Urban expansion pattern recognition method based on mathematical morphology and landscape connectivity

Abstract

The application relates to a city expansion mode recognition method fusing mathematical morphology and landscape connectivity, and belongs to the cross field of city planning and geographic information technology. The method comprises the following steps: recognizing corridor expansion based on road data; recognizing city holes through seed filling and morphological closing operation, and then recognizing infill expansion located in the holes; recognizing enclaves through the calculation of the minimum expansion step number based on eight-neighborhood iterative expansion; and fusing connectivity and linear morphology to recognize corridor expansion and edge expansion. The method provided by the application, namely morphological landscape expansion index MLEI, has significantly improved classification accuracy compared with traditional landscape expansion index LEI, and is suitable for multi-scale city evolution analysis.

Description

Technical Field

[0001] The present invention belongs to the interdisciplinary field of urban planning and geographic information technology, and specifically relates to an urban expansion pattern recognition method that integrates mathematical morphology and landscape connectivity. Background Art

[0002] Traditional landscape expansion index analysis methods (such as the LEI index) rely on buffer zone algorithms and have the following limitations:

[0003] Scale sensitivity: The buffer distance parameter affects the stability of the criterion. A too small buffer distance parameter leads to misjudgment (overestimation) of enclave expansion, while a too large buffer distance parameter leads to misjudgment (overestimation) of infill expansion.

[0004] Morphological interference: Differences in patch geometry cause variations in the spatial configuration of the buffer zone, leading to pattern misclassification (e.g., complex boundary patches are misclassified as infill patches);

[0005] Process representation bias: failure to distinguish active infill from passive sprawl (e.g., non-contact infill expansion due to green space spacing is misclassified as enclave expansion);

[0006] Blind spots in the existing classification system: lack of ability to quantitatively characterize landscape connectivity (such as axial expansion guided by transportation corridors) (the corridor expansion proposed by Forman has not been included in the mainstream system). Summary of the Invention

[0007] The purpose of the present invention is to address the identification defects of traditional landscape expansion indices in terms of scale sensitivity, morphological interference, process representation deviation and landscape connectivity, fill the blind spots in the identification of corridor expansion, and provide an urban expansion pattern identification method that integrates mathematical morphology and landscape connectivity.

[0008] To achieve the above objectives, the technical solution of the present invention is: a method for identifying urban expansion patterns that integrates mathematical morphology and landscape connectivity, comprising:

[0009] Identify corridor expansion based on road data;

[0010] Urban holes are identified through seed filling and morphological closing operations, and then the infill expansion inside the holes is identified;

[0011] Identify enclave expansion based on eight-neighborhood iterative expansion to calculate the minimum number of expansion steps;

[0012] Integrate connectivity and linear morphology to identify corridor expansion and edge expansion.

[0013] Furthermore, corridor expansion is extracted based on road data, and the following is achieved:

[0014] Road data cleaning;

[0015] Extract corridor expansion based on road data;

[0016] Four-neighborhood connectivity iterative marking corridor expansion is adopted.

[0017] Furthermore, corridor expansion is extracted based on road data. The specific implementation steps are as follows:

[0018] S11. Road data cleaning:

[0019] Calculate road length;

[0020] Identify dead-end roads: Locate line segments whose ends are not connected to other roads through topological inspection;

[0021] Eliminate dead-end roads with a length of less than 1 km;

[0022] S12. Corridor expansion based on road data extraction:

[0023] Road buffer generation: With the road processed in step S11 as the center, create a parallel buffer with the spatial resolution of the urban land image as the width (the minimum must not be less than 10m, alignment error);

[0024] Extract pixels from the newly added urban land imagery that are completely or partially within the buffer zone;

[0025] S13, using four-neighborhood connectivity iterative marking corridor expansion:

[0026] Based on the pixels in the newly added urban land image extracted by S12, the four-neighborhood connectivity is used to iteratively extract the pixels in the newly added urban land image, and the number of iterations is S:

[0027] S=2K / L

[0028] K is the spatial resolution of urban land images; L is the maximum width of urban roads;

[0029] All pixels in the extracted newly added urban land images are marked as corridor expansion and assigned a morphological landscape expansion index (MLEI) of 0.5.

[0030] Furthermore, urban holes are identified through seed filling and morphological closing operations, and then the in-fill expansion within the holes is identified, achieving the following:

[0031] The four-neighborhood connectivity is used to scan the pixels belonging to non-urban areas in the original urban land image;

[0032] Establish a mechanism to eliminate pseudo holes at the boundary contact point to filter out pseudo holes at the boundary;

[0033] Use closing operation to form effective holes;

[0034] Based on the pixels in the newly added urban land image that are not marked as corridor-type expansion, infill expansion located within valid holes is identified.

[0035] Furthermore, urban holes are identified through seed filling and morphological closing operations, and then the in-fill expansion within the holes is identified. The specific implementation steps are as follows:

[0036] S21. Generate background template

[0037]

[0038] Among them, (x, y) is the pixel coordinate; Represents the pixel value of the original urban land image, where the pixel value of the pixel belonging to the urban area is 1, and the pixel value of the pixel belonging to the non-urban area is 0; The pixel value representing the pixel in the background;

[0039] S22. Identify hole candidate areas

[0040] Use the four-neighborhood connectivity rule to traverse the background set I bg For all cells in , do the following:

[0041] S221, initialize the marking matrix M, whose size is the same as I bg Same, the initial value is not visited;

[0042] S222, Sequential Scan I bg For each pixel in, if the current pixel coordinate (x, y) satisfies I bg If (x, y) = 1 and M(x, y) is unvisited, the seed filling algorithm is executed:

[0043] S2221. Create a hole candidate area set H=0;

[0044] S2222, create a new hole candidate area H k ;

[0045] S2223. Add (x, y) to queue Q;

[0046] S2224. Loop when queue Q is not empty:

[0047] (1) Take pixel coordinates from Q (P x ,P y );

[0048] (2) If (P x ,P y ) is within the image range and I bg (P x ,P y) = 1 and it has not been visited, then mark it as visited;

[0049] (3) (P x ,P y ) Add hole candidate area H k ; will (P x ,P y ) The four neighboring pixels are added to the queue Q;

[0050] S2225, Save H k The set of candidate holes H={H1,H2,…,H k};

[0051] S23, filter out pseudo holes at the boundary

[0052] S231, for each hole candidate region H obtained in step S22 k Perform boundary contact detection. If there is any pixel coordinate (x, y)∈H k If any of the following conditions is met, then H k For pseudo holes, remove H from H k ;

[0053] (1) x = 1 or x = W, where W is the number of image columns;

[0054] (2) y = 1 or y = H, where H is the number of image rows;

[0055] S232: After the elimination operation in step S231, the remaining set is recorded as the effective hole set H. valid ={H valid,1 ,…,H valid,m};

[0056] S24, hole morphology optimization

[0057] For each effective hole H valid,m Perform morphological closing:

[0058]

[0059] in represents the expansion operation, Indicates corrosion operation; Represented as a binary rectangle of preset size and geometry, used to control the optimization range of the hole boundary by dilation and erosion operations;

[0060] S25, internal filling expansion marker

[0061] For newly added urban land images that are not marked as corridor expansion, infill expansion is determined:

[0062] For the newly added urban land image binary raster data I that is not marked as corridor expansion new For each pixel in , if it satisfies: I new (x,y) H valid,m , then mark I new The corresponding pixel in the image is expanded by filling in, and the morphological landscape expansion index MLEI=0 is assigned. new (x, y) is the binary raster data of the newly added urban land image that is not marked as corridor expansion. new The pixel value of the pixel in .

[0063] Furthermore, the minimum expansion steps are calculated based on the eight-neighborhood iterative expansion, that is, the eight-neighborhood rule is used to set the expansion direction, and the minimum expansion steps N is defined as the number of expansion steps at which the expansion buffer intersects the original urban patch for the first time.

[0064] Furthermore, the minimum number of expansion steps is calculated based on the iterative expansion of the eight-neighborhood to identify the enclave expansion. The specific implementation method is as follows:

[0065] S31. Identify candidate patches

[0066] For the pixels in the newly added urban land image that are not marked as corridor expansion and infill expansion, a four-neighborhood connected component analysis is performed:

[0067] S311, initializing the access tag matrix;

[0068] S312, traverse all pixels in the newly added urban land image that are not marked as corridor expansion and infill expansion, use the four-neighborhood rule to identify and obtain candidate patches P new ;

[0069] S32, Initialization

[0070] S321, for each candidate patch P new , create a dilated result region X n , initially X0=P new ;

[0071] S322, define the structural element S dil : Use 8-neighborhood structure elements;

[0072] S323, set the iteration counter n=0;

[0073] S324, set Target to the original urban land image binary raster data I old ;

[0074] S33, Iterative Expansion

[0075] S331, cycle starts:

[0076] (1) Perform an expansion operation:

[0077]

[0078] Where P is P new The midpoint of a single independent patch, S dil,p The structural element S is translated with P as the center dil ;

[0079] (2) Increasing: n=n+1;

[0080] S332, intersection check

[0081] (1) Check the expanded area X n Whether there is an intersection with Target, calculation formula:

[0082]

[0083] (2) Conditional judgment:

[0084] If Intersect≠ : Stop expansion and record the minimum number of expansion steps N=n, that is, X after N steps of expansion n Contact with the original city;

[0085] If Intersect= : Return to step S331 to continue the expansion process;

[0086] S34. Preliminary classification

[0087] S341, record the minimum number of expansion steps N;

[0088] i. S342, if N ≥ 2, the corresponding candidate patch P new It is initially marked as enclave expansion and assigned the morphological landscape expansion index MLEI = N; if N = 1, the corresponding candidate patch P new Labeled as potential edge expansion or corridor expansion.

[0089] Furthermore, the perimeter area ratio PAR is calculated as follows:

[0090]

[0091] in is the candidate patch P new The circumference of is the candidate patch P new area.

[0092] Furthermore, the eccentricity ECC is calculated as follows:

[0093]

[0094] Where Length major is the candidate patch P new The major axis length of the minimum circumscribed rectangle MBR, Length minor is the minor axis length of the MBR.

[0095] Furthermore, the method of the present invention constructs the morphological landscape expansion index MLEI, whose mathematical expression is as follows:

[0096]

[0097] Where MLEI is the morphological landscape expansion index of the newly added urban patch, MLEI = 0 is infill expansion, MLEI = 0.5 is corridor expansion, MLEI = 1 is edge expansion; MLEI ≥ 2 is enclave expansion; 0 is a constant, indicating that the newly added patch is completely nested in the hole formed by the patch; N is the minimum number of expansion steps, and the value range is a positive integer with an interval of 1; The value range is 0 and 1. According to the definition of MLEI, the smaller the MLEI value, the closer the spatial relationship between the new urban patch and its adjacent existing urban patch, and the smaller the degree of expansion.

[0098] Compared with the prior art, the present invention has the following beneficial effects:

[0099] 1. Eliminate scale sensitivity: 8-neighborhood expansion based on mathematical morphology replaces the fixed buffer distance to reduce the interpretation bias caused by pixel boundary misalignment in traditional methods;

[0100] 2. Suppressing morphological interference: The hole recognition mechanism avoids misclassification caused by complex boundaries;

[0101] 3. Breakthrough in connectivity characterization: integrating corridor expansion for the first time and improving the classification system;

[0102] 4. Process characterization optimization: distinguish between active filling and passive spreading. BRIEF DESCRIPTION OF THE DRAWINGS

[0103] Figure 1 It is the technical roadmap of the present invention;

[0104] Figure 2 The process of hole identification in the present invention;

[0105] Figure 3 This is the process of the expansion path in the present invention;

[0106] Figure 4 Optimize for buffer boundaries;

[0107] Figure 5 It is a result schematic diagram of the present invention. DETAILED DESCRIPTION

[0108] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings.

[0109] The present invention provides an urban expansion pattern recognition method integrating mathematical morphology and landscape connectivity, comprising:

[0110] Identify corridor expansion based on road data;

[0111] Urban holes are identified through seed filling and morphological closing operations, and then the infill expansion inside the holes is identified;

[0112] Identify enclave expansion based on eight-neighborhood iterative expansion to calculate the minimum number of expansion steps;

[0113] Integrate connectivity and linear morphology to identify corridor expansion and edge expansion.

[0114] The following is a specific implementation process of the present invention.

[0115] like Figure 1 As shown in the figure, the present invention integrates mathematical morphology and landscape connectivity to identify urban expansion patterns. Based on the acquired images of newly added urban land and existing urban land in the same region at the same phase, the system is constructed through the four steps of "corridor extraction - hole identification - expansion path - connectivity structure". The specific steps are as follows:

[0116] 1. Corridor extraction

[0117] S11. Road data cleaning

[0118] Calculate road length;

[0119] Identify dead-end roads: Use topology inspection to locate line segments whose ends are not connected to other roads (node ​​connection number = 1);

[0120] Eliminate dead-end roads with a length of less than 1 km;

[0121] S12. Corridor expansion based on road data extraction

[0122] Road buffer generation: With the road processed in step S11 as the center, create a parallel buffer with the spatial resolution of the urban land image as the width (the minimum must not be less than 10m, alignment error);

[0123] Extract new urban land patches that are completely or partially within the buffer zone;

[0124] S13, using four-neighborhood connectivity iterative marking corridor expansion

[0125] Based on the newly added urban land patches extracted by S12, the four-neighborhood connectivity (i.e., upper, lower, left, and right adjacent pixels) is used to iteratively extract the newly added urban land patches. The number of iterations is S:

[0126] S=2K / L

[0127] K is the spatial resolution of urban land images; L is the maximum width of urban roads;

[0128] All extracted patches were marked as corridor expansion (MLEI = 0.5).

[0129] 2. Hole identification (see Figure 2 ,in Figure 2 (a) in the middle is the original urban patch. Figure 2 (b) is hole identification. Figure 2 (c) is a filled expanded plaque)

[0130] S21. Generate background template

[0131]

[0132] Among them, (x, y) is the pixel coordinate; Represents the pixel value of the original urban land image, where the pixel value of the pixel belonging to the urban area is 1, and the pixel value of the pixel belonging to the non-urban area is 0; The pixel value representing the pixel in the background;

[0133] S22. Identify hole candidate areas

[0134] Use the four-neighborhood connectivity rule to traverse the background set I bg For all cells in , do the following:

[0135] S221, initialize the marking matrix M, whose size is the same as I bg Same, the initial value is not visited;

[0136] S222, Sequential Scan I bg For each pixel in, if the current pixel coordinate (x, y) satisfies I bg If (x, y) = 1 and M(x, y) is unvisited, the seed filling algorithm is executed:

[0137] S2221. Create a hole candidate area set H=0;

[0138] S2222, create a new hole candidate area H k ;

[0139] S2223. Add (x, y) to queue Q;

[0140] S2224. Loop when queue Q is not empty:

[0141] (1) Take pixel coordinates from Q (P x ,P y );

[0142] (2) If (P x ,P y ) is within the image range and I bg (P x ,P y ) = 1 and it has not been visited, then mark it as visited;

[0143] (3) (P x ,P y ) Add hole candidate area H k ; will (P x ,P y ) The four neighboring pixels are added to the queue Q;

[0144] S2225, Save H k The set of candidate holes H={H1,H2,…,H k};

[0145] S23, filter out pseudo holes at the boundary

[0146] S231, for each hole candidate region H obtained in step S22 k Perform boundary contact detection. If there is any pixel coordinate (x, y)∈H k If any of the following conditions is met, then H k For pseudo holes, remove H from H k ;

[0147] (1) x = 1 or x = W, where W is the number of image columns;

[0148] (2) y = 1 or y = H, where H is the number of image rows;

[0149] S232: After the elimination operation in step S231, the remaining set is recorded as the effective hole set H. valid ={H valid,1 ,…,H valid,m};

[0150] S24, hole morphology optimization

[0151] For each effective hole H valid,m Perform morphological closing:

[0152]

[0153] in represents the expansion operation, Indicates corrosion operation; Represented as a binary rectangle of preset size and geometry, used to control the optimization range of the hole boundary by dilation and erosion operations;

[0154] S25, internal filling expansion marker

[0155] For newly added urban land images that are not marked as corridor expansion, infill expansion is determined:

[0156] For the newly added urban land image binary raster data I that is not marked as corridor expansion new For each pixel in , if it satisfies: I new (x,y) H valid,m , then mark I new The corresponding pixel in the image is expanded by filling in, and the morphological landscape expansion index MLEI=0 is assigned. new (x, y) is the binary raster data of the newly added urban land image that is not marked as corridor expansion. new The pixel value of the pixel in .

[0157] 3. Expansion Path (see Figure 3 ,in Figure 3 The middle (1) is the urban expansion patch, Figure 3 (2) is the expansion buffer, Figure 3 (3) is the minimum expansion buffer)

[0158] S31. Identify candidate patches

[0159] For the pixels in the newly added urban land image that are not marked as corridor expansion and infill expansion, a four-neighborhood connected component analysis is performed:

[0160] S311, initializing the access tag matrix;

[0161] S312, traverse all pixels in the newly added urban land image that are not marked as corridor expansion and infill expansion, use the four-neighborhood rule to identify and obtain candidate patches P new ;

[0162] i. Connected domain marking: Traverse all pixels in the newly added urban land imagery that are not marked as infill expansion, and use the four-neighborhood connectivity rule (i.e., only consider the upper, lower, left, and right adjacent pixels) to identify all connected components;

[0163] ii. Definition of independent patches: Each set of pixels connected by four consecutive neighborhoods is recorded as an independent candidate new patch ;

[0164] iii. Spatial isolation principle: Any two different and There is no four-neighborhood connected path between them;

[0165] iv. Get candidate patch P new ;

[0166] S32, Initialization

[0167] S321, for each candidate patch P new , create a dilation result layer X n , initially X0=P new ;

[0168] S322, define the structural element S dil : Use an 8-neighborhood (i.e., 3×3 square) structure element (see Figure 4 ); In the key step of generating the minimum buffer zone by the morphological dilation algorithm, the neighborhood-based buffer zone (solid-line frame) of the present invention is iteratively generated according to the 8-domain dilation path, and the buffer zone boundary is strictly aligned with the grid resolution.

[0169] S323, set the iteration counter n=0;

[0170] S324, set Target to the original urban land image binary raster data I old ;

[0171] S33, Iterative Expansion

[0172] S331, cycle starts:

[0173] (1) Perform an expansion operation:

[0174]

[0175] Where P is P new The midpoint of a single independent patch, S dil,p The structural element S is translated with P as the center dil ;

[0176] (2) Increasing: n=n+1;

[0177] S332, intersection check

[0178] (1) Check the expanded area X n Whether there is an intersection with Target, calculation formula:

[0179]

[0180] (2) Conditional judgment:

[0181] If Intersect≠ : Stop expansion and record the minimum number of expansion steps N=n, that is, X after N steps of expansionn Contact with the original city;

[0182] If Intersect= : Return to step S331 to continue the expansion process;

[0183] S34. Preliminary classification

[0184] S341, record the minimum number of expansion steps N;

[0185] S342, if N≥2, the corresponding candidate patch P new It is initially marked as enclave expansion and assigned the morphological landscape expansion index MLEI = N; if N = 1, the corresponding candidate patch P new Labeled as potential edge expansion or corridor expansion.

[0186] Furthermore, we integrate connectivity with linear morphology to identify corridor expansion and edge expansion, which can be achieved as follows:

[0187] S41, the original urban land image binary raster data I old Perform four-neighborhood connected component analysis

[0188] S411, initializing the access tag matrix;

[0189] S412, using the four-neighborhood rule to identify all connected components C old , each connected component is regarded as an independent original urban patch.

[0190] i. Connected domain marking: Traverse all original urban patches in the original urban land image and use the four-neighborhood connection rule (i.e., only consider the upper, lower, left, and right adjacent pixels) to identify all connected components.

[0191] ii. Definition of independent patches: Each set of pixels connected by four consecutive neighborhoods is recorded as an independent candidate new patch .

[0192] iii. Spatial isolation principle: Any two different and There is no four-neighborhood connected path between them.

[0193] iv. Obtain candidate patch C old .

[0194] S42. Identify bridging effects

[0195] S421, in P new Connected component analysis (CCA) is performed on the C old The number of independent connected components C;

[0196] S422, bridging condition judgment:

[0197] i. If C=2: indicates plaque P new It connects two previously disconnected urban area segments. This condition is met before proceeding to the next step of morphological analysis;

[0198] ii. If C≠2: indicates that the patch P new If the bridging effect is not satisfied, it is directly marked as marginal expansion, and its MLEI=1.

[0199] S43. Morphological feature calculation

[0200] S431, two Cs old Area calculation;

[0201] S432, Perimeter-Area Ratio (PAR):

[0202]

[0203] in It is plaque P new The perimeter of (the length of the pixel boundary), It is plaque P new The larger the PAR value, the narrower and longer the patch shape.

[0204] S433, Eccentricity (ECC):

[0205]

[0206] Where Length major It's P new The major axis length (maximum side length) of the minimum circumscribed rectangle MBR, Length minor is the short axis length (minimum side length) of the MBR. The closer the ECC value is to 1, the more elongated the plaque is.

[0207] S44. Corridor Expansion Determination

[0208] S441. For the plaque P that meets the bridging condition new , check its morphological characteristics:

[0209] i. Condition 1: Two Cs old The area is larger than 0.5ha;

[0210] ii. Condition 2: PAR>4.3 (optimal threshold determined experimentally);

[0211] iii. Condition 3: ECC>0.83 (optimal threshold determined experimentally).

[0212] S442. If all three morphological conditions in 1) are met simultaneously, it is marked as corridor expansion and assigned an MLEI of 0.5. If the patch does not meet any of the morphological conditions in 1), it is marked as edge expansion and assigned an MLEI of 1.

[0213] The method of the present invention constructs the morphological landscape expansion index MLEI, whose mathematical expression is as follows:

[0214]

[0215] Where MLEI is the morphological landscape expansion index of the newly added urban patch, MLEI = 0 is infill expansion, MLEI = 0.5 is corridor expansion, MLEI = 1 is edge expansion; MLEI ≥ 2 is enclave expansion; 0 is a constant, indicating that the newly added patch is completely nested in the hole formed by the patch; N is the minimum number of expansion steps, and the value range is a positive integer with an interval of 1; The value range is 0 and 1. According to the definition of MLEI, the smaller the MLEI value, the closer the spatial relationship between the new urban patch and its adjacent existing urban patch, and the smaller the degree of expansion.

[0216] Figure 5 Schematic diagram of the results of applying the method of the present invention, wherein Figure 5 Middle (a), Figure 5 Middle (b), Figure 5 The middle (c) are all filling expansions; Figure 5 Middle (d) is marginal expansion, Figure 5 Middle (e) is enclave expansion, Figure 5 Middle (f) is corridor expansion.

[0217] Based on the above method of the present invention, the MLEI value corresponding to each newly added urban patch is output; the newly added patches can be classified into infill, corridor, edge and enclave types according to the MLEI value; and then spatial pattern statistical analysis and visualization are performed.

[0218] The above are preferred embodiments of the present invention. Any changes made according to the technical solution of the present invention, as long as the resulting functions and effects do not exceed the scope of the technical solution of the present invention, shall fall within the scope of protection of the present invention.

Claims

1. A method for identifying urban expansion patterns by integrating mathematical morphology and landscape connectivity, characterized by: include: Identify corridor expansion based on road data; Urban holes are identified through seed filling and morphological closing operations, and then the infill expansion inside the holes is identified; Based on the eight-neighborhood iterative expansion, the minimum expansion step is calculated to identify enclave expansion. The specific implementation method is as follows: S31. Identify candidate patches For the pixels in the newly added urban land image that are not marked as corridor expansion and infill expansion, a four-neighborhood connected component analysis is performed: S311, initializing the access tag matrix; S312, traverse all pixels in the newly added urban land image that are not marked as corridor expansion and infill expansion, use the four-neighborhood rule to identify and obtain candidate patches P new ; S32, Initialization S321, for each candidate patch P new , create a dilated result region X n , initially X0=P new ; S322, define the structural element S dil : Use 8-neighborhood structure elements; S323, set the iteration counter n=0; S324, set Target to the original urban land image binary raster data I old ; S33, Iterative Expansion S331, cycle starts: (1) Perform an expansion operation: Where P is P new The midpoint of a single independent patch, S dil,p The structural element S is translated with P as the center dil ; (2) Increasing: n=n+1; S332, intersection check (1) Check the expanded area X n Whether there is an intersection with Target, calculation formula: (2) Conditional judgment: If Intersect≠∅: stop expansion and record the minimum expansion steps N=n, that is, X after N steps of expansion n Contact with the original city; If Intersect=∅: return to step S331 to continue the expansion process; S34. Preliminary classification S341, record the minimum number of expansion steps N; S342, if N≥2, the corresponding candidate patch P new It is initially marked as enclave expansion and assigned the morphological landscape expansion index MLEI = N; if N = 1, the corresponding candidate patch P new Labeled as potential edge-type expansion or corridor-type expansion; The MLEI mathematical expression is as follows: Among them, MLEI = 0 is infill expansion; MLEI = 0.5 is corridor expansion; MLEI = 1 is edge expansion; MLEI ≥ 2 is enclave expansion; 0 is a constant, indicating that the new patch is completely nested in the hole formed by the patch; The value is 0 or 1. According to the definition of MLEI, the smaller the MLEI value, the closer the spatial relationship between the newly added urban patch and its adjacent existing urban patch, and the smaller the expansion degree. Integrate connectivity and linear morphology to identify corridor expansion and edge expansion.

2. The urban expansion pattern recognition method integrating mathematical morphology and landscape connectivity according to claim 1 is characterized in that: Corridor expansion is extracted based on road data and implemented as follows: Road data cleaning; Extract corridor expansion based on road data; Four-neighborhood connectivity iterative marking corridor expansion is adopted.

3. The urban expansion pattern recognition method integrating mathematical morphology and landscape connectivity according to claim 1 or 2, characterized in that: Extract corridor expansion based on road data. The specific implementation steps are as follows: S11. Road data cleaning: Calculate road length; Identify dead-end roads: Locate line segments whose ends are not connected to other roads through topological inspection; Eliminate dead-end roads with a length of less than 1 km; S12. Corridor expansion based on road data extraction: Road buffer generation: with the road processed in step S11 as the center, create a parallel buffer with the spatial resolution of the urban land image as the width; Extract pixels from the newly added urban land imagery that are completely or partially within the buffer zone; S13, using four-neighborhood connectivity iterative marking corridor expansion: Based on the pixels in the newly added urban land image extracted by S12, the four-neighborhood connectivity is used to iteratively extract the pixels in the newly added urban land image, and the number of iterations is S: S=2K / L K is the spatial resolution of urban land images; L is the maximum width of urban roads; All pixels in the extracted newly added urban land images are marked as corridor expansion and assigned a morphological landscape expansion index (MLEI) of 0.

5.

4. The urban expansion pattern recognition method integrating mathematical morphology and landscape connectivity according to claim 1 is characterized in that: Urban holes are identified through seed filling and morphological closing operations, and then the infill expansion within the holes is identified, which is achieved as follows: The four-neighborhood connectivity is used to scan the pixels belonging to non-urban areas in the original urban land image; Establish a mechanism to eliminate pseudo holes at the boundary contact point to filter out pseudo holes at the boundary; Use closing operation to form effective holes; Based on the pixels in the newly added urban land image that are not marked as corridor-type expansion, infill expansion located within valid holes is identified.

5. The urban expansion pattern recognition method integrating mathematical morphology and landscape connectivity according to claim 1 or 4, characterized in that: Urban holes are identified through seed filling and morphological closing operations, and then the in-fill expansion within the holes is identified. The specific implementation steps are as follows: S21. Generate background template Among them, (x, y) is the pixel coordinate; Represents the pixel value of the original urban land image, where the pixel value of the pixel belonging to the urban area is 1, and the pixel value of the pixel belonging to the non-urban area is 0; The pixel value representing the pixel in the background; S22. Identify hole candidate areas Use the four-neighborhood connectivity rule to traverse the background set I bg For all cells in , do the following: S221, initialize the marking matrix M, whose size is the same as I bg Same, the initial value is not visited; S222, Sequential Scan I bg For each pixel in, if the current pixel coordinate (x, y) satisfies I bg If (x, y) = 1 and M(x, y) is unvisited, the seed filling algorithm is executed: S2221. Create a hole candidate area set H=0; S2222, create a new hole candidate area H k ; S2223. Add (x, y) to queue Q; S2224. Loop when queue Q is not empty: (1) Take pixel coordinates from Q (P x ,P y ); (2) If (P x ,P y ) is within the image range and I bg (P x ,P y ) = 1 and it has not been visited, then mark it as visited; (3) (P x ,P y ) Add hole candidate area H k ; will (P x ,P y ) The four neighboring pixels are added to the queue Q; S2225, Save H k The set of candidate holes H={H1,H2,…,H k }; S23, filter out pseudo holes at the boundary S231, for each hole candidate region H obtained in step S22 k Perform boundary contact detection. If there is any pixel coordinate (x, y)∈H k If any of the following conditions is met, then H k For pseudo holes, remove H from H k ; (1) x = 1 or x = W, where W is the number of image columns; (2) y = 1 or y = H, where H is the number of image rows; S232: After the elimination operation in step S231, the remaining set is recorded as the effective hole set H. valid ={H valid,1 ,…,H valid,m }; S24, hole morphology optimization For each effective hole H valid,m Perform morphological closing: in represents the expansion operation, Indicates corrosion operation; Represented as a binary rectangle of preset size and geometry, used to control the optimization range of the hole boundary by dilation and erosion operations; S25, internal filling expansion marker For newly added urban land images that are not marked as corridor expansion, infill expansion is determined: For the newly added urban land image binary raster data I that is not marked as corridor expansion new For each pixel in , if it satisfies: I new (x,y)⊆H valid,m , then mark I new The corresponding pixel in the image is expanded by filling in, and the morphological landscape expansion index MLEI=0 is assigned. new (x, y) is the binary raster data of the newly added urban land image that is not marked as corridor expansion. new The pixel value of the pixel in .

6. The urban expansion pattern recognition method integrating mathematical morphology and landscape connectivity according to claim 1 is characterized in that: The minimum expansion steps are calculated based on the eight-neighborhood iterative expansion. That is, the eight-neighborhood rule is used to set the expansion direction, and the minimum expansion steps N is defined as the number of expansion steps when the expansion buffer first intersects with the original urban patch.

7. The urban expansion pattern recognition method integrating mathematical morphology and landscape connectivity according to claim 1 is characterized in that: Connectivity and linear morphology are combined to identify corridor expansion and edge expansion. This is achieved as follows: S41, to I old Perform four-neighborhood connected component analysis to determine the independent patch C of the original urban land image old ; S42. For all candidate patches P with N=1 new , perform the following operations: S421, perform four-neighborhood connected component analysis, statistics and C old The number of connected components C; S422, if C≠2, the corresponding candidate patch P new Marked as edge expansion, the morphological landscape expansion index MLEI = 1; if C = 2, calculate the corresponding candidate patch P new Two independent patches C of the original urban land images connected old Their respective areas and corresponding candidate patches P new The perimeter area ratio PAR and eccentricity ECC, if the two C old The area is larger than 0.5ha, PAR>4.3 and ECC>0.83, and the corresponding candidate patches P new Marked as corridor expansion, the landscape expansion index MLEI = 0.5 is assigned. Otherwise, the corresponding candidate patch P new It is marked as marginal expansion and assigned the morphological landscape expansion index MLEI = 1.

8. The urban expansion pattern recognition method integrating mathematical morphology and landscape connectivity according to claim 7 is characterized in that: The perimeter area ratio PAR is calculated as follows: Perimeter Pnew is the candidate patch P new Perimeter, Area Pnew is the candidate patch P new area.

9. The urban expansion pattern recognition method integrating mathematical morphology and landscape connectivity according to claim 7 is characterized in that: The eccentricity ECC is calculated as follows: Where Length major is the candidate patch P new Lengthminor is the length of the major axis of the minimum bounding rectangle MBR, and Lengthminor is the length of the minor axis of MBR.

Images