verification rule: ∀P∈A,Class(Img pre (P))∈{non-arable land} ∧ Class(Img post (P))∈{arable land} This rule ensures that the source of each piece of identified arable land is legal and that the project is genuine and valid.

6. The method for identifying supplementary cultivated land based on the fusion of spatiotemporal data and remote sensing imagery according to claim 1, characterized in that: Regarding the precise calculation model for the net increase in cultivated land area: Let A be the i-th supplementary cultivated land parcel involved in the identification. i The area of ​​the corresponding patch is S(A) i Within this patch: Refined area subtraction of field ridges: For field ridges wider than 1 meter, the area is subtracted based on their image texture features and continuous length l. ij It can be divided into m line segment units; the deducted area of ​​each unit takes into account not only its average width w ij An effectiveness coefficient α based on vegetation cover is also introduced. ij (0≤α ij ≤1), used to characterize whether the field ridge is necessary for agricultural production; completely bare or hardened field ridges α≈1, field ridges partially covered with crops α<1); then the total area of ​​field ridges to be deducted within the i-th plot is: The total area of ​​the field ridges in the project, after deducting the area, is: Non-arable land area deduction: For areas U within the scope of supplementary arable land parcels that are still non-arable land after the calculation of newly added arable land areas, the relationship is U⊂P∩A; in, For the first The area of ​​a non-cultivated map patch; The comprehensive formula for net newly added cultivated land area is: Therefore, the total net newly added cultivated land area S Net The calculations are based on the following models:

7. The method for identifying supplementary cultivated land based on the fusion of spatiotemporal data and remote sensing imagery according to claim 1, characterized in that: Regarding multi-level data consistency: ① Range inclusion constraint (hard constraint): Where R represents the project's boundary line, and L represents the scope of the land leveling work to be carried out by the project. To ensure project coverage, the planned area should, in principle, be covered by the project implementation area by at least 1. Error minimization model for area consistency check: The ledger record value S record (K), S record (U) Compare with the spatially calculated value, and define the permissible relative error threshold between the ledger recorded area and the spatially calculated area as ϵ; Consistency of field ridge area: Minimize the objective function F. K : Constraints: F K ≤ϵ;where K i It falls within the identified map patch A i Within the field ridges; consistency of non-cultivated land area: minimizing the objective function F U : Constraints: F U ≤ϵ;where P i To be with A i The corresponding part of the original plan.

Images