A method and device for typifying a group of combined straight-line buildings

Abstract

The present invention discloses a method and device for typifying a combined linear pattern building complex. The method includes: Step 1, randomly select an ungeneralized combined linear pattern building complex CP, perform a merging operation on the buildings with spatial conflicts within CP, that is, the pattern units, and then perform Step 2; Step 2, calculate the minimum on-map distance d between the pattern units within CP ex , and judge the size of d ex and the distance constraint threshold δ dis . If d ex < δ dis , then perform Step 3; if d ex ≥ δ dis , then perform Step 6; Step 3, perform a deletion operation on CP, and then perform Step 4; Step 4, perform an exaggeration operation on CP, and then perform Step 5; Step 5, perform a shifting operation on CP, and then perform Step 2; Step 6, perform a splitting operation on CP, and judge whether there is an ungeneralized combined linear pattern building complex. If so, go back to Step 1; if not, the typification ends. The present invention realizes the generalization of building complexes taking into account the combined linear pattern.

Description

Technical Field

[0001] The invention belongs to the technical field of comprehensive urban building complex maps, and in particular relates to a method and device for typifying a combined straight line model building complex. Background Art

[0002] Typification of building complexes refers to compressing data details while maintaining the location accuracy, distribution pattern and spatial structure of the building complex as much as possible. Therefore, typification, as the focus and difficulty in comprehensive research on buildings, is often used in the comprehensive mapping of building complexes with pattern distribution.

[0003] However, existing research on the typification of building complexes based on spatial distribution pattern characteristics mainly considers the synthesis of pattern building complexes with single buildings as pattern units, but there is little research on the synthesis of combined linear pattern building complexes with combined buildings as pattern units. Summary of the invention

[0004] The present invention aims at the existing research on the typification of building complexes supported by spatial distribution pattern characteristics, which mainly considers the synthesis of pattern building complexes with single buildings as pattern units, while there is little research on the synthesis of combined straight line pattern building complexes with combined buildings as pattern units. A method and device for typifying combined straight line pattern building complexes are proposed, so that the integrated building complex can maintain the distribution and structural characteristics of the combined straight line pattern as much as possible.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] In one aspect, the present invention provides a method for typifying a combined straight line model building complex, comprising:

[0007] Step 1, randomly select an unintegrated combined straight line pattern building complex CP, perform a merge operation on the buildings, i.e., pattern units, that have spatial conflicts within the CP, and then perform step 2;

[0008] Step 2: Calculate the minimum spacing d between the mode units in the CP ex , judge d ex and the distance constraint threshold δ dis The size of d ex <δ dis , then execute step 3; if d ex ≥δ dis , then execute step 6;

[0009] Step 3, perform a delete operation on the CP, and then perform step 4;

[0010] Step 4, perform an exaggeration operation on CP, and then perform step 5;

[0011] Step 5, perform a shift operation on CP, and then execute Step 2;

[0012] Step 6, perform a splitting operation on CP, and determine whether there is an un-synthesized combined straight-line pattern building complex. If so, go to Step 1; if not, the typification ends.

[0013] Further, the merging operation includes:

[0014] Represent all buildings using their minimum-area circumscribed rectangles SMBR, and use the SMBR of the building set within the pattern unit as the merged building;

[0015] Establish a data structure Unit_Buildings that describes the spatial relationship and structure between the merged building and the original buildings; in the data structure Unit_Buildings, the pattern unit serial number, pattern unit elements, pattern unit minor axis, pattern unit major axis, building ID included in the pattern unit, and building anchor point are defined.

[0016] Further, the deletion operation shall follow the following rules:

[0017] The first and last pattern units of the combined straight-line pattern cannot be deleted;

[0018] Delete the pattern unit with the least impact; including: preferentially delete the pattern unit with the smallest sum of building areas; preferentially retain the pattern unit composed of combined buildings, and the higher the priority of retention for the pattern unit with a larger internal building spacing distance;

[0019] The deletion operation includes the following steps:

[0020] Step 3.1, traverse each pattern unit Unit in the combined straight-line pattern Pattern i , 1 < i < Num, where Num is the number of pattern units in Pattern, calculate the maximum value Dis of the distances between adjacent buildings max , and add the array [i, Dis max to the linked list List_Distance. If Unit i contains only one building, calculate the sum of building areas Area sum , and add the array [i, Area sum to the linked list List_Area; execute Step 3.2;

[0021] Step 3.2, if the linked list List_Area is not empty, then traverse and find the array [m, Area min], then the mth pattern unit needs to be deleted, and the deletion operation ends; if the linked list List_Area is empty, execute step 3.3;

[0022] Step 3.3, traverse and find the array [n, Dis min ], the nth pattern unit needs to be deleted and the deletion operation ends.

[0023] Further, the exaggeration operation includes:

[0024] Step 4.1, determine the exaggerated area of ​​each pattern unit; including: using an average distribution strategy to distribute the area of ​​the deleted pattern unit, and record the area of ​​the deleted pattern unit as Area del , then the remaining area allocated to each mode unit is Area ave The calculation formula is as follows:

[0025]

[0026] Where Num is the number of remaining pattern units;

[0027] Step 4.2, exaggeration operation of the pattern unit; including: the building is exaggerated along its major axis or minor axis; when exaggerating, the angles between the major axis, the minor axis and the main direction of the pattern are first calculated, and the axis direction with the smaller angle is the exaggeration direction; the rectangular sides of the building consistent with the exaggeration direction are extended outward from both ends respectively, and the extended vertices form the exaggerated building, thereby realizing the exaggeration of the building along the pattern direction.

[0028] Furthermore, the shift operation must follow the following rules:

[0029] Keep the spatial coverage of the pattern unchanged, that is, only the pattern units except the first and last ends can be shifted;

[0030] The shifted pattern unit tries to keep the trajectory consistent with the original combined straight line pattern;

[0031] The shift operation includes the following steps:

[0032] By making equal-interval differences on the lines connecting the original centroids of all pattern units, the coordinates of the shifted points of the centroids of each pattern unit can be obtained, and then the shift operation is completed.

[0033] Furthermore, the segmentation operation includes:

[0034] Step 6.1, update the pattern unit structure data: for the pattern unit elements, pattern unit short axis and pattern unit long axis in Unit_Buildings, update them according to the SMBR of the corresponding pattern unit; keep the ratio of the distance value between each anchor point and the pattern unit long axis length value equal before and after the update;

[0035] Step 6.2, reconstruction of buildings within the model unit: judge the distance between adjacent anchor points of adjacent buildings. If the distance is less than the distance constraint threshold δ dis The adjacent anchor points will be deleted.

[0036] Another aspect of the present invention provides a device for typifying a group of buildings in a combined straight line model, comprising:

[0037] The merging module is used to select any unintegrated combined straight line pattern building complex CP, perform a merging operation on the buildings, i.e., pattern units, that have spatial conflicts within the CP, and then execute the judgment module;

[0038] The judgment module is used to calculate the minimum spacing d between the pattern units in the CP ex , judge d ex and the distance constraint threshold δ dis The size of d ex <δ dis , then execute the deletion operation module; if d ex ≥δ dis , then execute the segmentation operation module;

[0039] A deletion operation module, used to perform a deletion operation on the CP and then execute the exaggeration operation module;

[0040] an exaggeration operation module, used for performing an exaggeration operation on the CP and then performing a shift operation module;

[0041] A shift operation module, used for performing a shift operation on the CP and then executing the judgment module;

[0042] The segmentation operation module is used to perform segmentation operation on CP to determine whether there is an unintegrated combined straight line pattern building complex. If so, it will go to the merging module, otherwise the typicalization ends.

[0043] Furthermore, the merging operation includes:

[0044] All buildings are represented by their minimum area circumscribed rectangle SMBR, and the SMBR of the building set in the model unit is used as the merged building;

[0045] Create a data structure Unit_Buildings that describes the spatial relationship and structure between the merged building and the original building; in the data structure Unit_Buildings, the serial number of the pattern unit, the elements of the pattern unit, the minor axis of the pattern unit, the major axis of the pattern unit, the building ID included in the pattern unit, and the building anchor point are defined.

[0046] Furthermore, the deletion operation shall follow the following rules:

[0047] The first and last pattern units of the combined straight-line pattern cannot be deleted;

[0048] Delete the pattern unit with the least impact; including: preferentially delete the pattern unit with the smallest sum of building areas; preferentially retain the pattern unit composed of combined buildings, and the higher the priority of retention for the pattern unit with a greater internal building spacing distance;

[0049] The deletion operation includes the following steps:

[0050] Step 3.1, traverse each pattern unit Unit in the combined straight-line pattern Pattern i , 1 < i < Num, where Num is the number of pattern units in Pattern, calculate the maximum value Dis of the distances between adjacent buildings max , and add the array [i, Dis max to the linked list List_Distance. If Unit i contains only one building, then calculate the sum of building areas Area sum , and add the array [i, Area sum to the linked list List_Area; execute Step 3.2;

[0051] Step 3.2, if the linked list List_Area is not empty, then traverse to find the array [m, Area min in List_Area that contains the minimum area, then the m-th pattern unit needs to be deleted, and the deletion operation ends; if the linked list List_Area is empty, then execute Step 3.3;

[0052] Step 3.3, traverse to find the array [n, Dis min in List_Distance that contains the minimum distance, then the n-th pattern unit needs to be deleted, and the deletion operation ends.

[0053] Furthermore, the exaggeration operation includes:

[0054] Step 4.1, determine the exaggerated area of each pattern unit; including: adopt an equal distribution strategy to distribute the area of the pattern unit after deletion. Denote the area of the deleted pattern unit as Area del, then the remaining area allocated to each mode unit is Area ave The calculation formula is as follows:

[0055]

[0056] Where Num is the number of remaining pattern units;

[0057] Step 4.2, exaggeration operation of the pattern unit; including: exaggeration of the building along its major axis or minor axis; when exaggerating, firstly calculating the angle between the major axis, the minor axis and the main direction of the pattern, and the axis direction with the smaller angle is the exaggeration direction; extending the rectangular sides of the building consistent with the exaggeration direction outward from both ends respectively, and forming the exaggerated building with the extended vertices, so as to realize the exaggeration of the building along the pattern direction;

[0058] The shift operation must follow the following rules:

[0059] Keep the spatial coverage of the pattern unchanged, that is, only the pattern units except the first and last ends can be shifted;

[0060] The shifted pattern unit tries to keep the trajectory consistent with the original combined straight line pattern;

[0061] The shift operation includes the following steps:

[0062] By making equal-interval differences on the lines connecting the original centroids of all pattern units, the coordinates of the shifted points of the centroids of each pattern unit can be obtained, and then the shift operation is completed;

[0063] The segmentation operation includes:

[0064] Step 6.1, update the pattern unit structure data: for the pattern unit elements, pattern unit short axis and pattern unit long axis in Unit_Buildings, update them according to the SMBR of the corresponding pattern unit; keep the ratio of the distance value between each anchor point and the pattern unit long axis length value equal before and after the update;

[0065] Step 6.2, reconstruction of buildings within the model unit: judge the distance between adjacent anchor points of adjacent buildings. If the distance is less than the distance constraint threshold δ dis The adjacent anchor points will be deleted.

[0066] Compared with the prior art, the present invention has the following beneficial effects:

[0067] (1) Combining the distribution pattern and spatial structure characteristics of the combined straight line pattern building complex, the present invention adopts the idea of ​​decomposing complex operators into simple operator combinations, and decomposes the typical operation of the combined straight line pattern building complex into a "gradual and iterative" process composed of five basic operators: merging, deleting, exaggerating, shifting and splitting. This provides a feasible new case in terms of operator collaboration and algorithm collaboration.

[0068] (2) The synthesis of building complexes that take into account the combined straight line pattern is achieved. The typicalization method proposed in the present invention can fully maintain the overall straight line pattern distribution characteristics of the building complex, and its spatial coverage does not change during the synthesis process. The method also takes into account the local spatial heterogeneity characteristics of the building complex. Since local heterogeneity belongs to low-level, local detail information, when the comprehensive scale is large, such details are no longer the main factor in distinguishing building complexes. In addition, the method can effectively implement typicalization operations on combined straight line pattern building complexes with different characteristics. In general, the proposed typicalization method can effectively and reasonably synthesize combined straight line pattern building complexes. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 This is an example diagram of a combined straight line mode of a building complex according to an embodiment of the present invention;

[0070] Figure 2 A basic flow chart of a typical method for a combined straight line model building complex according to an embodiment of the present invention;

[0071] Figure 3 This is an example diagram of the merging of buildings within a unit in an embodiment mode of the present invention;

[0072] Figure 4 Delete the example diagram for the mode unit of the embodiment of the present invention;

[0073] Figure 5 This is an exaggerated example diagram of a building according to an embodiment of the present invention;

[0074] Figure 6 This is an example diagram of updating structure data according to an embodiment of the present invention;

[0075] Figure 7 It is an example diagram of a straight line pattern of a building complex combination with different characteristics of an embodiment of the present invention;

[0076] Figure 8 This is an example diagram of comprehensive results of different combinations of straight line model building complexes according to an embodiment of the present invention;

[0077] Fig. 9 This is an example diagram of the original combined straight line model building complex and its comprehensive results according to an embodiment of the present invention;

[0078] Fig.10 This is a comparison diagram of the comprehensive results of combined straight line mode B according to an embodiment of the present invention;

[0079] Fig.11 This is a schematic diagram of the structure of a typical device for a combined straight line model building complex according to an embodiment of the present invention. DETAILED DESCRIPTION

[0080] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments:

[0081] First, let’s introduce the straight line pattern of building complex:

[0082] In large-scale maps, the typical linear pattern of building groups is a regular spatial arrangement and distribution of building groups at the macro level, while at the micro level it shows the similarity of buildings in spatial and geometric characteristics. However, the cognitive characteristics of the typical linear pattern are not fully applicable to phenomena that are locally heterogeneous but distributed in a linear pattern as a whole, such as Figure 1 In (a), although some buildings within the rectangular frame have obvious local feature differences, they can still be identified as a straight line pattern as a whole.

[0083] According to the Gestalt cognitive principle and visual perception topology theory, the phenomenon of buildings with obvious local heterogeneity and overall straight line distribution is defined as the concept of combined straight line pattern. Combined straight line pattern has the characteristics of multi-level cognition, such as Figure 1 As shown in the figure, at the macroscopic overall level, the building complex is distributed in a straight line pattern; at the mesoscopic pattern unit level, the straight line pattern is composed of multiple pattern units with similar characteristics, such as Figure 1 In (b), the rectangular frame that approximates the building is used as the pattern unit that composes the straight line pattern; at the microscopic pattern element layer, the pattern unit is a single building element or a combination of multiple buildings arranged in an extended and aligned manner.

[0084] The main task of building complex synthesis is to maintain the spatial distribution pattern, structure and other characteristics while compressing the details of the building complex. Especially in large-scale maps, it is necessary to pay attention to the implicit structured knowledge of the building complex. The synthesis of building complexes taking into account the spatial distribution pattern needs to solve two problems. The first is to obtain structured and characterized knowledge, such as determining which building complexes have interesting distribution characteristics through pattern recognition and determining their distribution characteristic types, that is, the characteristic information that needs to be maintained during synthesis. The second is to select appropriate operators and algorithms based on structured and characterized knowledge, and design reasonable and effective synthesis rules and strategies.

[0085] Generally, when a building group with a single building as a pattern unit forms a linear pattern distribution, map synthesis is required when its internal space is not enough to ensure readability and distinguishability at a larger scale. If a displacement synthesis strategy is adopted, the building can only be moved outward due to the limited internal space of the group, which will cause a loss of positioning accuracy and may also conflict with other buildings or spatial targets, and further lead to more secondary conflicts; if an overall merger is performed, the map synthesis degree will be too large, and information such as spatial distribution characteristics will be lost; while typicalization is a comprehensive operator that can take into account the spatial structure and distribution characteristics of the target group. Therefore, typicalization is often used for the synthesis of spatial targets with pattern distribution. Different from the typical linear pattern, the pattern unit in the combined linear pattern is usually a combined structure of multiple buildings. From the perspective of "linear pattern-pattern unit", if the space inside the pattern is not enough to clearly express all the pattern units, it is necessary to use the pattern unit as the processing object to implement the same typicalization operation as the typical linear pattern. From the perspective of "pattern unit-building", when a pattern unit contains multiple buildings and the internal space of the pattern unit is not sufficient to clearly express all the buildings, the pattern unit needs to be integrated. Since the buildings within the pattern unit are usually regarded as a whole to form a combined linear pattern, a merging strategy can be adopted for the buildings within the pattern unit.

[0086] In combination with the distribution pattern and spatial structural characteristics of the combined straight line model building complex, the present invention adopts the idea of ​​decomposing complex operators into simple operator combinations, and decomposes the typical operation of the combined straight line model building complex into a "progressive and iterative" process composed of five basic operators: merging, deleting, exaggerating, shifting, and splitting. Here, merging is to merge the buildings within the model unit to solve the spatial conflict between the buildings within the model unit; deleting is to reduce the number of model units to reserve extra space to express the entire building complex; exaggerating is to restore the overall spatial characteristics of the combined straight line building complex, such as area and density; shifting is to solve the spatial conflict between the model units and reconstruct the distribution pattern of the building complex; and splitting is to maintain the structural characteristics within the model unit.

[0087] For building complexes, no matter which comprehensive operators are used or how the comprehensive algorithms are designed, a good comprehensive result needs to maintain the legibility of buildings and spatial structures, the visual distinguishability of groups, and the harmony between the spatial structures of groups. This is the general guiding ideology of building complex synthesis. By summarizing the comprehensive knowledge in the existing literature on building complex synthesis, the principles that the operators and algorithms used in the synthesis of combined linear model building complexes should follow can be obtained, as shown in Table 1.

[0088] Table 1 Comprehensive rules for combined straight line pattern building complexes

[0089]

[0090]

[0091] A reasonable and effective synthesis process, in addition to selecting and designing operators and algorithms for different types of spatial conflicts, also needs to formulate reasonable synthesis strategies to coordinate the cooperation between operators and algorithms to complete the synthesis task. For the synthesis of combined straight line pattern building complexes, the conflicts between pattern units are eliminated by iteratively executing deletion, exaggeration and shifting. The spatial conflicts between buildings in the pattern units are eliminated by successively executing merging and splitting operations. The relationship between the distance on the map between the pattern units and the minimum spacing distance on the map is used as the control condition for the synthesis operation. In summary, the process of a method for typifying a combined straight line pattern building complex of the present invention is as follows. Figure 2 As shown, including:

[0092] Step 1, randomly select an unintegrated combined straight line pattern building complex CP, perform a merge operation on the buildings, i.e., pattern units, that have spatial conflicts within the CP, and then perform step 2;

[0093] Step 2: Calculate the minimum spacing d between the mode units in the CP ex , judge d ex and the distance constraint threshold δ dis The size of d ex <δ dis , then execute step 3; if d ex ≥δ dis , then execute step 6;

[0094] Step 3, perform a delete operation on the CP, and then perform step 4;

[0095] Step 4, perform an exaggeration operation on CP, and then perform step 5;

[0096] Step 5, perform a shift operation on CP, and then perform step 2;

[0097] Step 6, perform segmentation operation on CP to determine whether there is an unintegrated combined straight line pattern building complex. If so, go to step 1, otherwise the typicalization ends.

[0098] Furthermore, the merging operation of buildings within the model unit includes:

[0099] In the combined straight line pattern, there are pattern units composed of multiple buildings. When the internal space is not enough to clearly express all the buildings, it is necessary to merge the buildings with spatial conflicts. The primary task of typicalizing the pattern building group is usually to reduce the number of pattern units, and the deletion operation should be performed first. The present invention takes merging as the most prioritized operator mainly for two reasons.

[0100] (1) When performing the exaggeration operation, if the model unit contains multiple buildings and the main direction of the buildings is close to the main direction of the model, whether the buildings are exaggerated along the main direction to one side of the building or to both sides of the center of the building, the gap between the buildings in the model unit will be reduced. In this way, buildings that should not be merged may meet the merging conditions, resulting in the loss of structural information of the model unit. Therefore, in order to maintain the structural characteristics of the model unit, the present invention adopts a strategy of first merging all the buildings in the model unit, then exaggerating the model unit as a whole, and finally dividing the model unit.

[0101] (2) When performing the deletion operation, the characteristics of the buildings in each model unit before merging are mainly considered. Therefore, the execution order of the merge operation and the delete operation does not affect the final typicalization result. Prioritizing the merge operation can simplify the process of subsequent comprehensive operations, such as reducing the calculation of the distance between buildings in the model unit.

[0102] Considering that the combined straight line model is mainly for buildings with regular shapes and high rectangularity, and the buildings in the model unit are extensions of the main directions of each other, here, all buildings are represented by their minimum area circumscribed rectangle (SMBR), and the SMBR of the building set in the model unit is used as the merged building, as shown in Figure 3 The dotted box U is the building after the buildings B1, B2 and B3 are merged. All buildings mentioned below are expressed as SMBR and will not be explained in detail.

[0103] To facilitate the implementation of subsequent operators, a data structure describing the spatial relationship and structure between the merged building (i.e., the pattern unit) and the original building was established based on object-oriented programming ideas, as shown in the Unit_Buildings structure below.

[0104]

[0105] The present invention records the projection point from the beginning and end points of the long axis of the building to the long axis of the merged building as the building anchor point. The building anchor point list AnchorPTS in the structure Unit_Buildings refers to a list set of anchor points of each building organized in sequence. Figure 3 In, L U The auxiliary line after the long axis of the merged building is translated. Points s1 and e1 are the starting and ending points of the long axis L1 of building B1 to L U The projection point of U is the building anchor point list in the pattern unit structure corresponding to U is [s1,e1,s2,e2,s3,e3].

[0106] Furthermore, the deletion operation of the least impact mode unit includes:

[0107] The delete operation is the first operator executed after the merge. Due to the gradual and iterative strategy, only one pattern unit is deleted each time. According to the comprehensive rules in Table 1, the delete operation needs to follow the following two aspects:

[0108] (1) The first and last pattern units of the combined straight line pattern cannot be deleted. The pattern units at both ends of the pattern are the boundary marks of the entire pattern space coverage. Deleting them will change the spatial coverage of the pattern, resulting in the loss of positioning accuracy and distribution characteristics.

[0109] (2) Delete the pattern unit with the least impact. The optimal purpose of the deletion operation is to delete the pattern unit with the least impact on the characteristics of the combined straight line pattern, and provide sufficient expression space for other pattern units, that is, to maintain the spatial distribution characteristics of the original extended straight line pattern to the greatest extent. In addition, the deletion operation is limited by the number of pattern units. Generally, a pattern can only be formed when the number of pattern units is greater than two. The present invention believes that two pattern units can still reflect the distribution characteristics of the original building complex to a certain extent. Therefore, when performing the deletion operation, it is necessary to ensure that the number of pattern units is not less than two.

[0110] From a holistic perspective, the comprehensive mapping of pattern building complexes focuses on the overall shape, spatial distribution and spatial coverage of the building complex. The deletion of pattern units has limited impact on the perception of the pattern from characteristics such as direction and shape, while the deleted area will directly cause obvious perceptual differences within the spatial coverage of the pattern. Therefore, when deleting pattern units, it is necessary to focus on the impact of the area on the overall pattern building complex. From a local perspective, there are pattern units composed of multiple buildings in the combined straight line pattern, and the structures of each pattern unit are different. Therefore, the combined straight line pattern has certain heterogeneous characteristics. Although the perception level of this local heterogeneity is lower than the overall perception of the pattern, it is still an important feature of distinguishability at a certain scale. The visual characteristics of this local heterogeneity are mainly reflected in the size of the gaps between adjacent buildings within the pattern unit, such as Figure 4 In (a), when the gap between buildings in the pattern unit is larger, the difference in area between the building and the pattern unit is larger, and the visual difference is more obvious. The pattern unit composed of buildings that are closer to each other is more easily perceived as a whole visually, such as Figure 4 (a) Pattern unit 5. The pattern units composed of buildings that are farther away are visually distinguishable, e.g. Figure 4 (a) Pattern unit 2. Therefore, the deletion of pattern units also needs to consider the local distinguishability of the pattern. Here, the nearest anchor point (such as Figure 2The distance between the anchor points e1 and s2) in it is used as the building interval distance to measure the distinguishability. The greater the distance, the stronger the distinguishability. From the above analysis, it can be seen that there are mainly two rules for deleting the pattern unit with the least influence:

[0111] (1) Give priority to deleting the pattern unit with the smallest sum of building areas;

[0112] (2) Give priority to retaining the pattern unit composed of combined buildings, and the higher the priority of retaining the pattern unit with the larger internal building interval distance.

[0113] Since the larger the distance between buildings, the smaller the building area of the pattern unit, this causes a conflict between the rule of giving priority to deleting the pattern unit with the smallest area and the rule of giving priority to retaining the pattern unit with the largest building spacing. Considering that local heterogeneity and distinguishability are important features of the combined straight-line pattern, the present invention stipulates that the binding force of the area constraint rule is weaker than that of the spacing constraint rule.

[0114] As Figure 4 shown, Figure 4 (a) is the original combined straight-line pattern. The sorting of the pattern unit areas is: #1 = #4 > #5 > #3 = #2, and the sorting of the maximum internal interval distances is: #2 > #3 > #5 > #1 = #4 (the interval distance of the pattern unit with only one building is counted as 0). As Figure 4 (b) The wrong scheme deletes the pattern unit at the head of the combined straight-line pattern, resulting in a loss of the positioning accuracy and distribution characteristics of the pattern. As Figure 4 (c) The wrong scheme deletes the pattern unit with the largest interval distance, resulting in a loss of the local heterogeneity characteristics of the combined straight-line pattern. As Figure 4 (d) The wrong scheme in (d) does not delete the pattern unit with the largest interval distance, but the local heterogeneity characteristic belongs to the feature to be preferentially retained, so it also causes a certain degree of loss of the local heterogeneity characteristic of the combined straight-line pattern. In the correct scheme as Figure 4 (e), although the pattern unit with the largest area located inside the pattern is deleted, the straight-line distribution pattern and local spatial heterogeneity characteristics of the combined straight-line pattern are effectively maintained.

[0115] Based on the above analysis, the steps for determining the pattern unit to be deleted in each deletion operation are as follows:

[0116] Step 3.1, traverse each pattern unit Unit in the combined straight-line pattern Pattern i (1 < i < Num, Num is the number of pattern units in Pattern), calculate the maximum value Dis of the distances between adjacent buildings max , and the array [i, Dis max]Add to the linked list List_Distance, if Unit i If there is only one building in the sum , and the array [i,Area sum ]Add to the linked list List_Area; execute step 3.2;

[0117] Step 3.2, if the linked list List_Area is not empty, then traverse to find the array [m, Area in List_Area that contains the minimum area value min ], the mth pattern unit needs to be deleted and the program ends; if the linked list List_Area is empty, execute step 3.3;

[0118] Step 3.3, traverse and find the array [n, Dis min ], the nth pattern unit needs to be deleted and the program ends.

[0119] Furthermore, the exaggerated operations taking into account the positioning accuracy include:

[0120] Since the overall area of ​​the building complex before and after integration should remain similar, after the deletion operation is performed, the remaining pattern units need to be appropriately exaggerated, and the degree of exaggeration is mainly determined by the area of ​​the deleted pattern unit. Combined with the rules of the exaggeration operation in Table 1, the exaggeration operation is mainly completed through the following process.

[0121] Step 4.1, determine the exaggerated area of ​​each mode unit.

[0122] Here, the average distribution strategy is used to distribute the area of ​​the deleted pattern unit. The area of ​​the deleted pattern unit is denoted as Area del , then the remaining area allocated to each mode unit is Area ave The calculation formula is as follows:

[0123]

[0124] Where Num is the number of remaining pattern units.

[0125] Step 4.2, exaggeration operation of the pattern unit.

[0126] The exaggeration of the pattern unit is also the exaggeration of the merged building. The exaggeration needs to be done along the main direction of the pattern, because if the exaggeration is done in a direction perpendicular to the main direction of the pattern, the spatial coverage of the pattern will become larger. Considering that the major axis or minor axis of the building is usually consistent with the main direction of the pattern, the exaggeration of the building is done along its major axis or minor axis. When exaggerating, first calculate the angle between the major axis, minor axis and the main direction of the pattern, and the axis direction with the smaller angle is the exaggeration direction. Extend the rectangular sides of the building consistent with the exaggeration direction outward from both ends, and use the extended vertices to form the exaggerated building, thereby achieving the exaggeration of the building along the pattern direction. For example Figure 5 (a) Midpoint A E , B E Point C is the endpoint of the extended side AB. E , D E is the endpoint of the extended edge CD, then Figure 5 Rectangle A in (b) E B E C E D E An exaggerated building.

[0127] Combination Figure 5 As shown, for building ABCD, it is known that the exaggeration direction is the direction of side AB, and the coordinates of the vertices of building ABCD are (x A ,y A )、(x B ,y B )、(x C ,y C ) and (x D ,y D ), and the exaggerated area Area determined by formula (1) ave , then the exaggeration operation algorithm flow of building ABCD is as follows:

[0128] (1) Calculate the exaggerated length of the building. Since the building is a regular rectangle at this time and is only exaggerated along the exaggeration direction (sides AB and CD), the length of the sides perpendicular to the exaggeration direction (sides AD and BC) remains unchanged. Sides AB and CD are recorded as exaggerated sides. The length Len of side AD or BC (sides AD and BC are equal) can be calculated from the vertex coordinates. The calculation formula for the length Len_Ex of the exaggerated side extending outward from the two end points is as follows.

[0129]

[0130] (2) Calculate the exaggerated endpoint. Take the example of extending the edge AB from the endpoint A. Let the coordinates of the extended endpoint E be (x E ,y E ), it is easy to know that the distance between point A and E is Len_Ex, then we can get

[0131]

[0132] From the analysis, we know that points A, B and E are all on the same straight line. By using equation (3), we can get the coordinate expression of point E:

[0133]

[0134] From formula (4), we can see that the coordinates of point E (x E ,y E ) has two values, such as Figure 5 Point A in (a) E and A I They are all on the straight line AB and their distance from point A is Len_Ex. Point A E and A I They correspond to the two values ​​in equation (4), so we need to further determine the coordinates of point E. Since point E is the point where edge AB extends outward from endpoint A, the distance between point E and point B must be greater than the distance between point A and point B. Based on this, we can determine the coordinates of point E, as follows: Figure 5 Point A in (a) E It is the point after the endpoint A of side AB is extended.

[0135] For the first and last pattern units, since the pattern boundary must not change, the first and last pattern units need to be exaggerated along the main direction of the pattern and toward the pattern center. The implementation idea is to determine the distance between the endpoints on both sides of the exaggerated edge and the pattern center point, and only exaggerate the side with the smallest distance to the endpoint, and the exaggerated length is 2*Len_Ex.

[0136] (3) Building reconstruction: Replace the vertices in the building with the exaggerated endpoints in order to complete the exaggeration operation of the pattern unit.

[0137] Further, the shift operation of maintaining the distribution mode characteristics includes:

[0138] Deletion and exaggeration operations will cause the spacing between pattern units to be different, resulting in the loss of the spatial distribution pattern of the building complex. Therefore, it is necessary to shift the pattern units to restore the equally spaced distribution characteristics of the straight line pattern. The shift of pattern units needs to take into account two principles:

[0139] (1) Keep the spatial coverage of the pattern unchanged, that is, only the pattern units except the first and last ones can be shifted;

[0140] (2) The shifted pattern unit should try to keep the trajectory consistent with the original combined straight line pattern.

[0141] Therefore, by performing equal-interval differences on the lines connecting the original centroids of all pattern units, the coordinates of the shifted points of the centroids of each pattern unit can be obtained, and then the shift operation can be completed.

[0142] Furthermore, the segmentation operation that maintains local heterogeneity includes:

[0143] Since the merging operation adopts the strategy of merging the buildings in the pattern unit as a whole, this results in the loss of the structural characteristics of the pattern unit, especially the local heterogeneity characteristics of the combined straight line pattern that are different from the typical straight line pattern. Therefore, it is necessary to segment the pattern unit so that it can restore the structural characteristics before the integration as much as possible. In combination with the characteristics of high regularity and extended alignment distribution of buildings in the combined straight line pattern unit, the pattern unit can be segmented in the direction orthogonal to the main direction of the pattern unit based on the building anchor points in the structural data corresponding to the pattern unit. The segmentation of the pattern unit can be achieved by distinguishing the building polygons and the interval area polygons according to the order of the anchor points.

[0144] Based on the above analysis, the segmentation operation is mainly completed through the following process.

[0145] Step 6.1, the pattern unit structure data is updated.

[0146] After the merge operation, the deletion, exaggeration and shift operations were performed on the pattern unit as the object, but the data of the structure Unit_Buildings corresponding to the pattern unit was not updated accordingly. Unit_Index and BuildingFIDList in Unit_Buildings are fixed attributes and do not need to be updated. Unit_Polygon, Unit_SMBRL and Unit_SMBRW are updated according to the SMBR of the corresponding pattern unit. The building anchor point can reflect the spatial position relationship between the building and the pattern unit through its position relationship with the long axis of the pattern unit. In order to maintain a similar position relationship between each anchor point and the long axis of the pattern unit before and after the update, it is necessary to keep the ratio of the distance value between each anchor point and the long axis length value of the pattern unit equal, such as Figure 6 In, s1e1 and L before Length ratio and s1'e1' and L update The length ratio is equal. Combining the combined straight line pattern features, it can be seen that the first and last anchor points in AnchorPTS coincide with the endpoints of the long axis of the pattern unit. Therefore, the endpoints of the long axis of the updated pattern unit can be recorded as the first and last anchor points P0' and P n '. Then, according to the other anchor points P in AnchorPTS i The ratio of the distance from the anchor point P0 to the length of the major axis of the original pattern unit is used to calculate the updated anchor point P iThe distance between ' and P0' can be calculated by combining formula (4) to get the anchor point P i 'Specific coordinate value.

[0147] Step 6.2, reconstruction of buildings within the model unit.

[0148] Although the segmentation operation needs to restore the structural features of the pattern unit as much as possible, considering that the adjacent buildings still need to be merged if they are too close after reconstruction, the distance between the adjacent anchor points of the adjacent buildings is judged here, and if the distance is too small, the adjacent anchor points are deleted. Let the target scale be Scale, and the minimum distance threshold on the graph be δ dis , then the specific steps of building reconstruction are as follows:

[0149] Step 6.2.1: Take the anchor point P in AnchorPTS i and P i+1 Building B i The first and last anchor points of B i Add to B_List, set i = 1, 3, ..., count-1 (count is the length of AnchorPTS); execute step 6.2.2;

[0150] Step 6.2.2: Take any set of adjacent buildings B in B_List i and B i+1 ; Calculate B i The end anchor point of B i+1 The distance d on the graph of the first anchor point in , if d in ≤δ dis , then B i The end anchor point is replaced by B i+1 The end anchor point of the B_List is deleted. i+1 , execute step 6.2.2; if d in >δ dis , repeat step 6.2.2 until all adjacent buildings are traversed, and then execute step 6.2.3;

[0151] Step 6.2.3: If the number of buildings in B_List is 1, there is no need to reconstruct the buildings. If the number of buildings in B_List is greater than 1, take any building B in B_List. i , respectively through B i The first anchor point S and the end anchor point E are used as perpendicular lines to the major axis of the model unit, and the intersection points of the perpendicular lines and the model unit boundaries are calculated, which are recorded as S1 and S2, E1 and E2 respectively; determine whether the line segments S1E1 and S2E2 intersect. If they do, the polygon S1E2E1S2 is building B. iIf they do not intersect, polygon S1E1E2S2 is the building B. i ; Continue to execute step 6.2.3 until all buildings in B_List are traversed.

[0152] In order to verify the effectiveness and applicability of the method proposed in the present invention, the above key algorithms were implemented on the VisualStudio2012 platform based on ArcGIS Engine 10.2 using C# programming language, and typical experiments on typical combined straight line pattern building complexes and building groups were carried out using actual data.

[0153] (1) Typical combination straight line model building complex typification test

[0154] To verify the effectiveness of the proposed algorithm, the present invention selected four groups of combined straight line pattern buildings with typical characteristics from the 1:0000 scale topographic map of the Netherlands for experiment, such as Figure 7 As shown. The selected combined straight line patterns have different distribution and structural characteristics. From the perspective of spatial correspondence, pattern A and pattern B have spatial correspondences of 1:1 and 1:n, pattern B has a spatial correspondence of m:n, and pattern D has an n:n type of spatial correspondence, that is, the pattern units of pattern C and pattern D are all composed of buildings. From the perspective of the relationship between the overall pattern and the main direction of the pattern unit, patterns A, B and D are approximately orthogonal, and pattern C is approximately parallel. Therefore, the selected test data better summarizes the combined straight line patterns of different complexity existing in real data. Assume that the minimum spacing distance δ on the graph dis =0.5mm, the target scale is 1:40000, and the results of calling the proposed method to perform typical Figure 8 shown.

[0155] From the comprehensive results, it can be seen that the four groups of combined straight line model building complexes have been integrated to varying degrees, and can effectively maintain the characteristics of straight line distribution as a whole, and can also maintain the local heterogeneity characteristics of the building complex to a certain extent, such as model C. Therefore, the proposed method can effectively implement typical operations on combined straight line model building complexes with different characteristics, achieving the purpose of building complex integration.

[0156] (2) Experiment on Typing of Building Groups

[0157] In order to verify the rationality and applicability of the proposed method, the local building complex data with a scale of 1:5000 in Guangzhou, China, downloaded from OpenStreetMap, was used for experimental analysis. The experimental area contains a total of 288 buildings, such as Fig. 9As shown in (a). Most of the buildings in the selected test area have regular shapes and high rectangularity, and present obvious combined straight line patterns in visual perception. Since the method proposed in the present invention is designed for the integration of combined straight line pattern building complexes, this experiment only uses the proposed method to typify the building complexes that form combined straight line patterns in the test area. Before conducting the typification experiment, the combined straight line patterns of the test area were identified and extracted using the method proposed in the literature [Xing Ruixing, Wu Fang, Gong Xianyong, et al. Template matching method for identifying combined straight line patterns of building complexes [J]. Journal of Geodaetics and Cartography, 2021, 50(6): 800-811. DOI: 10.11947 / j.AGCS.2021.20200298.] Fig. 9 The buildings highlighted in (a) are the identified combined straight line patterns. Let the minimum separation distance δ dis =0.5mm, the target scales are set to 1:15000, 1:25000 and 1:35000 respectively, and the results of calling the proposed method to perform the synthesis are as follows Fig. 9 As shown in (b), (c) and (d).

[0158] In order to quantitatively evaluate the comprehensive results of the method of the present invention, the number, area and direction of the combined straight line patterns before and after the comprehensive analysis were statistically analyzed. Since the comprehensive analysis of pattern characteristics no longer emphasizes the individual, but considers the attribute of the element group as a group, the relevant information of the pattern as a whole is statistically analyzed here, and the statistical results are shown in Table 2. According to the number of buildings before and after the comprehensive analysis, it can be calculated that the comprehensive degree of different target scales is 50.00%, 68.75% and 75.00% respectively. By comparing and calculating the area of ​​the buildings before and after the comprehensive analysis, it can be seen that the area changes of the comprehensive analysis of each target scale are 16.86%, 18.89% and 18.89% respectively. The area of ​​the building complex after the comprehensive analysis is larger than that before the comprehensive analysis, but it is still within an acceptable range. There are two main reasons for the increase in area: one is that the present invention uses SMBR to represent the building; the other is that the merger of buildings in the pattern unit incorporates the gaps between the buildings into the merged buildings. In the experiment, the area change from the original data to 1:35000 is equal to that from the original data to 1:25000. This is because when the comprehensive scale is 1:25000, the combined buildings as model units in the original combined straight line model will be merged. When the comprehensive scale is less than 1:25000, the overall area of ​​the building complex will not continue to increase. Fig.10 In the figure, when the target scale is 1:25000, there is no combined building as a model unit in the model building complex after the integration of the original combined straight line model B. Comparison and analysis of the mean values ​​of the building directions before and after the integration show that the change in the mean value of the building directions after the integration does not exceed 2%, indicating that the building directions before and after the integration are basically consistent.

[0159] Table 2 Statistics of results before and after typification of combined straight line model

[0160]

[0161]

[0162] Combination Fig. 9 From the test results and the above analysis, it can be seen that the typicalization method proposed in this invention has the following advantages: ① It fully maintains the overall linear distribution characteristics of the building complex, and its spatial coverage does not change during the integration process; ② It also maintains the local spatial heterogeneity characteristics of the building complex. Since local heterogeneity belongs to low-level, local detail information, when the integration scale is large, this detail is no longer the main factor in distinguishing the building complex.

[0163] Based on the above embodiments, Fig.11 As shown, the present invention further provides a typical device for a combined straight line model building group, comprising:

[0164] The merging module is used to select any unintegrated combined straight line pattern building complex CP, perform a merging operation on the buildings, i.e., pattern units, that have spatial conflicts within the CP, and then execute the judgment module;

[0165] The judgment module is used to calculate the minimum spacing d between the pattern units in the CP ex , judge d ex and the distance constraint threshold δ dis The size of d ex <δ dis , then execute the deletion operation module; if d ex ≥δ dis , then execute the segmentation operation module;

[0166] A deletion operation module, used to perform a deletion operation on the CP and then execute the exaggeration operation module;

[0167] an exaggeration operation module, used for performing an exaggeration operation on the CP and then performing a shift operation module;

[0168] A shift operation module, used for performing a shift operation on the CP and then executing the judgment module;

[0169] The segmentation operation module is used to perform segmentation operation on CP to determine whether there is an unintegrated combined straight line pattern building complex. If so, it will go to the merging module, otherwise the typicalization ends.

[0170] Furthermore, the merging operation includes:

[0171] All buildings are represented by their minimum-area circumscribed rectangles (SMBRs), and the SMBR of the set of buildings within the pattern unit is used as the merged building.

[0172] A data structure Unit_Buildings is established to describe the spatial relationship and structure between the merged building and the original buildings. The data structure Unit_Buildings defines the pattern unit serial number, pattern unit elements, minor axis of the pattern unit, major axis of the pattern unit, building IDs included in the pattern unit, and building anchor points.

[0173] Furthermore, the deletion operation shall follow the following rules:

[0174] The first and last pattern units of the combined line pattern cannot be deleted.

[0175] Delete the pattern unit with the least impact, including: preferentially delete the pattern unit with the smallest sum of building areas; preferentially retain the pattern unit composed of combined buildings, and the higher the priority of retention for the pattern unit with a greater distance between internal buildings.

[0176] The deletion operation includes the following steps:

[0177] Step 3.1, traverse each pattern unit Unit in the combined line pattern Pattern i , 1 < i < Num, where Num is the number of pattern units in Pattern, calculate the maximum value Dis of the distances between adjacent buildings max , and add the array [i, Dis max to the linked list List_Distance. If Unit i contains only one building, calculate the sum of building areas Area sum , and add the array [i, Area sum to the linked list List_Area; execute Step 3.2.

[0178] Step 3.2, if the linked list List_Area is not empty, traverse and find the array [m, Area min in List_Area that contains the minimum area. Then the m-th pattern unit needs to be deleted, and the deletion operation ends. If the linked list List_Area is empty, execute Step 3.3;

[0179] Step 3.3, traverse and find the array [n, Dis min in List_Distance that contains the minimum distance. Then the n-th pattern unit needs to be deleted, and the deletion operation ends.

[0180] Furthermore, the exaggeration operation includes:

[0181] Step 4.1, determine the exaggerated area of ​​each pattern unit; including: using an average distribution strategy to distribute the area of ​​the deleted pattern unit, and record the area of ​​the deleted pattern unit as Area del , then the remaining area allocated to each mode unit is Area ave The calculation formula is as follows:

[0182]

[0183] Where Num is the number of remaining pattern units;

[0184] Step 4.2, exaggeration operation of the pattern unit; including: exaggeration of the building along its major axis or minor axis; when exaggerating, firstly calculating the angle between the major axis, the minor axis and the main direction of the pattern, and the axis direction with the smaller angle is the exaggeration direction; extending the rectangular sides of the building consistent with the exaggeration direction outward from both ends respectively, and forming the exaggerated building with the extended vertices, so as to realize the exaggeration of the building along the pattern direction;

[0185] The shift operation must follow the following rules:

[0186] Keep the spatial coverage of the pattern unchanged, that is, only the pattern units except the first and last ends can be shifted;

[0187] The shifted pattern unit tries to keep the trajectory consistent with the original combined straight line pattern;

[0188] The shift operation includes the following steps:

[0189] By making equal-interval differences on the lines connecting the original centroids of all pattern units, the coordinates of the shifted points of the centroids of each pattern unit can be obtained, and then the shift operation is completed;

[0190] The segmentation operation includes:

[0191] Step 6.1, update the pattern unit structure data: for the pattern unit elements, pattern unit short axis and pattern unit long axis in Unit_Buildings, update them according to the SMBR of the corresponding pattern unit; keep the ratio of the distance value between each anchor point and the pattern unit long axis length value equal before and after the update;

[0192] Step 6.2, reconstruction of buildings within the model unit: judge the distance between adjacent anchor points of adjacent buildings. If the distance is less than the distance constraint threshold δ dis The adjacent anchor points will be deleted.

[0193] In summary, combined with the distribution pattern and spatial structure characteristics of the combined straight line model building complex, the present invention adopts the idea of ​​decomposing complex operators into simple operator combinations, and decomposes the typical operation of the combined straight line model building complex into a "progressive, iterative" process composed of five basic operators such as merging, deleting, exaggerating, shifting and splitting. A feasible new case is provided in terms of operator collaboration and algorithm collaboration. The present invention realizes the integration of building complexes that take into account the combined straight line model. The typicalization method proposed in the present invention can fully maintain the overall straight line model distribution characteristics of the building complex, and its spatial coverage does not change during the integration process. The method also takes into account the local spatial heterogeneity characteristics of the building complex. Since local heterogeneity belongs to low-level, local detail information, when the comprehensive scale is large, this detail is no longer the main factor in distinguishing the building complex. In addition, the method can effectively implement typical operations for combined straight line model building complexes with different characteristics. In general, the proposed typicalization method can effectively and reasonably integrate the combined straight line model building complex.

[0194] The above is only a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A method for typifying a group of buildings in a combined straight line pattern, characterized in that: include: Step 1, randomly select an unintegrated combined straight line pattern building complex CP, perform a merge operation on the buildings, i.e., pattern units, that have spatial conflicts within the CP, and then perform step 2; The merging operation includes: All buildings are represented by their minimum area circumscribed rectangle SMBR, and the SMBR of the building set in the model unit is used as the merged building; Establish a data structure Unit_Buildings that describes the spatial relationship and structure between the merged building and the original building; the data structure Unit_Buildings defines a pattern unit sequence number, a pattern unit element, a pattern unit short axis, a pattern unit long axis, a building ID included in the pattern unit, and a building anchor point; Step 2: Calculate the minimum spacing d between the mode units in the CP ex , judge d ex and the distance constraint threshold δ dis The size of d ex <δ dis , then execute step 3; if d ex ≥δ dis , then execute step 6; Step 3, perform a delete operation on the CP, and then perform step 4; The deletion operation must follow the following rules: The first and last mode units of the combined straight line mode cannot be deleted; Delete the model unit with the least impact; including: give priority to deleting the model unit with the smallest sum of building areas; give priority to retaining the model unit composed of combined buildings, and the larger the distance between internal buildings, the higher the retention priority of the model unit; The deletion operation includes the following steps: Step 3.1, traverse each pattern unit Unit in the combined straight-line pattern Pattern i , where 1 < i < Num and Num is the number of pattern units in Pattern, calculate the maximum value Dis of the distances to adjacent buildings max , and add the array [i, Dis max to the linked list List_Distance. If Unit i contains only one building, calculate the sum of the building areas Area sum , and add the array [i, Area sum to the linked list List_Area; execute Step 3.2; Step 3.2, if the linked list List_Area is not empty, then traverse to find the array [m, Area in List_Area that contains the minimum area value min ], then the mth pattern unit needs to be deleted, and the deletion operation ends; if the linked list List_Area is empty, execute step 3.3; Step 3.3, traverse and find the array [n, Dis min ], then the nth pattern unit needs to be deleted, and the deletion operation ends; Step 4, perform an exaggeration operation on CP, and then perform step 5; Step 5, perform a shift operation on CP, and then perform step 2; Step 6, perform segmentation operation on CP to determine whether there is an unintegrated combined straight line pattern building complex. If so, go to step 1, otherwise the typicalization ends.

2. A method for typifying a combined straight line pattern building complex according to claim 1, characterized in that: The exaggeration operation includes: Step 4.1, determine the exaggerated area of ​​each pattern unit; including: using an average distribution strategy to distribute the area of ​​the deleted pattern unit, and record the area of ​​the deleted pattern unit as Area del , then the remaining area allocated to each mode unit is Area ave The calculation formula is as follows: Where Num is the number of remaining pattern units; Step 4.2, exaggeration operation of the pattern unit; including: the building is exaggerated along its major axis or minor axis; when exaggerating, the angles between the major axis, the minor axis and the main direction of the pattern are first calculated, and the axis direction with the smaller angle is the exaggeration direction; the rectangular sides of the building consistent with the exaggeration direction are extended outward from both ends respectively, and the extended vertices form the exaggerated building, thereby realizing the exaggeration of the building along the pattern direction.

3. The method for typifying a combined straight line pattern building complex according to claim 1, characterized in that: The shift operation must follow the following rules: Keep the spatial coverage of the pattern unchanged, that is, only the pattern units except the first and last ends can be shifted; The shifted pattern unit tries to keep the trajectory consistent with the original combined straight line pattern; The shift operation includes the following steps: By making equal-interval differences on the lines connecting the original centroids of all pattern units, the coordinates of the shifted points of the centroids of each pattern unit can be obtained, and then the shift operation is completed.

4. The method for typifying a combined straight line pattern building complex according to claim 1, characterized in that: The segmentation operation includes: Step 6.1, update the pattern unit structure data: for the pattern unit elements, pattern unit short axis and pattern unit long axis in Unit_Buildings, update them according to the SMBR of the corresponding pattern unit; keep the ratio of the distance value between each anchor point and the pattern unit long axis length value equal before and after the update; Step 6.2, reconstruction of buildings within the model unit: judge the distance between adjacent anchor points of adjacent buildings. If the distance is less than the distance constraint threshold δ dis The adjacent anchor points will be deleted.

5. A typical device for combining straight line model buildings, characterized in that: include: The merging module is used to select any unintegrated combined straight line pattern building complex CP, perform a merging operation on the buildings, i.e., pattern units, that have spatial conflicts within the CP, and then execute the judgment module; The merging operation includes: All buildings are represented by their minimum area circumscribed rectangle SMBR, and the SMBR of the building set in the model unit is used as the merged building; Establish a data structure Unit_Buildings that describes the spatial relationship and structure between the merged building and the original building; the data structure Unit_Buildings defines a pattern unit sequence number, a pattern unit element, a pattern unit short axis, a pattern unit long axis, a building ID included in the pattern unit, and a building anchor point; The judgment module is used to calculate the minimum spacing d between the pattern units in the CP ex , judge d ex and the distance constraint threshold δ dis The size of d ex <δ dis , then execute the deletion operation module; if d ex ≥δ dis , then execute the segmentation operation module; A deletion operation module, used to perform a deletion operation on the CP and then execute the exaggeration operation module; The deletion operation must follow the following rules: The first and last mode units of the combined straight line mode cannot be deleted; Delete the model unit with the least impact; including: give priority to deleting the model unit with the smallest sum of building areas; give priority to retaining the model unit composed of combined buildings, and the larger the distance between internal buildings, the higher the retention priority of the model unit; The deletion operation includes the following steps: Step 3.1, traverse each pattern unit Unit in the combined line pattern Pattern i , where 1 < i < Num and Num is the number of pattern units in Pattern, calculate the maximum value Dis of the distances to adjacent buildings max , and add the array [i, Dis max to the linked list List_Distance. If Unit i contains only one building, calculate the sum of the building areas Area sum , and add the array [i, Area sum to the linked list List_Area; execute Step 3.2; Step 3.2, if the linked list List_Area is not empty, then traverse to find the array [m, Area in List_Area that contains the minimum area value min ], then the mth pattern unit needs to be deleted, and the deletion operation ends; if the linked list List_Area is empty, execute step 3.3; Step 3.3, traverse and find the array [n, Dis min ], then the nth pattern unit needs to be deleted, and the deletion operation ends; an exaggeration operation module, used for performing an exaggeration operation on the CP and then performing a shift operation module; A shift operation module, used for performing a shift operation on the CP and then executing the judgment module; The segmentation operation module is used to perform segmentation operation on CP to determine whether there is an unintegrated combined straight line pattern building complex. If so, it will go to the merging module, otherwise the typicalization ends.

6. The device for typifying a group of combined straight-line buildings according to claim 5, characterized in that: The exaggeration operation includes: Step 4.1, determine the exaggerated area of ​​each pattern unit; including: using an average distribution strategy to distribute the area of ​​the deleted pattern unit, and record the area of ​​the deleted pattern unit as Area del , then the remaining area allocated to each mode unit is Area ave The calculation formula is as follows: Where Num is the number of remaining pattern units; Step 4.2, exaggeration operation of the pattern unit; including: exaggeration of the building along its major axis or minor axis; when exaggerating, firstly calculating the angle between the major axis, the minor axis and the main direction of the pattern, and the axis direction with the smaller angle is the exaggeration direction; extending the rectangular sides of the building consistent with the exaggeration direction outward from both ends respectively, and forming the exaggerated building with the extended vertices, so as to realize the exaggeration of the building along the pattern direction; The shift operation must follow the following rules: Keep the spatial coverage of the pattern unchanged, that is, only the pattern units except the first and last ends can be shifted; The shifted pattern unit tries to keep the trajectory consistent with the original combined straight line pattern; The shift operation includes the following steps: By making equal-interval differences on the lines connecting the original centroids of all pattern units, the coordinates of the shifted points of the centroids of each pattern unit can be obtained, and then the shift operation is completed; The segmentation operation includes: Step 6.1, update the pattern unit structure data: for the pattern unit elements, pattern unit short axis and pattern unit long axis in Unit_Buildings, update them according to the SMBR of the corresponding pattern unit; keep the ratio of the distance value between each anchor point and the pattern unit long axis length value equal before and after the update; Step 6.2, reconstruction of buildings within the model unit: judge the distance between adjacent anchor points of adjacent buildings. If the distance is less than the distance constraint threshold δ dis The adjacent anchor points will be deleted.

Images