A method for constructing a 15-minute urban life circle boundary based on multi-constraint paths

Abstract

This invention discloses a method for constructing the boundary of a 15-minute urban living circle based on multi-constraint paths, specifically relating to the fields of urban spatial analysis and edge computing. The method includes inputting community units, road segments, pedestrian crossings, passageway segments, open segments, entry edges, and blocking segments into edge nodes. An edge graph is generated within the edge nodes according to their connectivity relationships, and corresponding connections are severed using blocking segments. The edge graph and access points are then output. Using the access points and edge graph as input, walking time is accumulated segment by segment along the edge graph within the edge nodes. Road segments, pedestrian crossings, passageway segments, and open segments with a cumulative walking time not exceeding 15 minutes and directly connected to the preceding segment are sequentially written into the receiving layer. By constructing the edge graph within the edge nodes, generating the receiving layer, and using the search results of the area to be inspected to reverse-filter out variant outer edge chains that include unsearched objects, the method can relatively suppress the inclusion of non-continuously accessible internal spaces into the living circle boundary, making the obtained outer edge more consistent with the actual walking receiving relationship.

Description

Technical Field

[0001] This invention relates to the field of urban spatial analysis and edge computing technology, and more specifically, to a method for constructing the boundary of a 15-minute urban living circle based on multi-constraint paths. Background Technology

[0002] The delineation of a 15-minute living circle in a city refers to identifying the spatial range within a limited walking time range where residents can reliably obtain basic public services and daily living services, based on their daily walking activities. The results of this delineation are usually used to support the assessment of community-level facility configuration, the identification of service shortcomings, and the optimization and renewal of the area. In the delineation of the 15-minute living circle in cities, existing technologies usually focus on determining the spatial range that community residents can cover within a limited walking time, and carry out facility configuration evaluation, shortcoming identification and update planning accordingly. In engineering, the approach is often to first calculate the reachable range based on the road network, pedestrian passages and crosswalk nodes, and then generate the boundary from the reachable result, or after the boundary is formed, combine hole filling, local trimming and smoothing to modify the boundary shape. However, in old urban areas, mountainous cities, and areas with a mix of large, enclosed blocks, there are often spatial barriers such as railway barriers, enclosed campuses, enclosed courtyards, continuous main roads that cannot be crossed, and elevation differences. Furthermore, the boundaries of the living area still need to meet the constraints of limited walking time, accessible outer facilities, and continuous access to the internal space. In this case, although the boundaries obtained by the current method can form a closed outer edge on the map, and the surrounding commercial, medical, cultural and sports facilities may fall within the boundary range, there will be repeated instances of large areas of space being completely enclosed inside the boundary, making it impossible for residents to walk in segments from the community starting point. This manifests as large areas of seemingly covered but actually unconnected inner voids inside the boundary. Whether the voids are retained or filled in later, it can only correct the graphic representation and cannot explain why the outer edge has been established but the internal connection chain is broken. Consequently, the judgment of facility coverage, the identification of shortcomings, and the analysis of supplementary points deviate from the actual usage relationship. The technical problem this application aims to solve is: how to avoid including internal spaces that cannot be continuously accessed from the community starting point into the boundary during the construction of the 15-minute living circle boundary in the city, so as to obtain the outer edge of the living circle that conforms to the actual walking access relationship. Summary of the Invention

[0003] To overcome the aforementioned deficiencies in the prior art, embodiments of the present invention provide a method for constructing the boundary of a 15-minute urban living circle based on multi-constraint paths. This method involves constructing a community perimeter edge map within edge computing nodes, generating a receiving layer layer by layer according to access points, performing chain-like mutation reconstruction on the outer edge segments, and deleting the outer edge chains that include unreceived spaces by combining connectivity search of the area to be inspected. This solves the problem mentioned in the background art where the internal space that cannot be continuously received from the community starting point is included in the living circle boundary.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a method for constructing the boundary of a 15-minute urban living circle based on multi-constraint paths, comprising: S1. Input community units, road segments, pedestrian crossings, passage segments, open segments, entry edges, and barrier segments into edge nodes. Generate an edge graph within the edge nodes according to the connection relationship, and cut off the corresponding connection with the barrier segment. Output the edge graph and access point. S2. Using the access point and edge map as input, accumulate the walking time segment by segment along the edge map within the edge node. Write the road segments, crossing points, passage segments and open segments with a cumulative walking time of no more than 15 minutes that are directly connected to the previous segment into the receiving layer in sequence, and output the receiving layer and the outer edge segment group. S3. Taking the outer edge segment group as input, perform neighbor segment search, start and end splicing and chain position replacement on each outer edge segment within the edge node to obtain the outer edge candidate chain, and perform outer side chaining and inner side shrinking on each outer edge candidate chain to output the mutated outer edge chain group. S4. Using the mutated outer edge chain group, the receiving layer and the access point as input, each mutated outer edge chain and its inner receiving layer are enclosed in the inspection area within the edge node. Then, starting from the access point, each segment in the inspection area is searched segment by segment along the receiving layer. The mutated outer edge chains with unsearched segments are deleted. The remaining mutated outer edge chains are retained and the replacement, chain merging, shrinking and searching are repeated. The target outer edge chain is output. S5. Input the target outer edge chains output by each edge node into the convergence node. In the convergence node, splice the target outer edge chains that are connected end to end, remove duplicates from the overlapping parts, and perform closed connection on the splicing result to output the boundary of the city's 15-minute living circle.

[0005] In a preferred embodiment, S1 includes: S1-1. Input the road segment, crossing point, passage segment, open segment and entry edge into the edge node. Within the edge node, take the endpoints of each segment as the start and end points, the crossing point as the cross-segment connection point, and the entry edge as the segment side access edge. First, establish the start and end connection for the road segment, passage segment and open segment with overlapping endpoints. Then, establish the cross-point connection for each segment connected on both sides of the crossing point. Finally, establish the segment edge connection for the open segment or passage segment that intersects with the entry edge. Output the initial edge graph. S1-2. Input the blocking segment and the initial edge graph into the edge node. In the edge node, check whether each connection passes through the blocking segment, intersects with the blocking segment, or whether the two ends of the connection are located on both sides of the blocking segment. For the connection that satisfies any of the above conditions, cut it off. Number each set of independent connected segments formed after the cut off, and output the split edge graph and the set of connected segments. S1-3. Input the community unit, the segmented edge graph, and the connected segment set into the edge node. Search for the road segment, passage segment, and open segment that intersect with the outer boundary of the community unit within the edge node. Record the intersection point of each segment located on the outer boundary of the community unit as the initial access point. Then retain the initial access point that is in the same connected segment set as the initial access point and is not separated by the blocking segment as the access point. Output the edge graph and the access point.

[0006] In a preferred embodiment, S2 includes: S2-1. Input the access point and edge graph into the edge node. Within the edge node, search outwards along the edge graph segment by segment, starting from each access point. For each search path, accumulate the segment length and the crossing time to form the path time. Record the minimum time of each segment in each search path as the arrival time and output the arrival time segment set. S2-2. Input the arrival time segment set into the edge node. Read each segment in the edge node in order of arrival time from smallest to largest. Write the road segment, crossing point, passage segment and open segment with an arrival time of no more than 15 minutes and whose preceding connected segment has been written into the corresponding receiving layer. Write the segments with the same arrival time into the same receiving layer and output the receiving layer sequence. S2-3. Input the receiving layer sequence into the edge node. In the edge node, for each receiving layer, search for each segment that is directly connected to a segment that is not in this receiving layer. Record each road segment, crossing point, passage segment and open segment that is directly connected to the next receiving layer or a segment that has not been written and is located outside the current receiving layer as the outer edge segment. Output the receiving layer and the outer edge segment group.

[0007] In a preferred embodiment, S3 includes: S3-1. Input the outer edge segment group into the edge node. Extract the coordinates of the two endpoints, direction vector and segment length for each outer edge segment within the edge node. Search for the adjacent segments that connect to the two endpoints of each outer edge segment. Construct an adjacency matrix according to the endpoint connection relationship between adjacent segments. Construct a direction matrix according to the dot product of the direction vector of the adjacent segment and the direction vector of the corresponding outer edge segment. Perform matrix multiplication on the adjacency matrix and the direction matrix to obtain a two-step expansion matrix. Use the adjacent segments corresponding to the non-zero elements in the two-step expansion matrix as the neighbor segment set. Delete the adjacent segments that intersect the inner side of the corresponding outer edge segment, form a reverse fold with the corresponding outer edge segment, or generate self-intersection after connection. Output the neighbor segment set.

[0008] In a preferred embodiment, S3 further includes: S3-2. Input the neighbor segment set into the edge node. Within the edge node, take each outer edge segment as the initial chain. Select neighbor segments to be connected from both ends of the initial chain. For each neighbor segment to be connected, calculate the number of breaks, crosses, backtracks, total length increment, and the sum of distances from the chain endpoints to the two ends of the corresponding outer edge segment after connection. Construct a five-element cost sequence using the number of breaks, crosses, backtracks, total length increment, and distance sum. Select the connected segments in rounds according to the lexicographical order of break number priority, crosses second priority, backtracks third priority, total length increment third priority, and last element of distance sum priority. After each round of connection, compare the five-element cost sequence of the newly generated chain with the five-element cost sequence of the previous round item by item. Repeat the two-end connection until the two ends of the current chain reach the two ends of the corresponding outer edge segment, or the five-element cost sequence after all neighbor segments to be connected are connected is not better than the current chain. Output the outer edge candidate chain.

[0009] In a preferred embodiment, S3 further includes: S3-3. Input the candidate outer edge chains and the outer edge segment group into the edge node. Replace the corresponding outer edge segment with the candidate outer edge chain within the edge node. Construct a chain connection matrix for each replaced chain according to the head-to-tail connection relationship. Construct a chain conflict matrix according to the chain intersection, enclosing, and inner penetration relationships. Perform connectivity decomposition on the chain connection matrix to obtain candidate chain clusters. After resolving the conflict chains in the chain conflict matrix of each candidate chain cluster, perform head-to-tail splicing to obtain the chain-joined result. Then, calculate the cosine value of the included angle between two adjacent segments and the perpendicular distance from the endpoint of each segment to the head-to-tail connection line for each chain-joined result. Delete the concave segments with negative included angle cosine values ​​and which are still connected head-to-tail after deletion. Repeat the conflict resolution, head-to-tail splicing, and concave segment deletion for the deleted chains until the chain connection matrix no longer changes and the chain conflict matrix is ​​all zero. Output the mutated outer edge chain group.

[0010] In a preferred embodiment, S4 includes: S4-1. Input the mutated outer edge chain group, the receiving layer, and the access point into the edge node. Extract the chain segment endpoint sequence for each mutated outer edge chain within the edge node. Select the receiving layer segment set located inside each mutated outer edge chain and connected to the beginning and end of each mutated outer edge chain. Close the beginning and end of each mutated outer edge chain and the receiving layer segment set to form the inspection area. Construct the intra-area adjacency matrix according to the beginning and end connection relationship of each segment in the inspection area. Construct the starting vector according to the case where the access point falls into each segment in the inspection area. Output the inspection area, the intra-area adjacency matrix, and the starting vector.

[0011] In a preferred embodiment, S4 further includes: S4-2. Input the region to be inspected, the adjacency matrix within the region, and the starting vector into the edge node. Within the edge node, use the starting vector as the first-round search vector. Multiply the first-round search vector by the adjacency matrix within the region to obtain the next-round search vector. Record the non-zero values ​​of the next-round search vector as 1 and the zero values ​​as 0, and add it bit by bit to the cumulative search vector of the previous round. Record the non-zero bits in the addition result as 1 and the zero bits as 0 to obtain the new cumulative search vector. Repeat the left multiplication, bit by bit addition, and binarization until the new cumulative search vector has the same values ​​as the previous cumulative search vector. Output the cumulative search vector and the set of unsearched segments. S4-3. Input the cumulative search vector, the set of unsearched segments, and the group of mutated outer edges into the edge node. Delete mutated outer edges that are not empty in the set of unsearched segments within the edge node. Count the total number of segments in the inspection area of ​​each mutated outer edge after retention, the number of non-zero bits in the cumulative search vector, and the number of connected components in the set of unsearched segments. Subtract the non-zero bits from the total number of segments to obtain the number of missing segments. Form an adjudication sequence using the number of missing segments and the number of connected components. Arrange the mutated outer edges after retention in ascending order of the number of missing segments and the number of connected components. Output the group of retained chains and the adjudication sequence.

[0012] In a preferred embodiment, S4 further includes: S4-4. Input the retained chain group, adjudication sequence, receiving layer and access point into the edge node. Read the retained chain group one by one in the order of adjudication sequence within the edge node. Repeat the replacement, chain merging, shrinking, inspection area reconstruction, intra-area adjacency matrix reconstruction, starting vector reset, cumulative search vector recalculation and unsearched segment set recalculation for each retained chain. Stop when the values ​​of each item in the adjudication sequence of two adjacent rounds are consistent, and output the target outer edge chain.

[0013] In a preferred embodiment, S5 includes: S5-1. Input the target outer edge chains output by each edge node into the convergence node. Extract the first endpoint, the last endpoint, and the chain segment sequence of each target outer edge chain in the convergence node. Construct the endpoint connection matrix according to the overlap relationship between the first endpoint and the last endpoint. Construct the overlapping segment matrix according to the overlap relationship of the chain segment coordinates. Perform connectivity decomposition on the endpoint connection matrix to obtain spliced ​​chain groups. In each spliced ​​chain group, splice the target outer edge chains in the order of first and last connection. Delete the overlapping parts according to the overlapping segment matrix. Output the deduplicated spliced ​​chain group. S5-2. Input the deduplication splicing chain group into the aggregation node. Calculate the connection segment between the first and last endpoints for each deduplication splicing chain group within the aggregation node. Write the connection segment into the corresponding deduplication splicing chain group to form a closed chain. Then, delete the duplicate closed chains located inside the outer closed chains according to the inclusion relationship between the closed chains. Output the remaining closed chains in the order of continuous first and last, to obtain the boundary of the city's 15-minute living circle.

[0014] The technical effects and advantages of this invention are as follows: By constructing an edge graph within the edge nodes, generating a connection layer, and using the search results of the area to be inspected to reverse-filter out the variant outer edge chains that include unsearched objects, it is possible to relatively suppress the writing of non-continuously accessible internal spaces into the boundary of the living circle, making the obtained outer edge more consistent with the actual walking connection relationship. By first cutting off the connection according to the blocking section, then extracting the access point from the outer boundary of the community unit and expanding the receiving layer layer by layer along the access point, it is possible to distinguish the spatial objects blocked by the closed boundary from the spatial objects that can be entered, so that the pedestrian propagation results are relatively close to the actual passage conditions. By performing adjacent segment search, chain replacement, chaining and retraction on the outer edge segment group, and combining the intersection, enclosing and inner penetration relationships to resolve conflict chains, it is possible to constrain the breakage, folding and repeated expansion of the outer boundary, thereby relatively improving the continuity of the outer edge and the stability of the boundary morphology. By performing splicing, deduplication, and closure connections on the target outer edge chains output by each edge node within the aggregation node, the partitioning results can be integrated into a continuous boundary, and the bias caused by repeated writing or disconnection at cross-node boundary positions on facility coverage judgment and area update analysis can be relatively reduced. Attached Figure Description

[0015] Figure 1 This is a flowchart of the method steps of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Refer to the instruction manual appendix Figure 1 The present invention provides a method for constructing the boundary of a 15-minute urban living circle based on multi-constraint paths, comprising: S1. Input community units, road segments, pedestrian crossings, passage segments, open segments, entry edges, and barrier segments into edge nodes. Generate an edge graph within the edge nodes according to the connection relationship, and cut off the corresponding connection with the barrier segment. Output the edge graph and access point. In one specific implementation, S1 is used to generate an edge graph within the edge node for direct reading by subsequent layer expansion, and deletes connections interrupted by blocked segments during the generation process, while extracting access points corresponding to the community unit; wherein, road segments, passage segments, open segments, entry edges, and blocked segments are represented by endpoint coordinate sequences in the same plane coordinate system, and street crossings are represented by point coordinates; the edge node is deployed in the edge computing processing unit corresponding to the community unit, and is responsible for completing the connection establishment, connection disconnection, and access point extraction of the surrounding spatial objects of the community unit. This implementation process includes the following steps: S1-1 is used to organize road segments, passageways, open sections, crossings, and entry edges into an initial edge map. The processing method is as follows: First, the edge nodes read the road segments, crossings, passageways, open sections, and entry edges. Then, for each road segment, passageway, and open section, the first and last endpoints are extracted, and the sum of squared coordinate differences between any two endpoints is calculated. When the sum of squared coordinate differences is not greater than the squared value corresponding to the base map acquisition accuracy, the corresponding two road segments, passageways, or open sections are written into the first-to-last connection table. Next, each crossing is read, and the positions where the crossing coordinates coincide with or fall within the lines of the road segment, passageway, and open section endpoints are retrieved. The object pairs located on both sides of the crossing are then written into the connection table. Cross-point connection table; then read each entering edge, retrieve the open segment and channel segment that intersect with the entering edge, and write the intersection position into the segment edge connection table; when there are multiple intersection positions between the entering edge and the open segment or channel segment, record each intersection position and establish segment edge connection respectively; when there is an overlapping interval between the entering edge and the open segment or channel segment, split the overlapping interval into the first endpoint and the last endpoint and then establish segment edge connection; finally, write the first and last connection table, the cross-point connection table and the segment edge connection table into the initial edge graph for S1-2 to read; when no connection relationship is found for the road segment, channel segment or open segment, the edge node retains the corresponding object and writes it into the initial edge graph as an isolated object; S1-2 is used to remove connections blocked by obstructed segments from the initial edge map. The process is as follows: Edge nodes read the obstructed segments and the initial edge map. For each end-to-end connection, cross-point connection, and segment-edge connection in the initial edge map, the coordinates of the two ends of the connection and the connection path are extracted. For each obstructed segment, the connection path is calculated to determine if it intersects with the obstructed segment, whether it crosses the boundary of the obstructed segment, and whether the signs of the two ends of the connection are reversed when substituted into the line containing the obstructed segment. When any calculation result is true, the corresponding connection is deleted from the initial edge map. After all obstructed connections are deleted, the edge nodes process the remaining connections for road segments and passages. Segments, open segments, and incoming edges undergo connectivity search, and a set of mutually connected objects is written as a connected segment set, which is numbered sequentially according to the generation order to obtain the segmentation edge graph and connected segment set for S1-3 to read; when a blocking segment coincides with a connection endpoint but does not cross the connection path, the edge node continues to determine whether the blocking segment is located between the extension directions of the connection path. If it is located between the extension directions of the connection path, the corresponding connection is deleted; if it is not located between the extension directions of the connection path, the corresponding connection is retained; when a road segment, channel segment, open segment, or incoming edge is no longer connected to other objects after deleting the blocking connection, the edge node numbers the corresponding object as a connected segment set separately; S1-3 is used to extract access points corresponding to community units from the segmented edge map. The processing method is as follows: the edge node reads the community unit, the segmented edge map, and the connected segment set. It then searches for the intersection points of the community unit's outer boundary with road segments, passage segments, and open segments segment by segment, and writes the intersection point coordinates into the initial access point set. When the same road segment, passage segment, or open segment forms multiple intersection points with the community unit's outer boundary, all intersection points are written into the initial access point set in the order along the community unit's outer boundary. When there is an overlapping section between the road segment, passage segment, or open segment and the community unit's outer boundary, the two endpoints of the overlapping section are written into the initial access point set. Subsequently, the edge node searches for each initial access point... The system identifies the connected segment set to which the corresponding object belongs and checks whether there is a blocking segment cutoff relationship between the outer boundary of the community unit and the initial access point. Only the initial access points whose connected segment set has been numbered and which do not have a blocking segment cutoff relationship between the outer boundary of the community unit and the initial access point are retained as access points. The retained results are written into the access point set, and the segmented edge map and the access point set are output together as the edge map and access point for S2 to read. When the object corresponding to the initial access point is not written into any connected segment set, the edge node deletes the corresponding initial access point. When the coordinates of two initial access points coincide and they correspond to the same connected segment set, the edge node merges the two initial access points into one access point. In practical applications, edge nodes can be deployed within community-level edge computing devices. They first read data from the surrounding road network, neighborhood access routes, open spaces, facility access edges, and construction barriers, railway barriers, or closed boundaries to generate an initial edge graph. Then, they delete connections blocked by obstructed segments and reconstruct the connected segment set. Subsequently, they extract initial access points from the outer boundary of the community unit, deleting initial access points not written into the connected segment set or blocked by obstructed segments, thus obtaining an edge graph and access points that can be directly read by the extension layer. In scenarios where closed courtyards, railway barriers, and internal passages coexist in old urban areas, after the above processing, the input objects read in subsequent steps have excluded connections blocked by closed boundaries and retained access locations that the community unit can actually access.

[0018] S2. Using the access point and edge map as input, accumulate the walking time segment by segment along the edge map within the edge node. Write the road segments, crossing points, passage segments and open segments with a cumulative walking time of no more than 15 minutes that are directly connected to the previous segment into the receiving layer in sequence, and output the receiving layer and the outer edge segment group. In one specific implementation, S2 is used to write road segments, pedestrian crossings, passageway segments, and open segments in the edge map into the receiving layer according to the actual walking propagation order, starting from the access point within the edge node. After the receiving layer is formed, the outer edge segment group is extracted for subsequent outer edge replacement, chaining, and shrinkage processing. The edge map uses the output result of S1, the access point uses the extraction result of S1, and the walking time for road segments, passageway segments, and open segments is obtained by dividing the length of the corresponding object by the walking speed. The walking speed uses the walking parameters written to the edge node during system deployment. The pedestrian crossing time is obtained by adding the walking time corresponding to the crossing length to the waiting time. The waiting time is taken from intersection signal control data, on-site collected data, or area basic configuration data. All times are uniformly converted to seconds before being accumulated. This implementation process includes the following steps: S2-1 is used to calculate the arrival time of each object within the edge graph from the access point outwards. Its function is to convert discrete connectivity relationships into sortable walking propagation results. In specific execution, the edge node reads the access point and the edge graph, establishing a search queue and arrival time table for each access point as the starting point. It calculates the first-round path time for road segments, passageways, and open segments directly connected to the access point. The path time for road segments, passageways, and open segments is equal to the length of the corresponding object divided by the walking speed. The path time for a street crossing is equal to the cumulative path time of the object before entering the street crossing plus the crossing time. Subsequently, the edge node reads the current path sequentially according to the search queue, continuing to expand outwards for the road segments, street crossings, passageways, and open segments directly connected to the end of the current path, and recording the path time of the newly expanded objects as the current path time. The length is the sum of the walking time or street crossing time corresponding to the newly extended object; when the same object is reached by multiple search paths, the edge node compares the path lengths corresponding to each search path, retains only the path length with the smaller value as the arrival time of the object, and updates the arrival time table with this arrival time; when the path length of the newly extended object is greater than the recorded arrival time, the edge node deletes the current extension result and stops searching along the current extension result; when no new objects are written to the search queues corresponding to all access points, the edge node ends the search, outputs the road segments, street crossings, passage segments, and open segments written to the arrival time table and their arrival times as an arrival time segment set for S2-2 to read; when there are closed loop connections in the edge graph, the edge node relies on the rule that the arrival time is only updated when the value decreases to prevent repeated loops; S2-2 is used to convert the arrival time segment set into a continuation layer sequence. Its function is to organize the continuous arrival time results into the hierarchical results that can be directly read by subsequent boundary extraction. In specific execution, the edge node reads the arrival time segment set, sorts all road segments, crossings, passageways, and open segments in ascending order of arrival time, and reads the sorted results in sequence. For the current read object, the edge node first determines whether the arrival time of the current read object does not exceed 900 seconds, and then searches whether the preceding connected object corresponding to the current read object has been written into the continuation layer. The preceding connected object refers to the road segment, crossing, passageway, or open segment located at the previous position of the current read object along the search path that generated the arrival time of the current read object. When the arrival time of the current read object does not exceed 900 seconds, and at least one preceding connected object has been written into the continuation layer, the edge node will... The edge node writes the read object to the corresponding receiving layer. When multiple objects have the same arrival time, the edge node writes all objects to the same receiving layer. When the current read object is a directly connected object of the access point, the edge node writes the current read object directly to the first receiving layer. When the arrival time of the current read object exceeds 900 seconds, the edge node stops writing. When the current read object does not have a preceding connected object that has been written to the receiving layer, the edge node postpones writing and continues to read the next object. After this round of sorting and reading, it searches for the postponed object again. If the corresponding preceding connected object has been written to the receiving layer during the second search, the corresponding postponed object is written to the receiving layer corresponding to the arrival time. If the preceding connected object still does not exist after the second search, the corresponding postponed object is deleted. Finally, the edge node outputs the receiving layer sequence in the order of arrival time for S2-3 to read. S2-3 is used to extract the outer edge segment group from the bearing layer sequence. Its function is to identify objects distributed on the outer edge of the bearing layer that directly face the unwritten area as the starting objects for subsequent boundary deformation processing. In specific execution, the edge node reads the bearing layer sequence and processes each bearing layer in sequence according to the bearing layer number. For each road segment, crossroads, passage segments, and open segments in the current bearing layer, it searches for objects that are directly connected to it but do not belong to the current bearing layer and records the layer to which the corresponding object belongs. Subsequently, the edge node calculates the direction of the line connecting the center point of the community unit to the center coordinates of the corresponding object for each road segment, crossroads, passage segments, and open segments in the current bearing layer. Then, it calculates the distribution position of the center coordinates of non-current bearing layer objects that are directly connected to it relative to the line. The edge node then selects the object that is far away from the center point of the community unit and is connected to the next bearing layer object or an unwritten object. Directly connected road segments, crossings, passageways, and open sections are denoted as outer edge segments. When an object is simultaneously connected to the previous layer object, the next layer object, and unwritten objects, the edge node only retains the object facing the next layer object or the unwritten object as an outer edge segment. When an object is only connected to objects within the current layer and not connected to the next layer object or unwritten objects, the edge node does not write the object into the outer edge segment group. After each layer is processed, the edge node writes the outer edge segments in the corresponding layer into the outer edge segment group sequence according to the layer number, and outputs both the layer sequence and the outer edge segment group for S3 to read. When multiple outer edge segments are connected end to end to form a continuous outer boundary, the edge node maintains the original order of each outer edge segment and writes it into the corresponding outer edge segment group of the same layer to avoid order jumps during subsequent outer edge candidate chain generation. In practical applications: edge nodes can be deployed in community-level edge computing devices. They first read the access points and edge maps output by S1, and then calculate the arrival time layer by layer outward, taking the roads around the community unit, the internal passages of the block, the passage lines of open spaces, and the pedestrian crossing facilities as the propagation objects. For example, when there is an open block passage on the north side of the community, a signal-controlled intersection on the east side, and an open space with fixed pedestrian access time on the west side, the edge node first calculates the walking time of the road segment, passage segment, and open segment according to the object length and walking speed. Then, it adds the waiting time at the intersection to the walking time corresponding to the crossing length to obtain the crossing time. Subsequently, it writes the object into the receiving layer according to the arrival time, and extracts the outer edge segment facing the next receiving layer or the area not written from each receiving layer. Finally, it forms the receiving layer sequence and outer edge segment group for S3 to read. Through the above processing, the input results read by the subsequent outer edge replacement and the search of the inspection area already contain the access point propagation order, hierarchical affiliation, and outer boundary position, and it is no longer necessary to recalculate the basic walking propagation relationship in subsequent steps.

[0019] S3. Taking the outer edge segment group as input, perform neighbor segment search, start and end splicing and chain position replacement on each outer edge segment within the edge node to obtain the outer edge candidate chain, and perform outer side chaining and inner side shrinking on each outer edge candidate chain to output the mutated outer edge chain group. In one specific implementation, S3 is used to perform chain-like reconstruction of the outer edge segment group within the edge node, transforming the discrete outer edge segments outside the receiving layer into a mutated outer edge chain group that can be directly read for the construction of the inspection area. The processing sequence is as follows: first, extract the neighbor segment set around each outer edge segment; then, extend outwards to both ends with the corresponding outer edge segment as the starting object to form an outer edge candidate chain; subsequently, perform conflict resolution, head-tail splicing, and inward deletion between the outer edge candidate chains to obtain a continuously expanding mutated outer edge chain group that does not penetrate into the inner region. The outer edge segment, adjacent segments, and outer edge candidate chain all use the output results of S2 and the results generated in this step. The edge node is deployed within the edge computing processing unit corresponding to the community unit, responsible for completing the local outer edge reconstruction and writing the mutated outer edge chain group into the subsequent inspection area construction process. This implementation process includes the following steps: S3-1 is used to extract the set of neighboring segments for chain-like extension reading from the outer edge segment group. Its function is to first filter out candidate connection objects compatible with each outer edge segment according to spatial connection relationship and outward expansion direction. In specific execution, the edge node reads the outer edge segment group and extracts the coordinates of the first endpoint, the coordinates of the last endpoint, the direction vector and the segment length for each outer edge segment. The direction vector points from the first endpoint to the last endpoint, and the segment length is calculated from the length of the broken line between the first endpoint and the last endpoint. Subsequently, the edge node retrieves adjacent segments that coincide with the first or last endpoint, and constructs an adjacency matrix according to the endpoint connection relationship between the outer edge segment and the adjacent segment. In the matrix, the value of the i-th and j-th bits is 1, which indicates that the i-th outer edge segment and the j-th adjacent segment have a connection relationship at the first or last endpoint; otherwise, the value is zero. Based on this, the edge node calculates the dot product of the direction vector of each adjacent segment and the direction vector of the corresponding outer edge segment, and writes the dot product value into the corresponding position of the direction matrix. Then, the adjacency matrix and the direction matrix are multiplied to obtain a two-step expansion matrix. The adjacent segments corresponding to the non-zero elements in the matrix are written into the adjacent segment set. Before writing into the adjacent segment set, the edge node continues to perform three deletion judgments on each adjacent segment: the first is the inner intersection judgment, that is, taking the side where the community unit center point is located as the inner side, if the adjacent segment has an intersection point with the inner area of ​​the corresponding outer edge segment, the corresponding adjacent segment is deleted; the second is the reverse folding judgment, that is, if the dot product of the direction vector of the adjacent segment and the direction vector of the corresponding outer edge segment is less than zero, the corresponding adjacent segment is deleted; the third is the self-intersection judgment, that is, after the adjacent segment is connected to the corresponding outer edge segment, if the newly connected polyline intersects with a non-adjacent object in the existing outer edge segment group, the corresponding adjacent segment is deleted; the adjacent segments retained after deletion are written into the adjacent segment set for S3-2 to read; when an outer edge segment does not find an adjacent segment, or the found adjacent segments are empty after deletion, the edge node records the corresponding outer edge segment as the single chain start object and writes it into the empty marker of the adjacent segment set so that S3-2 can directly output the unexpanded chain; S3-2 is used to extend outwards from the outer edge segment to both ends to form an outer edge candidate chain. Its function is to select the access object with a more suitable connection relationship with the corresponding outer edge segment from multiple neighboring segments to be accessed in rounds. In specific execution, the edge node reads the neighboring segment set, uses each outer edge segment as the initial chain, and establishes a list of neighboring segments to be accessed at the beginning and end of the initial chain respectively. For each neighboring segment to be accessed, the edge node first tries to connect the neighboring segment to one end of the current chain to form a trial connection chain, and then calculates the number of breaks, intersections, and folds of the trial connection chain. The number of reversals, the total length increment, and the sum of distances are: the number of breaks (the number of endpoints in the trial link that do not close to each other), the number of intersections (the number of times each object in the trial link intersects with a line segment of a non-adjacent object), the number of reversals (the number of times the dot product of the direction vectors of two adjacent objects in the trial link is less than zero), the total length increment (the difference between the total length of the trial link and the length of the corresponding outer edge segment), and the sum of distances (the sum of the Euclidean distance from the first endpoint of the trial link to the first endpoint of the corresponding outer edge segment and the sum of the Euclidean distance from the last endpoint of the trial link to the last endpoint of the corresponding outer edge segment). Subsequently, the edge nodes construct a five-element cost sequence based on the number of breaks, intersections, backtracking, total length increment, and distance. All trial connection chains are then sorted lexicographically in the following order: break count priority, intersection count second priority, backtracking count priority, total length increment priority, and distance and last element priority. The neighboring segment corresponding to the trial connection chain at the top of the sorted list is taken as the current round's access segment, and written into the current chain. After each round of access, the edge nodes compare the five-element cost sequence of the newly generated chain with the five-element cost sequence of the previous round item by item. If the current round's five-element cost sequence is lexicographically superior to the previous round's, then a new chain is generated. Both ends read the next round of neighboring segments to be connected. If the current round's quinary cost sequence is not lexicographically higher than the previous round's, the current round's connection result is cancelled and the extension of that end is stopped. Edge nodes process the head and tail ends synchronously according to the same rules until the head end of the current chain closes with the head end of the corresponding outer edge segment, the tail end of the current chain closes with the tail end of the corresponding outer edge segment, or no quinary cost sequence with a higher ranking is formed after all neighboring segments to be connected have been connected. The result is written as an outer edge candidate chain for S3-3 to read. When there are no neighboring segments to be connected at the head or tail end, the edge node keeps that end unchanged and only continues to perform the connection operation on the other end. S3-3 is used to organize the candidate outer edge chains into a continuous group of variant outer edge chains that do not penetrate into the inner region. Its function is to transform adjacent candidate outer edge chains into chain objects that can be directly closed in the subsequent inspection area through conflict resolution, start-end splicing, and concave deletion. In specific execution, the edge node reads the candidate outer edge chains and the outer edge segment group, replaces the corresponding outer edge segments with the candidate outer edge chains, and extracts the start endpoint, end endpoint, and chain object sequence for each chain after replacement. A chain connection matrix is ​​constructed according to the overlap relationship between the start endpoint and the end endpoint. In the matrix, the value of the i-th and j-th bits is 1, which means that the i-th chain and the j-th chain are connected end-to-end; otherwise, the value is zero. At the same time, the edge node constructs a chain conflict matrix according to the chain intersection, enclosing, and inner penetration relationships. Among them, the chain intersection is determined by whether the non-adjacent objects in any two chains intersect, the enclosing is determined by whether the closed area formed by one chain completely falls into the closed area formed by another chain, and the inner penetration is determined by whether the chain object enters the inner region of the corresponding outer edge segment. Subsequently, the edge nodes perform connectivity decomposition on the chain connection matrix, classifying interconnected chains as the same candidate chain cluster. For each candidate cluster, the chain conflict matrix is ​​read item by item. When a non-zero item exists in the chain conflict matrix, the edge nodes compare the number of objects, total length, and average distance to the community unit center point of the corresponding conflicting chain, deleting the conflicting chain with more objects. If the number of objects is the same, the conflicting chain with the longer total length is deleted; if the total length is the same, the conflicting chain with the shorter average distance is deleted. After deleting conflicting chains, the edge nodes perform head-to-tail concatenation on the remaining chains to form a merged chain result. Then, the edge nodes calculate the cosine of the angle between adjacent objects for each object in the merged chain result. The edge node calculates the perpendicular distance from the two endpoints of the object to the line connecting the first and last endpoints of the chain result. When the cosine of the included angle is less than zero, and the first and last ends of adjacent objects can still be directly connected after deleting the object, the corresponding object is recorded as an indented object and deleted from the chain result. After completing one round of deletion, the edge node reconstructs the chain connection matrix and the chain conflict matrix, and repeats the conflict resolution, first and last end splicing and indented deletion until the chain connection matrices of two adjacent rounds are identical and the chain conflict matrix is ​​all zero. The final result is written as a mutated outer edge chain group for S4 to read. When the candidate chain cluster contains only one chain and the chain conflict matrix is ​​all zero, the edge node directly writes the corresponding chain into the mutated outer edge chain group. Through the above processing, the mutated outer edge chain group output by the edge node is no longer a set of discrete outer edge segments directly read from the outside of the bearing layer, but a continuous chain object after neighbor segment screening, chain extension, conflict resolution and concave deletion. When constructing the inspection area, each mutated outer edge chain and its inner bearing layer can be directly closed to form the inspection area, thereby reducing the incorrect construction of the inspection area caused by outer edge breakage, outer edge folding and chain penetration. In practical applications: When there are pedestrianized commercial interfaces, diagonal open passages, and cross-street connecting spaces on the periphery of a community unit, the outer edge segment group output by S2 often contains multiple outer edge segments that are not closed at the beginning and end and have local foldbacks. The edge nodes first filter out adjacent segments with the same connection direction as each outer edge segment through the adjacency matrix and direction matrix, and then select access objects in round by round using a five-element cost sequence to form an outer edge candidate chain. Subsequently, the intersection, enclosing, and inner penetration relationships between the outer edge candidate chains are deleted, the beginning and end of the retained chains are spliced, and the concave objects in the chain-joining results are deleted. Finally, a group of mutated outer edge chains that are continuously distributed along the outside of the community unit and do not penetrate into the inner area is obtained, which is used for the next step to continue to build the area to be inspected and perform the judgment of unsearched segments in the edge computing environment.

[0020] S4. Using the mutated outer edge chain group, the receiving layer and the access point as input, each mutated outer edge chain and its inner receiving layer are enclosed in the inspection area within the edge node. Then, starting from the access point, each segment in the inspection area is searched segment by segment along the receiving layer. The mutated outer edge chains with unsearched segments are deleted. The remaining mutated outer edge chains are retained and the replacement, chain merging, shrinking and searching are repeated. The target outer edge chain is output. In one specific implementation, S4 is used to determine within the edge node whether a mutated outer edge chain can be retained as the outer edge of the living circle. The processing principle is to close each mutated outer edge chain with its inner supporting layer to form an inspection area. Then, starting from the access point, it searches segment by segment along the supporting layer for all objects within the inspection area. If there are objects within the inspection area that cannot be continuously reached from the access point through the supporting layer, the corresponding mutated outer edge chain does not participate in the outer edge retention. If all objects within the inspection area can be found, the corresponding mutated outer edge chain enters the subsequent adjudication and reconstruction process. The mutated outer edge chain group uses the output result of S3, and the supporting layer and access point use the output results of S2 and S1. The edge node is deployed within the edge computing processing unit corresponding to the community unit, responsible for completing the construction of the inspection area, the search within the area, the chain retention adjudication, and the output of the target outer edge chain. This implementation process includes the following steps: S4-1 is used to construct the corresponding region to be inspected and its search input for each mutated outer edge chain. Its function is to convert the chain-like outer edge result into an intra-region structure that can perform a connectivity search. In specific execution, the edge node reads the mutated outer edge chain group, the connecting layer, and the access point. For each mutated outer edge chain, it extracts the chain segment endpoint sequence and determines the first and last endpoints according to the chain segment order. Subsequently, it retrieves the connecting layer objects directly connected to the first endpoint and the connecting layer objects directly connected to the last endpoint from the connecting layer. Using the corresponding objects as the starting and ending objects, it continuously traces along the beginning and end of the connecting layer to obtain the region located inside the mutated outer edge chain and connected to the first endpoint. A set of receiving layer objects whose endpoints and tail endpoints are connected is denoted as the receiving layer segment set. When there are multiple receiving paths with consecutive beginnings and ends, the edge nodes first compare the number of objects in each receiving path and retain the one with fewer objects. When the number of objects is the same, the total length of each receiving path is compared and the one with the shorter total length is retained. After the receiving layer segment set is determined, the edge nodes close the variable outer edge chain and the receiving layer segment set according to the beginning and end endpoints to form the inspection area, and sequentially number all road segments, crossings, passage segments, open segments, variable outer edge chain objects and receiving layer objects in the inspection area. Subsequently, the edge nodes construct an intra-region adjacency matrix according to the head-to-tail connection relationship of each object in the inspection area. The rows and columns of the matrix are arranged in numerical order. The value of the i-th and j-th bits is one, which indicates that the i-th object and the j-th object have a head-to-tail connection relationship. Otherwise, the value is zero. Next, the edge node reads the access point one by one and determines whether the access point coincides with the first and last endpoints of an object in the inspection area, or is located inside the corresponding object. If either condition is met, the corresponding position of the starting vector is recorded as one; otherwise, it is recorded as zero. The starting vector is obtained, and the inspection area, the adjacency matrix within the area, and the starting vector are written into the search cache for S4-2 to read. When a certain mutated outer edge chain does not find a set of connecting segments that are continuously connected to the first and last endpoints, the edge node directly records the corresponding mutated outer edge chain as a chain to be deleted and does not construct an inspection area. S4-2 is used to search for continuously reachable objects within the inspection area starting from the access point. Its function is to expand the head-to-tail connection relationship in the inspection area into verifiable arrival results. Specifically, the edge node reads the inspection area, the intra-area adjacency matrix, and the starting vector, using the starting vector as the first-round search vector and the first-round cumulative search vector. Then, the first-round search vector is multiplied by the intra-area adjacency matrix to obtain the next-round search vector. The next-round search vector is then binarized bit by bit, recording non-zero values ​​as 1 and zero values ​​as 0. Finally, the binarized next-round search vector is added bit by bit to the previous round cumulative search vector, and the result is binarized again. Non-zero values ​​are marked as 1, and zero values ​​are marked as 0, resulting in a new round of cumulative search vector. After obtaining the new round of cumulative search vector, the edge node continues to use the binarized next round of search vector as the new search vector, repeating left multiplication, bit-by-bit addition, and binarization. When the new round of cumulative search vector is identical to the previous round of cumulative search vector in every bit comparison, the edge node stops iterating and writes the objects corresponding to the zero values ​​in the cumulative search vector as the unsearched segment set. If all positions in the cumulative search vector are 1, the edge node records the unsearched segment set as an empty set. If all positions in the initial vector are zero, the edge node directly writes all objects in the area to be inspected as the unsearched segment set. Finally, the edge nodes write the cumulative search vector and the set of unsearched segments into the adjudication cache for S4-3 to read; within the inspection area where a closed loop exists, the above iterative process stops after the cumulative search vector stabilizes bit by bit, and no longer continues to expand repeatedly; S4-3 is used to delete and sort mutated outer edge chains based on the search results of the area to be inspected. Its function is to first exclude mutated outer edge chains with broken connecting objects inside, and then determine the subsequent reconstruction order of the remaining mutated outer edge chains. In specific execution, the edge node reads the cumulative search vector, the set of unsearched segments, and the group of mutated outer edge chains, and checks whether the set of unsearched segments of each mutated outer edge chain is empty. When the set of unsearched segments is not empty, the edge node deletes the corresponding mutated outer edge chain and writes the deletion result into the chain adjudication record. When the set of unsearched segments is empty, the edge node retains the corresponding mutated outer edge chain and continues to count the total number of segments in the area to be inspected, the number of non-zero bits in the cumulative search vector, and the number of connected components in the set of unsearched segments. The total number of segments is the total number of numbers of all objects in the area to be inspected, the number of non-zero bits is the number of positions with a value of 1 in the cumulative search vector, and the number of connected components is obtained by reconstructing the graph according to the head-tail connection relationship inside the set of unsearched segments and then performing a connected search. Subsequently, the edge nodes subtract the non-zero bits from the total number of segments to obtain the number of missing segments, and form a decision sequence by combining the number of missing segments and the number of connected components. After the decision sequences of all retained chains are generated, the edge nodes first sort them by the number of missing segments in ascending order, and then sort them by the number of connected components in ascending order if the number of missing segments is the same, to obtain the arrangement order of the retained chain groups. The results after the order is arranged are written as the retained chain groups and decision sequences for S4-4 to read. When there is only one mutated outer edge chain after retention, the edge nodes directly write the mutated outer edge chain to the beginning of the retained chain group. When all mutated outer edge chains are deleted, the edge nodes record the retained chain group as an empty group and the decision sequence as an empty sequence. S4-4 is used to repeatedly reconstruct the retained chain group and output the target outer edge chain. Its function is to correct the retained chains one by one along the adjudication order, so that the retained outer edge chains can converge stably after repeated replacement, chain merging, shrinking and inspection of the inspection area. In specific execution, the edge node reads the retained chain group, adjudication sequence, connecting layer and access point, reads each retained chain one by one in the order of adjudication sequence, and, based on the currently read retained chain, re-invokes the replacement, chain merging and shrinking processing rules in S3 to obtain the updated retained chain. Subsequently, the edge node re-executes the inspection area reconstruction, intra-area adjacency matrix reconstruction and starting vector reset for the updated retained chain. Among them, the inspection area reconstruction, intra-area adjacency matrix reconstruction and starting vector reset all follow the construction rules of S4-1. After completing the above reconstruction, the edge node re-executes the cumulative search vector recalculation and the unsearched segment set recalculation. Among them, the cumulative search vector recalculation and the unsearched segment set recalculation follow the search rules of S4-2. Next, the edge node recalculates the number of missing segments and connected components based on the recalculation results, generates the current round of adjudication sequence, and compares the current round of adjudication sequence with the previous round of adjudication sequence item by item. When the number of missing segments and connected components at corresponding positions are all the same in the two rounds of adjudication sequences, the edge node stops repeating the processing of the current read retain chain and writes the current read retain chain into the target outer edge chain. When the values ​​at any position in the two rounds of adjudication sequences are different, the edge node continues to repeat the replacement, chain merging, shrinking, inspection area reconstruction, intra-area adjacency matrix reconstruction, starting vector reset, cumulative search vector recalculation, and unsearched segment set recalculation for the current read retain chain. After all retain chain processing is completed, the edge node writes the result as the target outer edge chain for S5 to read. When the retain chain group is empty, the edge node directly outputs the empty target outer edge chain and writes an empty chain mark to S5. Through the above processing, the target outer edge chain output by the edge node is no longer just a continuous geometric chain on the outside, but an outer edge result after internal connectivity checks, missing segment adjudication, and repeated reconstruction. Objects that cannot be continuously reached from the access point through the receiving layer will not be included in the corresponding outer edge chain. Therefore, the living circle boundary obtained by subsequent aggregation nodes can correspond to the spatial range that the community unit gradually receives outwards. In practical applications: when there are closed courtyards, isolation green belts, inaccessible spaces under overpasses, and partially open passages around the community unit, the variant outer edge chain output by S3 may have formed a continuous outer edge on the map, but the edge In S4, each node first closes each mutated outer edge chain with its inner connecting layer to form an inspection area. Then, it performs a matrix iterative search along the connecting layer starting from the access point. If there is a road segment, passage segment, or open segment in an inspection area that is blocked by a closed wall or isolation green belt and cannot be continuously reached from the access point, the corresponding mutated outer edge chain is deleted when the unsearched segment set is not empty. For the mutated outer edge chains that are not deleted, the edge nodes continue to sort them according to the number of missing segments and the number of connected components, and repeatedly perform replacement, chain merging, shrinking, and inspection area re-examination on the retained chains until the decision sequence remains unchanged in the two rounds before and after, and finally obtain the target outer edge chain for the convergence node to close and connect.

[0021] S5. Input the target outer edge chains output by each edge node into the convergence node. In the convergence node, splice the target outer edge chains that are connected end to end, remove duplicates from the overlapping parts, and perform closed connection on the splicing results to output the boundary of the city's 15-minute living circle. In one specific implementation, S5 is used to perform cross-node splicing, overlap reduction, and closed connection of the target outer edge chains output by each edge node within the convergence node to obtain the boundary of the city's 15-minute living circle. Specifically, the convergence node reads the target outer edge chains and corresponding chain segment sequences returned by each edge node. For the target outer edge chains distributed across edge nodes, it first restores the continuous chain according to the endpoint connection relationship, then deletes duplicate parts according to the chain segment coordinate overlap relationship, and finally, it adds connection segments to the unclosed deduplicated spliced ​​chain groups and deletes duplicate closed chains, so that the output result corresponds to the complete boundary formed segment by segment outward from the community unit. This implementation process includes the following steps: S5-1 is used to splice the target outer edge chains scattered in different edge nodes into a continuous chain and delete overlapping parts. Its function is to first restore the boundary continuity relationship that is broken across edge nodes, and then remove the chain segments that are repeatedly written. In specific execution, the sink node reads the target outer edge chains output by each edge node, extracts the first endpoint, the last endpoint, and the chain segment sequence for each target outer edge chain, and constructs an endpoint connection matrix according to whether the sum of the squares of the coordinate differences of all first endpoints and last endpoints is not greater than the square value of the coordinate coincidence determination precision. The i-th and j-th bits of the matrix are set to one, which means that the last endpoint of the i-th target outer edge chain coincides with the first endpoint of the j-th target outer edge chain. Otherwise, the value is zero. Subsequently, the pooling node constructs a coincident segment matrix based on whether the coordinates of the first and last endpoints of each segment in the outer edge chain of the target are respectively coincident, or whether the first and last endpoints of one segment coincide with the last and first endpoints of another segment. In the matrix, the value of the m-th and n-th bits is 1, indicating that the m-th segment and the n-th segment are coincident; otherwise, the value is zero. After the endpoint connection matrix and the coincident segment matrix are constructed, the pooling node performs connectivity decomposition on the endpoint connection matrix, dividing the target outer edge chains that can be continuously connected end to end into the same splicing chain group. Within each splicing chain group, the target outer edge chains are spliced ​​sequentially according to the end-to-end connection order indicated by the endpoint connection matrix to obtain the splicing result. Then, the aggregation node reads the overlapping segment matrix corresponding to the splicing result item by item. For the overlapping chain segments, it retains the chain segment written for the first time and deletes the chain segments written repeatedly in the future to obtain the deduplicated splicing chain group. The deduplicated splicing chain group is written to the closure processing buffer for S5-2 to read. When a target outer edge chain does not form a connection relationship with other target outer edge chains in the endpoint connection matrix, the aggregation node writes the corresponding target outer edge chain as a deduplicated splicing chain group. When there are multiple first and last sorted paths in the same splicing chain group, the aggregation node first compares the total number of chain segments of each sorted path and retains the one with the smaller total number of chain segments. If the total number of chain segments is the same, the aggregation node compares the total length of each sorted path and retains the one with the shorter total length. S5-2 is used to close the deduplicated splicing chain group and output the boundary of the city's 15-minute living circle. Its function is to fill in the endpoint connections of the unclosed splicing chain group and delete the closed chain that is repeatedly surrounded by the outer closed chain. In specific execution, the aggregation node reads the deduplicated splicing chain group, extracts the first endpoint and the last endpoint for each deduplicated splicing chain group, and searches for the connection segment connecting the first endpoint and the last endpoint from the chain segment set that has not yet been written into the current deduplicated splicing chain group. The search rule for the connection segment is as follows: first, search for candidate connection paths that are continuous from beginning to end and do not intersect with the existing chain segments in the current deduplicated splicing chain group without non-endpoint intersections; then compare the total length of each candidate connection path and retain the one with the shorter total length; if the total lengths are the same, retain the one with fewer chain segments. After the connection segment is determined, the aggregation node writes the connection segment into the corresponding deduplicated splicing chain group to form a closed chain. Subsequently, the aggregation node calculates the closed region for each closed chain and compares them according to the inclusion relationship between the closed chains. When the closed region formed by one closed chain completely falls within the closed region formed by another closed chain, and the two closed chains contain overlapping chain segments, the closed chain located inside the outer closed chain is deleted, and the outer closed chain is retained. After the duplicate closed chains are deleted, the aggregation node outputs the retained closed chains in the order of continuous beginning and end, thus obtaining the boundary of the city's 15-minute living circle. When the beginning and end endpoints of a deduplicated spliced ​​chain group coincide, the aggregation node no longer searches for connecting segments and directly writes the corresponding deduplicated spliced ​​chain group as a closed chain. When a deduplicated spliced ​​chain group does not find a connecting segment that satisfies the condition of continuous beginning and end and does not intersect with existing chain segments at non-endpoints, the aggregation node records the corresponding deduplicated spliced ​​chain group as an unclosed chain and returns to the previous round of spliced ​​chain groups, reselects the second shortest sorting path in total length, and performs the connecting segment search again. Through the above processing, the boundary of the urban 15-minute living circle output by the aggregation node is no longer a simple parallel set of local results from the edge nodes, but rather an overall boundary result after cross-node splicing, duplicate chain deletion, head-tail closure, and duplicate closed chain removal. This can restore the target outer edge chains formed within each edge node to a continuously distributed outer edge structure around the community unit. In practical applications: when the space surrounding the community unit is processed by multiple edge nodes, and different edge nodes output a portion of the target outer edge chain at the intersection, the aggregation node first splices the target outer edge chains across edge nodes into a continuous chain based on the coincidence of the head and tail endpoints, and then deletes the chain segments repeatedly written at the intersection based on the coincidence of the chain segment coordinates. For the deduplicated spliced ​​chain group that is still not closed, the aggregation node continues to search for the connecting segment between the head and tail endpoints in the unwritten chain segments and writes the connecting segment into the deduplicated spliced ​​chain group to form a closed chain. When multiple closed chains have an inner and outer enclosing relationship and duplicate chain segments exist, the aggregation node deletes the closed chain located inside the outer closed chain, and finally outputs the urban 15-minute living circle edge formed around the community unit.

[0022] Working principle: This scheme first reads the road segments, crossing points, passage segments, open segments, entry edges, and blocking segments around the community unit within the edge nodes, establishes an edge graph, and cuts off the connections separated by the blocking segments. Then, it extracts the access point from the outer boundary of the community unit. Subsequently, starting from the access point, it accumulates the walking time segment by segment along the edge graph to form the receiving layer and outer edge segment group. On this basis, it performs neighbor segment search, chain replacement, chain merging, and shrinking on the outer edge segment group to obtain the variant outer edge chain group. Then, it closes the variant outer edge chain with the inner receiving layer to form the inspection area. Starting from the access point, it searches for each object in the inspection area, deletes the variant outer edge chain that includes the unsearched object, and repeatedly corrects the retained chain to obtain the target outer edge chain. Finally, the convergence node performs splicing, deduplication, and closure connection on the target outer edge chains output by each edge node to form the boundary of the city's 15-minute living circle. Therefore, the output boundary is not obtained directly by geometric outlining, but by the actual entry relationship of the community unit receiving segments outward. For example, in an old urban area, when a community unit is surrounded by enclosed courtyards, railway barriers, pedestrian walkways, crossroads, and open plazas, the edge nodes first connect the roads and passageways into an edge graph and delete connections blocked by walls, barriers, or enclosed boundaries. Then, they extract access points from the community unit boundary that can actually enter the road network. Next, they expand outward layer by layer according to walking time to find the outer edge segments and reconstruct the scattered outer edge segments into continuous outer edge chains. If there are roads or open spaces within the area enclosed by an outer edge chain that are located inside the outer edge but cannot be continuously entered from the community unit, then that outer edge chain will be deleted. Only outer edge chains that can be reached step by step from the access points through the connecting layers will be retained. Finally, the target outer edge chains retained by each edge node are spliced ​​and closed in the convergence node to obtain the living circle boundary corresponding to the actual walking range of residents.

[0023] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for constructing the boundary of a 15-minute urban living circle based on multi-constraint paths, characterized in that, include: S1. Input community units, road segments, pedestrian crossings, passage segments, open segments, entry edges, and barrier segments into edge nodes. Generate an edge graph within the edge nodes according to the connection relationship, and cut off the corresponding connection with the barrier segment. Output the edge graph and access point. S2. Using the access point and edge map as input, accumulate the walking time segment by segment along the edge map within the edge node. Write the road segments, crossing points, passage segments and open segments with a cumulative walking time of no more than 15 minutes that are directly connected to the previous segment into the receiving layer in sequence, and output the receiving layer and the outer edge segment group. S3. Taking the outer edge segment group as input, perform neighbor segment search, start and end splicing and chain position replacement on each outer edge segment within the edge node to obtain the outer edge candidate chain, and perform outer side chaining and inner side shrinking on each outer edge candidate chain to output the mutated outer edge chain group. S4. Using the mutated outer edge chain group, the receiving layer and the access point as input, each mutated outer edge chain and its inner receiving layer are enclosed in the inspection area within the edge node. Then, starting from the access point, each segment in the inspection area is searched segment by segment along the receiving layer. The mutated outer edge chains with unsearched segments are deleted. The remaining mutated outer edge chains are retained and the replacement, chain merging, shrinking and searching are repeated. The target outer edge chain is output. S5. Input the target outer edge chains output by each edge node into the convergence node. In the convergence node, splice the target outer edge chains that are connected end to end, remove duplicates from the overlapping parts, and perform closed connection on the splicing result to output the boundary of the city's 15-minute living circle.

2. The method for constructing the boundary of a 15-minute urban living circle based on multi-constraint paths according to claim 1, characterized in that: S1 includes: S1-1. Input the road segment, crossing point, passage segment, open segment and entry edge into the edge node. Within the edge node, take the endpoints of each segment as the start and end points, the crossing point as the cross-segment connection point, and the entry edge as the segment side access edge. First, establish the start and end connection for the road segment, passage segment and open segment with overlapping endpoints. Then, establish the cross-point connection for each segment connected on both sides of the crossing point. Finally, establish the segment edge connection for the open segment or passage segment that intersects with the entry edge. Output the initial edge graph. S1-2. Input the blocking segment and the initial edge graph into the edge node. In the edge node, check whether each connection passes through the blocking segment, intersects with the blocking segment, or whether the two ends of the connection are on opposite sides of the blocking segment. For connections that meet any of the conditions, cut them off. Number each set of independent connected segments formed after the cut off, and output the split edge graph and the set of connected segments. S1-3. Input the community unit, the segmented edge graph, and the connected segment set into the edge node. Search for the road segment, passage segment, and open segment that intersect with the outer boundary of the community unit within the edge node. Record the intersection point of each segment located on the outer boundary of the community unit as the initial access point. Then retain the initial access point that is in the same connected segment set as the initial access point and is not separated by the blocking segment as the access point. Output the edge graph and the access point.

3. The method for constructing the boundary of a 15-minute urban living circle based on multi-constraint paths according to claim 2, characterized in that: S2 includes: S2-1. Input the access point and edge graph into the edge node. Within the edge node, search outwards along the edge graph segment by segment, starting from each access point. For each search path, accumulate the segment length and the crossing time to form the path time. Record the minimum time of each segment in each search path as the arrival time and output the arrival time segment set. S2-2. Input the arrival time segment set into the edge node. Read each segment in the edge node in order of arrival time from smallest to largest. Write the road segment, crossing point, passage segment and open segment with an arrival time of no more than 15 minutes and whose preceding connected segment has been written into the corresponding receiving layer. Write the segments with the same arrival time into the same receiving layer and output the receiving layer sequence. S2-3. Input the receiving layer sequence into the edge node. In the edge node, for each receiving layer, search for each segment that is directly connected to a segment that is not in this receiving layer. Record each road segment, crossing point, passage segment and open segment that is directly connected to the next receiving layer or a segment that has not been written and is located outside the current receiving layer as the outer edge segment. Output the receiving layer and the outer edge segment group.

4. The method for constructing the boundary of a 15-minute urban living circle based on multi-constraint paths according to claim 3, characterized in that: S3 includes: S3-1. Input the outer edge segment group into the edge node. Extract the coordinates of the two endpoints, direction vector and segment length for each outer edge segment within the edge node. Search for the adjacent segments that connect to the two endpoints of each outer edge segment. Construct an adjacency matrix according to the endpoint connection relationship between adjacent segments. Construct a direction matrix according to the dot product of the direction vector of the adjacent segment and the direction vector of the corresponding outer edge segment. Perform matrix multiplication on the adjacency matrix and the direction matrix to obtain a two-step expansion matrix. Use the adjacent segments corresponding to the non-zero elements in the two-step expansion matrix as the neighbor segment set. Delete the adjacent segments that intersect the inner side of the corresponding outer edge segment, form a reverse fold with the corresponding outer edge segment, or generate self-intersection after connection. Output the neighbor segment set.

5. The method for constructing the boundary of a 15-minute urban living circle based on multi-constraint paths according to claim 4, characterized in that: S3 further includes: S3-2. Input the neighbor segment set into the edge node. Within the edge node, take each outer edge segment as the initial chain. Select neighbor segments to be connected from both ends of the initial chain. For each neighbor segment to be connected, calculate the number of breaks, crosses, backtracks, total length increment, and the sum of distances from the chain endpoints to the two ends of the corresponding outer edge segment after connection. Construct a five-element cost sequence using the number of breaks, crosses, backtracks, total length increment, and distance sum. Select the connected segments in rounds according to the lexicographical order of break number priority, crosses second priority, backtracks third priority, total length increment third priority, and last element of distance sum priority. After each round of connection, compare the five-element cost sequence of the newly generated chain with the five-element cost sequence of the previous round item by item. Repeat the two-end connection until the two ends of the current chain reach the two ends of the corresponding outer edge segment, or the five-element cost sequence after all neighbor segments to be connected are connected is not better than the current chain. Output the outer edge candidate chain.

6. The method for constructing the boundary of a 15-minute urban living circle based on multi-constraint paths according to claim 5, characterized in that: S3 further includes: S3-3. Input the candidate outer edge chains and the outer edge segment group into the edge node. Replace the corresponding outer edge segment with the candidate outer edge chain within the edge node. Construct a chain connection matrix for each replaced chain according to the head-to-tail connection relationship. Construct a chain conflict matrix according to the chain intersection, enclosing, and inner penetration relationships. Perform connectivity decomposition on the chain connection matrix to obtain candidate chain clusters. After resolving the conflict chains in the chain conflict matrix of each candidate chain cluster, perform head-to-tail splicing to obtain the chain-joined result. Then, calculate the cosine value of the included angle between two adjacent segments and the perpendicular distance from the endpoint of each segment to the head-to-tail connection line for each chain-joined result. Delete the concave segments with negative included angle cosine values ​​and which are still connected head-to-tail after deletion. Repeat the conflict resolution, head-to-tail splicing, and concave segment deletion for the deleted chains until the chain connection matrix no longer changes and the chain conflict matrix is ​​all zero. Output the mutated outer edge chain group.

7. The method for constructing the boundary of a 15-minute urban living circle based on multi-constraint paths according to claim 6, characterized in that: S4 includes: S4-1. Input the mutated outer edge chain group, the receiving layer, and the access point into the edge node. Extract the chain segment endpoint sequence for each mutated outer edge chain within the edge node. Select the receiving layer segment set located inside each mutated outer edge chain and connected to the beginning and end of each mutated outer edge chain. Close the beginning and end of each mutated outer edge chain and the receiving layer segment set to form the inspection area. Construct the intra-area adjacency matrix according to the beginning and end connection relationship of each segment in the inspection area. Construct the starting vector according to the case where the access point falls into each segment in the inspection area. Output the inspection area, the intra-area adjacency matrix, and the starting vector.

8. The method for constructing the boundary of a 15-minute urban living circle based on multi-constraint paths according to claim 7, characterized in that: S4 further includes: S4-2. Input the region to be inspected, the adjacency matrix within the region, and the starting vector into the edge node. Within the edge node, use the starting vector as the first-round search vector. Multiply the first-round search vector by the adjacency matrix within the region to obtain the next-round search vector. Record the non-zero values ​​of the next-round search vector as 1 and the zero values ​​as 0, and add it bit by bit to the cumulative search vector of the previous round. Record the non-zero bits in the addition result as 1 and the zero bits as 0 to obtain the new cumulative search vector. Repeat the left multiplication, bit by bit addition, and binarization until the new cumulative search vector has the same values ​​as the previous cumulative search vector. Output the cumulative search vector and the set of unsearched segments. S4-3. Input the cumulative search vector, the set of unsearched segments, and the group of mutated outer edges into the edge node. Delete mutated outer edges that are not empty in the set of unsearched segments within the edge node. Count the total number of segments in the inspection area of ​​each mutated outer edge after retention, the number of non-zero bits in the cumulative search vector, and the number of connected components in the set of unsearched segments. Subtract the non-zero bits from the total number of segments to obtain the number of missing segments. Form an adjudication sequence using the number of missing segments and the number of connected components. Arrange the mutated outer edges after retention in ascending order of the number of missing segments and the number of connected components. Output the group of retained chains and the adjudication sequence.

9. The method for constructing the boundary of a 15-minute urban living circle based on multi-constraint paths according to claim 8, characterized in that: S4 further includes: S4-4. Input the retained chain group, adjudication sequence, receiving layer and access point into the edge node. Read the retained chain group one by one in the order of adjudication sequence within the edge node. Repeat the replacement, chain merging, shrinking, inspection area reconstruction, intra-area adjacency matrix reconstruction, starting vector reset, cumulative search vector recalculation and unsearched segment set recalculation for each retained chain. Stop when the values ​​of each item in the adjudication sequence of two adjacent rounds are consistent, and output the target outer edge chain.

10. The method for constructing the boundary of a 15-minute urban living circle based on multi-constraint paths according to claim 9, characterized in that: S5 includes: S5-1. Input the target outer edge chains output by each edge node into the convergence node. Extract the first endpoint, the last endpoint, and the chain segment sequence of each target outer edge chain in the convergence node. Construct the endpoint connection matrix according to the overlap relationship between the first endpoint and the last endpoint. Construct the overlapping segment matrix according to the overlap relationship of the chain segment coordinates. Perform connectivity decomposition on the endpoint connection matrix to obtain spliced ​​chain groups. In each spliced ​​chain group, splice the target outer edge chains in the order of first and last connection. Delete the overlapping parts according to the overlapping segment matrix. Output the deduplicated spliced ​​chain group. S5-2. Input the deduplication splicing chain group into the aggregation node. Calculate the connection segment between the first and last endpoints for each deduplication splicing chain group within the aggregation node. Write the connection segment into the corresponding deduplication splicing chain group to form a closed chain. Then, delete the duplicate closed chains located inside the outer closed chains according to the inclusion relationship between the closed chains. Output the remaining closed chains in the order of continuous first and last, to obtain the boundary of the city's 15-minute living circle.

Images