Three-dimensional road curve and surface model construction method based on two-dimensional stock mapping data

Abstract

The application discloses a kind of three-dimensional road curve, surface model construction method based on two-dimensional stock surveying and mapping data, comprising: first, extract road skeleton line and attribute information from two-dimensional stock surveying and mapping data, and extract elevation information to give skeleton line.Then, the abnormal elevation data of skeleton line section is corrected using bidirectional sniffing repair algorithm.After that, it is judged whether topological anomaly exists at road section connection, if exists, then according to the type to which topological anomaly case belongs, the appropriate correction method is selected to correct, so that the skeleton line with normal topology is obtained.Then, it is judged whether elevation anomaly exists at road section connection, if exists, then according to the skeleton line node distribution situation of different road at road section connection, the abnormal elevation is corrected.Finally, three-dimensional road curve model is constructed according to the corrected skeleton line.In this way, the road elevation fitting effect and elevation data precision of three-dimensional road curve model constructed based on road skeleton line can be ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of three-dimensional modeling of geographic models, and in particular to a three-dimensional road curve and surface model construction method based on two-dimensional stock surveying and mapping data. BACKGROUND

[0002] Compared with the traditional two-dimensional road element data mode, road three-dimension can fully combine the characteristics of roads such as horizontal, vertical and horizontal, and solve the problems of not being intuitive, real and accurate in the traditional road expression mode for a long time. The three-dimensional road data mode based on three-dimensional scene has stronger interactivity, easier analysis, stronger realism and more concrete expression ability. The conventional fine three-dimensional road model or oblique photography three-dimensional road model has long modeling period, high cost, large data volume and great updating difficulty.

[0003] Elevation anomaly is a key problem and difficulty that cannot be avoided in the process of road three-dimension. In the process of road three-dimension, the distribution and number of road nodes, the accuracy of DEM data and the difference of road elevation collection method may cause problems such as elevation anomaly and topological anomaly of connecting place of three-dimensional road, which reduces the modeling accuracy and fitting effect of three-dimensional road curve model. SUMMARY

[0004] In view of the deficiencies in the prior art, the present application provides a three-dimensional road curve and surface model construction method based on two-dimensional stock surveying and mapping data, which can correct the elevation anomaly of the road skeleton line and improve the modeling accuracy and fitting effect of the three-dimensional road curve model. The specific technical solutions are as follows:

[0005] In a first aspect, a three-dimensional road curve model construction method is provided. In a first implementation manner of the first aspect, the method comprises:

[0006] Extracting a skeleton line and attribute information of a road element from two-dimensional stock surveying and mapping data, and assigning the elevation information of the road element to the skeleton line according to the attribute information;

[0007] Identifying the elevation anomaly nodes of the skeleton line section, and correcting them by using a bidirectional sniffing repair algorithm;

[0008] Judging whether the skeleton line after elevation correction has topological anomaly at the connecting place of the road section, and in response to the existence of topological anomaly, selecting a corresponding correction method according to the topological anomaly condition for correction;

[0009] Judging whether the skeleton line after topological correction has elevation anomaly at the connecting place of the road section, and in response to the existence of elevation anomaly, correcting the abnormal elevation according to the distribution of the skeleton line nodes at the connecting place of the road section;

[0010] The three-dimensional road curve model of the road element is constructed based on the corrected skeleton line.

[0011] In the second implementation manner of the first aspect, the skeleton line of the road element is extracted, including:

[0012] According to the element form type of the road element, an adaptive extraction tool is selected to extract the skeleton line of the road element from the two-dimensional stock surveying and mapping data.

[0013] In the third implementation manner of the first aspect, the elevation information of the road element is extracted according to the attribute information and is given to the skeleton line, including:

[0014] The skeleton line of the road element is subjected to a densification operation, and the elevation information is given to the skeleton line after the densification.

[0015] In the fourth implementation manner of the first aspect, the elevation data of the elevation abnormal node is corrected by using a bidirectional sniffing repair algorithm, including:

[0016] Taking the elevation abnormal node as a reference point, the nearest elevation normal nodes in two directions of the road element are searched respectively from the reference point;

[0017] The slope between the two elevation normal nodes is calculated, and the slope is compared with a slope threshold value;

[0018] In response to the slope being less than the slope threshold value, an interpolation algorithm is used to generate the elevation data of all nodes between the two elevation normal nodes;

[0019] In response to the slope exceeding the slope threshold value, the former elevation normal node is taken as a fixed point, and in the direction from the fixed point to the latter elevation normal node, an elevation normal node with a slope less than the slope threshold value between the fixed point and the elevation normal node is searched, and the planar distance between the searched elevation normal node and the fixed point is calculated;

[0020] The planar distance is compared with a distance threshold value, and in response to the planar distance being less than the distance threshold value, an interpolation algorithm is used to generate the elevation data of all nodes between the former elevation normal node and the searched node;

[0021] In response to the planar distance exceeding the distance threshold value, manual intervention is performed.

[0022] In the fifth implementation manner of the first aspect, a corresponding correction method is selected according to a topological abnormality to correct the topological abnormality, including:

[0023] when the skeleton line node of the low-grade road does not reach the coverage range of the high-grade road, a buffer zone is generated with the skeleton line node as the origin and a set topological distance as the radius;

[0024] determining whether there is an intersection between the buffer zone and the high-grade road;

[0025] in response to the existence of the intersection, selecting an adaptive intersection from all intersections as a new node of the skeleton line of the low-grade road at the road segment connectivity;

[0026] in response to the non-existence of the intersection, manually checking the road element and the surveying and mapping data of the high-grade road.

[0027] when the skeleton line node is located within the coverage range of the high-grade road, the skeleton line node is reversely trimmed along the skeleton line to the boundary of the coverage range;

[0028] determining whether there is a skeleton line node of the skeleton line of the high-grade road at the boundary;

[0029] in response to the non-existence, adding a new node at the boundary and assigning the new node with elevation data;

[0030] in response to the existence, taking the skeleton line node as a new node of the skeleton line of the low-grade road.

[0031] In a sixth implementation manner of the first aspect, in the first implementation manner of the first aspect, the determining whether the skeleton line after the topological correction has an elevation abnormality at the road segment connectivity comprises:

[0032] performing consistency determination on the elevation data of the skeleton line of each connectivity road at the road segment connectivity, and if the elevation data of different connectivity roads is inconsistent, determining whether the skeleton line has an elevation abnormality at the road segment connectivity.

[0033] In a seventh implementation manner of the first aspect, in the first implementation manner of the first aspect, the correcting the abnormal elevation according to the distribution of the skeleton line node at the road segment connectivity comprises:

[0034] when the skeleton line node exists at the road segment connectivity for each connectivity road, determining whether each skeleton line node satisfies a slope condition;

[0035] in response to the non-satisfaction of the slope condition, calculating a repaired elevation by using a bidirectional sniffing repair algorithm;

[0036] assigning the repaired elevation to the skeleton line node of the low-grade road, and re-determining whether the skeleton line node satisfies the slope condition;

[0037] in response to the satisfaction of the slope condition, completing the correction;

[0038] In response to the slope condition not being met, the elevation value range corresponding to each skeleton line node under the road slope requirement is obtained;

[0039] The elevation value closest to the elevation of the high-grade road node is selected from the intersection of the elevation value ranges of different skeleton line nodes and is assigned to the low-grade road node;

[0040] When there is only one connected road or no connected road at the road segment connection, a node is added at the road segment connection, and the elevation data at the location is assigned to the added node;

[0041] It is judged whether the elevation of the added node meets the slope condition, and in response to the slope condition not being met, a repaired elevation is calculated using a bidirectional sniffing repair algorithm;

[0042] The repaired elevation is assigned to the skeleton line node of the low-grade road, and it is re-judged whether the skeleton line node meets the slope condition;

[0043] In response to the slope condition being met, the correction is completed;

[0044] In response to the slope condition not being met, the elevation value range corresponding to each skeleton line node under the road slope requirement is obtained;

[0045] The elevation value closest to the elevation of the high-grade road node is selected from the intersection of the elevation value ranges of different skeleton line nodes and is assigned to the low-grade road node.

[0046] In a second aspect, a three-dimensional road curve model construction method is provided.

[0047] The three-dimensional road curve model construction method of any one of the first to seventh implementable manners of the first aspect is used to construct a three-dimensional road curve model of a road element;

[0048] The three-dimensional road curve model and attribute information of the road element are combined to construct a road segment curve model and an intersection curve model, respectively;

[0049] All the road segment curve models and intersection curve models are merged to construct the three-dimensional road curve model.

[0050] In the second implementable manner of the second aspect, in combination with the first implementable manner of the second aspect, the elevation data of a digital elevation model is repaired using the elevation information of the three-dimensional road curve model.

[0051] In the third implementable manner of the second aspect, in combination with the first implementable manner of the second aspect, the method further includes:

[0052] The three-dimensional road curve model is divided into a complete route entity, a road section unit and a road intersection unit, and corresponding attribute information is respectively assigned to the complete route entity, the road section unit and the road intersection unit;

[0053] The complete route entity, the road section unit and the road intersection unit are respectively encoded, and a road semantic relationship network is constructed according to the encoding results;

[0054] Based on the road semantic relationship network, a corresponding knowledge graph is constructed with a semantic unit as a node, a relationship type as a context, a correlation strength and a constraint threshold as attributes.

[0055] Beneficial effects: The three-dimensional road curve and surface model construction method based on two-dimensional stock surveying and mapping data can obtain road skeleton lines and attribute information from two-dimensional stock surveying and mapping data, and correct the elevation abnormal nodes in the extracted road skeleton line section, the topological abnormality and the elevation abnormality at the road connection, so as to ensure the road elevation fitting effect and the elevation data precision of the three-dimensional road curve model constructed based on the road skeleton line. The surface model is constructed based on the three-dimensional road curve model, and the three-dimensional road surface model can intuitively reflect the three-dimensional shape of the road, so that the observer can perceive the spatial characteristics of the road from multiple angles and all directions. In combination with the three-dimensional road curve model, the spatial calculation and information carrying capacity of the model can be effectively improved. BRIEF DESCRIPTION OF DRAWINGS

[0056] In order to more clearly illustrate the specific embodiments of the present application, the drawings required to be used in the specific embodiments will be briefly introduced below. In all the drawings, the elements or parts are not necessarily drawn according to the actual proportions.

[0057] Figure 1 The flowchart of the three-dimensional road curve model construction method provided by an embodiment of the present application;

[0058] Figure 2 The flowchart of the three-dimensional road surface model construction method provided by an embodiment of the present application;

[0059] Figure 3 The schematic diagram of the topological abnormality of the road section connection with the node not reaching;

[0060] Figure 4 The schematic diagram of the topological abnormality of the road section connection with the node exceeding;

[0061] Figure 5 The node-node type elevation abnormal site node distribution;

[0062] Figure 6 The node-bar line type elevation abnormal site node distribution;

[0063] Figure 7 The distribution of the height anomaly points of the strip-to-strip type. DETAILED DESCRIPTION

[0064] The embodiments of the technical solutions of the present application will be described in detail below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present application, and therefore can only serve as examples, but cannot limit the protection scope of the present application.

[0065] As shown in the flow chart of the three-dimensional road curve model construction method, the construction method comprises: Figure 1

[0066] Step 1, extracting the skeleton line and attribute information of the road element from the two-dimensional stock surveying and mapping data, and assigning the elevation information of the road element to the skeleton line according to the attribute information;

[0067] Step 2, identifying the elevation anomaly nodes of the road segment part of the skeleton line, and correcting them by using the bidirectional sniffing repair algorithm;

[0068] Step 3, judging whether there is a topological anomaly at the road segment connection of the skeleton line after the elevation correction, and in response to the existence of the topological anomaly, selecting the corresponding correction method according to the topological anomaly condition for correction;

[0069] Step 4, judging whether there is an elevation anomaly at the road segment connection of the skeleton line after the topological correction, and in response to the existence of the elevation anomaly, correcting the abnormal elevation according to the skeleton line node distribution condition at the road segment connection;

[0070] Step 5, constructing a three-dimensional road curve model of the road element based on the corrected skeleton line.

[0071] Specifically, first, the road skeleton line and attribute information can be extracted from the two-dimensional stock surveying and mapping data corresponding to the road element, and the elevation information of the road element can be extracted from the digital elevation model and the digital surface model according to the attribute information of the road element and assigned to the road skeleton line. Then, the road skeleton line can be divided into road segment part and road segment connection part. For the road segment part, the skeleton line nodes with elevation anomalies can be identified according to the slope between adjacent nodes on the skeleton line, and the abnormal elevation data of the skeleton line nodes can be corrected by using the bidirectional sniffing repair algorithm. Then, for the road segment connection part, whether there is a topological anomaly at the road segment connection can be determined according to the distribution condition of the corrected different skeleton lines at the road segment connection, and if there is, the appropriate correction method can be selected according to the type of the topological anomaly condition for correction, so as to obtain a topologically normal skeleton line.

[0072] ​Due to the missing of the sampling nodes of the elevation information of the connected road, or the change of the elevation information after the processing of the two-way sniffing repair algorithm for the internal elevation anomaly of the road section, and the inconsistency of the elevation of the connected road with other road sections, etc. Therefore, after the topological anomaly of the connected road section is corrected, it is necessary to judge whether there is an elevation anomaly at the connected road section. If not, a three-dimensional road curve model is constructed according to the corrected skeleton line. If there is, the abnormal elevation can be corrected according to the distribution of the skeleton line nodes of different roads at the connected road section. Finally, a three-dimensional road curve model is constructed according to the corrected skeleton line. In this way, the road elevation fitting effect and the elevation data precision of the three-dimensional road curve model constructed based on the road skeleton line can be ensured.

[0073] Due to the strong time sequence and long time span of the construction of two-dimensional stock surveying and mapping data, there is a general phenomenon of non-uniform coordinate system. Therefore, when extracting the skeleton line and attribute information, the original spatial coordinate system of the two-dimensional stock surveying and mapping data needs to be read and recorded first, and the automatic conversion parameters are matched on the basis of data extraction, and are automatically converted to CGCS2000 coordinate system to realize the regularization and unification of the two-dimensional stock surveying and mapping data coordinate system.

[0074] In this embodiment, optionally, the skeleton line of the road element is extracted, including:

[0075] According to the element form type of the road element, an appropriate extraction tool is selected to extract the skeleton line of the road element from the two-dimensional stock surveying and mapping data.

[0076] Specifically, the road element can be divided into two types of surface and line, and the line road element can be divided into two types of double-line road element and single-line road element according to the data abstraction scale. Among them, the single-line road is the road skeleton line, which does not need to be processed.

[0077] For the surface road, the CenterlineReplacer tool can be used to convert the surface road into a skeleton line. For the double-line road, the CollapseDualLinesToCenterline function can be called to convert the double-line road into a skeleton line.

[0078] In this embodiment, optionally, the attribute information of the road element is extracted, including:

[0079] The extracted attribute information is cleaned, and the cleaned attribute information is merged according to the road name.

[0080] Specifically, in extracting the attribute information of the road element, new data attribute fields are created for important attribute information of the road such as road name, road grade, width, etc., the inheritance is regularized from the two-dimensional stock surveying and mapping data, and the AttributeManager is used for field renaming and deleting redundant fields to complete data cleaning and arrangement. The cleaned data is merged into continuous lines by using LineCombiner according to the road name, and then the LengthCalculator is used to calculate the road length attribute and record it in the attribute information.

[0081] In the embodiment, the elevation information of the road element is extracted according to the attribute information and is assigned to the skeleton line, including:

[0082] The densification operation is performed on the skeleton line of the road element, and the elevation information is assigned to the densified skeleton line.

[0083] Specifically, since the skeleton line with fewer nodes may lead to the loss of the height fluctuation information of the road after being three-dimensionalized. For example, if there are only two nodes, the original road with an uphill, a flat road and a downhill can only be represented in the form of an uphill, a downhill or a flat road. Therefore, after the skeleton line of the road element is extracted, the densification operation can be performed on the nodes of the skeleton line to increase the number of nodes of the skeleton line.

[0084] For example, in the python development language environment, the arcpy site package can be called to traverse all skeleton line elements, input the densification interval and other parameters, call the Densify function to perform node encryption on the skeleton line, and complete the densification operation of the skeleton line.

[0085] In the embodiment, the elevation information of the road element is extracted according to the attribute information and is assigned to the skeleton line, including:

[0086] According to the attribute information, the road type of the road element is determined, and the road type includes a flat road and a three-dimensional road;

[0087] For the flat road, the elevation information of the road element is obtained from the digital elevation model and is assigned to the skeleton line.

[0088] For the three-dimensional road, the elevation information of the road element is obtained from the digital elevation model and the digital surface model and is assigned to the skeleton line.

[0089] Specifically, the sources of obtaining elevation information include digital elevation model (DEM) and digital surface model (DSM). In obtaining the elevation information, various road classification systems can be integrated, and the road elements are classified, summarized and abstracted from the perspective of stereo height difference characteristics, and the road in the road element is abstracted into two categories of plane-like road and stereo road. The plane-like road is characterized by horizontal extension and no obvious vertical spatial structure, and the elevation mainly reflects the natural undulation of the ground, and the elevation information can be directly extracted from the digital elevation model (DEM) and assigned to the corresponding skeleton line. The stereo road has obvious vertical spatial hierarchy or spanning structure, and the elevation information of this kind of road is extracted from the digital elevation model (DEM) and the digital surface model (DSM) to accurately depict the height characteristics of the multi-layer structure. The specific expression is as follows:

[0090] ;

[0091] ;

[0092] 、 The elevation information corresponding to the plane-like road and the stereo road respectively, The elevation value of the node in the digital elevation model (DEM), The elevation value of the node in the digital surface model (DSM).

[0093] In this embodiment, optionally, the elevation abnormal nodes in the skeleton line segment part are identified, including:

[0094] Obtaining the slope corresponding to each skeleton line sub-segment of the skeleton line segment part, and comparing the slope of each skeleton line sub-segment with the preset slope threshold value respectively;

[0095] In response to the slope being greater than the slope threshold value, the subsequent node of the skeleton line sub-segment is an elevation abnormal node.

[0096] Specifically, after assigning the elevation information to the extracted skeleton line, in order to ensure the road elevation fitting effect and the elevation data precision, the skeleton line nodes with elevation abnormalities in the skeleton line segment part can be identified, so as to correct the skeleton line nodes with elevation abnormalities.

[0097] Specifically, the adjacent skeleton lines can be connected into a skeleton line sub-segment, and the slope of the skeleton line sub-segment is calculated, and the calculated slope value is compared with the preset slope threshold value. If the slope value does not exceed the slope threshold value, it indicates that the elevation of the subsequent skeleton line node is normal, otherwise, the elevation value of the subsequent skeleton line node is abnormal. For the skeleton line node with abnormal elevation value, a bidirectional sniffing repair algorithm can be used for correction.

[0098] In this embodiment, optionally, a bidirectional sniffing repair algorithm is used to correct the elevation data of elevation anomaly nodes, including:

[0099] Using the elevation anomaly node as a reference point, search for the nearest elevation normal node to the reference point along both directions of the road element;

[0100] Calculate the slope between two normal elevation nodes and compare the slope with a slope threshold;

[0101] In response to the slope being less than the slope threshold, an interpolation algorithm is used to generate elevation data for all nodes between two normal elevation nodes;

[0102] In response to the slope exceeding the slope threshold, taking the previous normal elevation node as a fixed point, along the direction from the fixed point to the next normal elevation node, search for normal elevation nodes with a slope less than the slope threshold between the fixed point and the fixed point, and calculate the planar distance between the searched normal elevation nodes and the fixed point.

[0103] The planar distance is compared with a distance threshold. In response to a planar distance being less than the distance threshold, an interpolation algorithm is used to generate elevation data for all nodes between the normal elevation node and the search node.

[0104] Manual intervention is initiated in response to a planar distance exceeding a distance threshold.

[0105] Specifically, firstly, using the skeleton line node with the elevation anomaly (i.e., the elevation anomaly node) as a reference point, search forward and backward for the nearest normal elevation node to the reference point. Then, calculate the slope between these two normal elevation nodes and compare the calculated slope value with a slope threshold. If the slope value is less than the slope threshold, an interpolation algorithm can be used to calculate the elevation value at the location of the skeleton line node with the elevation anomaly, and this value can be used to correct the elevation data of the skeleton line node. The specific calculation formula is as follows:

[0106] ;

[0107] ;

[0108] ;

[0109] in, From the skeleton line node of the elevation anomaly to the elevation normal node distance, Elevation normal node , The distance between them These are the elevation values ​​of the skeleton line nodes for geoid eccentricity. , These are the normal elevation nodes. , the elevation value of the skeleton line node.

[0110] If the slope value exceeds the slope threshold value, the elevation data of the skeleton line node cannot be corrected by the interpolation algorithm. At this time, the elevation normal node before the skeleton line node can be taken as a fixed point, and the slope between the elevation normal node after the elevation normal node and the elevation normal node is calculated in sequence until the slope value between the elevation normal node and the elevation normal node is less than the slope threshold value. Then, the planar distance between the elevation normal node and the elevation normal node is calculated. If the planar distance is greater than a pre-set distance threshold value, the elevation value of the elevation abnormal skeleton line node is assigned in a manually set manner. Otherwise, according to the newly searched elevation normal node and the elevation normal node , the interpolation algorithm is used to repair the elevation of the abnormal node.

[0111] After the elevation abnormality in the road section part of the skeleton line is corrected, the planar topological problem of the road section connection part needs to be identified and processed.

[0112] In this embodiment, the corresponding correction method can be selected according to the topological abnormality to correct, including:

[0113] According to the attribute information, the high-level road and the low-level road at the road section connection are determined, and according to the node coordinates of the low-level road and the coverage range of the high-level road, the topological abnormality at the road section connection is determined.

[0114] When the skeleton line node of the low-level road does not reach the coverage range of the high-level road, a buffer area is generated with the skeleton line node as the origin and a set topological distance as the radius.

[0115] It is judged whether there is an intersection between the buffer area and the high-level road.

[0116] In response to the existence of the intersection, an adaptive intersection is selected from all intersections as a new node of the skeleton line of the low-level road at the road section connection.

[0117] In response to the non-existence of the intersection, the surveying and mapping data of the road feature and the high-level road are manually checked.

[0118] Specifically, in identifying the topology anomaly, first, the three-dimensional coordinates of the skeleton line can be reduced to two-dimensional plane for simplified calculation to improve processing efficiency. Then, according to the road level in the attribute information, the high-level road and the low-level road involved in the road section connection are determined. If the involved roads are of the same level, the road read first is taken as the high-level road, and the road read later is taken as the low-level road. Then, according to the node coordinates of the low-level road at the road connection, the distance between the node of the skeleton line corresponding to the low-level road and the high-level road is calculated, and the calculated distance value is compared with the pre-set topology threshold value. If it is less than the distance threshold value, it can be determined that the topology problem belongs to node underrun, as shown in Figure 3 If it is greater than the distance threshold value, the topology problem belongs to node overrun, as shown in Figure 4 Different topology problems need to be corrected by different correction methods.

[0119] For node underrun, first, the skeleton node of the low-level road at the road connection is taken as the origin, and the topology threshold value is taken as the radius to form the corresponding buffer zone. Then, it is judged whether there is an intersection point between the buffer zone and the skeleton line of the high-level road. If there is no intersection point, it indicates that the road distance between the low-level road and the high-level road is too large, and the two-dimensional stock surveying and mapping data of the road feature needs to be checked manually. If there is only one intersection point, the origin of the low-level road is extended to the intersection point. If there are multiple intersection points, the intersection point distance between the origin and each intersection point is calculated, and the intersection point corresponding to the minimum intersection point distance is selected, and the origin is extended to the selected intersection point, thereby correcting the topology problem of node underrun.

[0120] For node overrun, the skeleton line node of the low-level road that overruns can be reversed to the high-level road along the skeleton line subsegment where the skeleton line node is located, and it is judged whether there is a skeleton line node at the cutting position of the skeleton line of the high-level road. If there is, the skeleton line node of the low-level road is cut to the skeleton line node of the high-level road. If there is not, a new skeleton line node is added on the skeleton line of the high-level road, and the new node is given the elevation information, thereby correcting the topology problem of node overrun.

[0121] After completing the internal elevation anomaly correction of the road section and the planar topology problem repair of the multi-section connection, the elevation information of the road processed by the internal elevation anomaly bidirectional sniffing repair algorithm changes, thereby causing inconsistency with the elevation at the connection with other road sections. There may also be missing sampling nodes of the connection road elevation information, or distorted situations caused by errors or insufficient precision of the basic data information. These will all cause three-dimensional elevation anomalies at the road section connection. Therefore, after the topology problem at the road connection is corrected, the elevation information at the road section connection also needs to be identified and corrected.

[0122] In the embodiment, optionally, it is judged whether the skeleton line after the topological correction has an elevation anomaly at the road section connection, including:

[0123] The elevation data of the skeleton line of each connected road at the road section connection is judged for consistency. If the elevation data of different connected roads is inconsistent, it is judged whether the skeleton line has an elevation anomaly at the road section connection.

[0124] Specifically, the elevation consistency can be judged at the road connection, that is, the elevation information of each site where the skeleton line of different roads is located at the road section connection is compared to judge whether the elevation information of each site is consistent, and the node with inconsistent elevation is identified as an abnormal site to form an abnormal site data set. In combination with the road connection information extracted from the two-dimensional stock mapping data, the sites that cannot be directly connected in practice, such as roads on viaducts, roads on bridges and tunnels, etc., are removed from the abnormal site data set, and the remaining sites in the abnormal site data set are the elevation abnormal sites that should be connected in practice but have inconsistent elevation information.

[0125] In the embodiment, optionally, the abnormal elevation is corrected according to the node distribution of the skeleton line at the road section connection, including:

[0126] When the skeleton line nodes exist at the road section connection of each connected road, it is calculated whether the skeleton line nodes satisfy the slope condition;

[0127] In response to not satisfying the slope condition, a bidirectional sniffing repair algorithm is used to calculate a repaired elevation;

[0128] The repaired elevation is assigned to the skeleton line node of the low-grade road, and it is re-judged whether the skeleton line node satisfies the slope condition;

[0129] In response to satisfying the slope condition, the correction is completed;

[0130] In response to not satisfying the slope condition, an elevation value range corresponding to each skeleton line node under the road slope requirement is obtained;

[0131] From the intersection of the elevation value ranges of different skeleton line nodes, an elevation value closest to the elevation of the high-grade road node is selected and assigned to the low-grade road node.

[0132] Specifically, for the elevation abnormal site at the road section connection, first, the node distribution of the skeleton line of the elevation abnormal site can be determined according to the node position distribution of the skeleton line of different roads. In the embodiment, the node distribution of the elevation abnormal site is divided into three types: node-node type, node-bar type, and bar-bar type.

[0133] Among them, roads connected in the node-node type all have skeleton line nodes at the elevation anomaly sites. Roads connected in the node-line type have only one skeleton line node at the elevation anomaly site. Roads connected in the line-line type have no skeleton line nodes at the elevation anomaly site.

[0134] like Figure 5 As shown, for node-to-node type geometries, two skeleton line nodes can be identified at the geometries. , The slope value is checked against the slope condition, i.e., the slope value is less than the slope threshold. If it is, the elevation of the high-grade road at the node is directly used as the elevation of the anomaly site. If it is not, the skeleton line node of the high-grade road located at the elevation anomaly site is used as the reference point, and the above-mentioned bidirectional sniffing repair algorithm is used to calculate the repair elevation. The repair elevation is then assigned to the skeleton line node of the low-grade road located at the elevation anomaly site, and the slope of the skeleton line node of the low-grade road is re-evaluated to see if it meets the slope condition.

[0135] If the conditions are met, then the elevation anomaly correction is completed. If not, then the elevation range of the two skeleton line nodes that meet the road slope requirements is calculated. and And determine two elevation ranges. and The intersection of the two elevation values ​​is used to select the elevation value of the skeleton line node closest to that of the higher-grade road and assign it to the skeleton line node of the lower-grade road, thus completing the elevation anomaly correction. If the two elevation value ranges... and If there is no intersection between them, it indicates that the elevation information initially assigned to the skeleton line is incorrect, and the initial elevation information needs to be manually checked.

[0136] like Figure 6 As shown, for the node-line type, only road A has skeleton line nodes at the elevation anomaly points. A new skeleton line node can be added at elevation anomaly points on existing skeleton lines, and the corresponding elevation information can be extracted from the relevant road data and assigned to the new node. Then, it is determined whether the slope of the new node meets the slope requirements. If it does, the correction is complete. If not, further correction is performed using a node-to-node correction method.

[0137] like Figure 7As shown, for the strip-line-strip line type, neither road A nor road B has a skeleton line node located at the elevation anomaly point. Similarly, a skeleton line node can be added at the elevation anomaly point of the road A skeleton line and the road B skeleton line, and the corresponding elevation information is extracted from the corresponding road data and assigned to the added node. Then, it is judged whether the slope of the added node meets the slope requirement. If it does, the correction is completed. If it does not, further correction is made according to the node-node type correction method.

[0138] As shown in the flowchart of the three-dimensional road curve model construction method, the construction method comprises: Figure 2

[0139] Step S1, using the three-dimensional road curve model construction method described above, a three-dimensional road curve model of a road element is constructed.

[0140] Step S2, combining the three-dimensional road curve model and the attribute information of the road element, a road section curve model and an intersection curve model are respectively constructed.

[0141] Step S3, merging all the road section curve models and intersection curve surfaces, the three-dimensional road curve model is constructed.

[0142] Specifically, the three-dimensional road curve model is a lightweight simulation model constructed on the basis of the three-dimensional curve road model in combination with road attribute information. First, the three-dimensional road curve model can be semantically segmented by comprehensively considering road geometric shape and spatial semantic relationship, and the model is divided into a road section model and an intersection model.

[0143] Then, for the road section model, the actual length of each road can be obtained based on the geometry.du function in the CityEngine, and the road units are divided along the road skeleton line direction by combining the split function with road lane number, road total width and other parameters to determine the initial layout of the road surface such as lane, curb, median strip, etc. The split function is used to segment the road cross section to generate components such as center median strip, guardrail, marking line, etc. Some road components such as traffic signal, tree, street lamp, etc. are inserted into the components using the insertAlongUV function along the road. Then, the road material is projected to the road surface using the projectUV and setupProjection functions, the material direction and proportion are adjusted using the scaleUV and rotateUV functions, the road is stretched using the extrude function, the road attributes are adjusted according to the display effect of the road model, and the three-dimensional road curve model without connection is formed.

[0144] ​For the intersection model, the intersection can be classified according to the geometric shape difference, and the intersection can be divided into a cross intersection, a T-shaped intersection, a roundabout, a Y-shaped intersection and an irregular intersection. Among them, the irregular intersection is further subdivided into a small angle intersection and a large angle intersection according to the angle between each road at the intersection. Different types of intersections use different curved surface model construction schemes: ① the crossing intersection uses Crossing to construct the intersection model; ② the Y-shaped intersection uses Junction to construct the intersection model; ③ the roundabout intersection uses Roundabout to construct the intersection model; ④ the small angle intersection uses Freeway to construct the intersection model to ensure that there is no extrusion deformation between the road models; and ⑤ the large angle intersection uses Junction to construct the intersection model to ensure the connectivity of the main road.

[0145] Finally, the three-dimensional road curved surface model of the constructed connection-free road and the intersection model can be fused to generate a three-dimensional road curved surface model.

[0146] In this embodiment, optionally further comprising: using the elevation information of the three-dimensional road curved surface model to repair the elevation data of the digital elevation model.

[0147] Specifically, after generating the three-dimensional road curved surface model, the MultipatchToRaster and MosaicToNewRaster functions of arcpy can be called to use the curved surface elevation to repair the DEM elevation, to ensure that the three-dimensional road curved surface model and the terrain can be completely matched, and finally form a "semantic analysis-fractal modeling-terrain matching" three-dimensional road curved surface model. The three-dimensional road curved surface model can intuitively reflect the three-dimensional shape of the road, is the core of the expression of three-dimensional road space visualization, and can enable the viewer to perceive the spatial characteristics of the road from multiple angles and in all directions.

[0148] In this embodiment, optionally further comprising:

[0149] The three-dimensional road curved surface model is divided into a complete route entity, a road segment unit and an intersection unit, and the corresponding attribute information is respectively assigned to the complete route entity, the road segment unit and the intersection unit;

[0150] The complete route entity, the road segment unit and the intersection unit are respectively encoded, and a road semantic relationship network is constructed according to the encoding result;

[0151] Based on the road semantic relationship network, a corresponding knowledge graph is constructed with a semantic unit as a node, a relationship type as a context, a correlation strength and a constraint threshold as an attribute.

[0152] Specifically, after the three-dimensional road curve model is constructed, the three-dimensional road curve model and the three-dimensional road curve model can be semantically transformed, so as to construct a multi-level semantic relationship of the three-dimensional road, and form a road semantic system, so as to solve the problem that the three-dimensional road is good-looking but not easy to use, and can be seen but not checked.

[0153] Specifically, the three-dimensional road curve model is the core of the semantic relationship construction, and the semantic transformation of the three-dimensional road curve model includes: first, the three-dimensional road curve model is semantically segmented, and the road curve is divided into complete route entities, road segment units and intersection units.

[0154] The complete route entity is an abstract entity of the road, corresponds to a road object in the real world with independent naming and complete functionality, and carries the comprehensive attribute information of the road. The intersection unit is a junction of multiple routes, has a clear spatial range and functional boundary. In the curve model, the intersection point of the road segment is selected as the abstract expression of the intersection.

[0155] The road segment unit is a linear segment formed after the route is cut by the intersection. The complete route binds the overall identification attribute (such as road name, unique code, administrative attribution) and the comprehensive attribute. The road segment unit inherits the core attribute of the route, and binds the specific physical attribute (such as actual length, number of lanes) and the traffic control attribute (such as speed limit value, lane function division).

[0156] The intersection unit binds the type attribute (such as planar intersection, vertical intersection) and the running attribute (such as a set of turning rules), and the intersection attribute needs to form a logical self-consistency with the attribute of the associated road segment, to realize the organic unity of the semantic attributes of the three. The semantic segmentation result formed by the above method not only maintains the geometric consistency with the three-dimensional road curve model, but also constructs a penetrable and inferable semantic relationship network, providing a structured semantic basis for subsequent semantic relationship construction.

[0157] After the three-dimensional road curve model is semantically segmented, a fixed length coding structure of "18-bit main code + 2-bit state code" can be used to code the curve and surface model. The first 6 bits are regional codes referring to the national administrative area code, the 7th-8th bits are road type codes, referring to the city road engineering design specification, 01 represents the main road, 02 represents the secondary road, etc. The 9th-18th bits are unique sequence codes generated by road incremental counting. The state code uses a combination of letters and numbers, the first letter identifies the category, L represents the curve model, M represents the surface model, the second number distinguishes the semantic form, 1 represents the route, 2 represents the road segment, and 3 represents the intersection. Through the splicing of "main code + state code", the coding association of different forms of the same road is realized. The curve model is used for road semantic relationship construction, and the surface model is used for data visualization display by coding connection with the curve model.

[0158] After the encoding is completed, the road semantic relationship network can be constructed according to the encoding, and the system can sort out the semantic relationship in three aspects and visualize it as a knowledge graph. For the semantic relationship between roads, first, the spatial semantic relationship of roads is defined, including connection relationship, intersection relationship, and inclusion relationship.

[0159] For the semantic relationship inside the road, the route is the highest level semantic entity, which forms an inclusion relationship with the road segment and the intersection. A complete route is composed of several continuous road segment units and the intersections between them, and the overall properties of the route will constrain the properties of the subordinate road segments and intersections.

[0160] The road segment unit is a linear component of the route, and through the series connection of “road segment-intersection-road segment”, it constructs a continuous traffic path. The intersection unit is a junction hub of multiple road segments or multiple routes, which not only bears the spatial connection function of connecting different road segments, but also regulates the traffic flow conversion between road segments through traffic rules (such as turning permission and priority order), forming the internal semantic relationship of “route governing road segment and intersection, road segment relying on intersection for connection, and intersection supporting route interconnection”. For the semantic relationship between roads and external elements, according to the needs of the application scenario, special semantic relationships are established, such as the service relationship between roads and buildings, the disaster risk relationship between roads and rivers or mountains, and the management and maintenance semantic relationship between roads and management units.

[0161] On the basis of the semantic relationship network, a knowledge graph of “node-vein-attribute” is constructed, with semantic units as nodes, relationship types as veins, and associated strength, constraint threshold, etc. as attributes. Cross-unit mapping is realized through node coding, forming the “bidirectional penetration” capability of nodes, i.e. “forward penetration” and “backward penetration”. “Forward penetration” supports penetration from high-level road network to low-level units, such as from “city road network” to a “intersection unit”, and then to the associated information related to the intersection unit. “Backward penetration” supports backtracking from the unit to the system it belongs to (such as from a “road segment unit” to the “route unit” it belongs to, and then to the entire “road network system”).

[0162] Meanwhile, by defining the relationship weight (such as the "connection" relationship weight is higher than the "adjacent"), the priority analysis of the semantic relationship is realized, and finally a three-dimensional road multi-level semantic relationship system with clear type, distinguishable unit, connectable relationship and penetrable level is formed. With the help of the semantic relationship network, the "visible and searchable" integrated application of the three-dimensional road can be realized: in the visualization layer, the semantic penetration capability is supported to drill down from the macro road network level to the specific road section, intersection, and even a single auxiliary facility, and the front end is called and visualized; in the query layer, based on the semantic association rule, the intelligent search across levels and types is realized, for example, inputting "the road section with emergency service relationship with the hospital", the associated road section can be automatically located, the query result is returned with the semantic relationship attribute, and the reverse tracing (from the road section to query the main road and the associated maintenance responsibility unit) is supported, and finally the three-dimensional road management mode of "what is seen can be associated, and what is searched can be penetrated" is formed, which not only meets the intuitive display demand, but also realizes the efficient search and association analysis of the semantic information.

[0163] Due to the variety of three-dimensional data formats at present, in order to meet the needs of three-dimensional curve and surface entity models in data sharing, efficient transmission, dynamic updating and the like, the mainstream three-dimensional data format can be selected to store the three-dimensional road curve model and the three-dimensional road surface model.

[0164] The three-dimensional curve model is stored by using gdb or shp, the three-dimensional surface model is stored by using obj or fbx, and the entity relationship is stored by using a graph database. After the storage format of the entity model is unified, a three-dimensional data service suitable for multiple engines is generated based on the model data, and the three-dimensional curve and surface model can be compatible with the i3s service specification and the 3dtiles service specification.

[0165] The 3dtiles service is flexible and can efficiently process large-scale three-dimensional data; the i3s service realizes the ordered management and rapid calling of data by virtue of the tree-shaped data organization form and the data description storage mode of the combination of json and binary files, and the two cooperate to provide a solid support for cross-platform applications. The entity relationship information is stored in the graph database, and the entity attribute and the semantic relationship are provided as services through the data interface, so as to facilitate the user to quickly query and deeply analyze the entity relationship, fully release the value of the three-dimensional road entity data, and promote the sharing and application of multi-field data.

[0166] The above embodiments are only used to illustrate the technical solutions of the present application, but not limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can be modified, or some or all of the technical features can be replaced equivalently; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application, and they should be covered in the scope of the claims and the description of the present application.

Claims

1. A method of constructing a three-dimensional road curve model, characterized by, The method comprises the following steps: extracting the skeleton line and attribute information of the road element from the two-dimensional stock surveying and mapping data, and assigning the skeleton line with the elevation information of the road element according to the attribute information; identifying the elevation abnormal nodes of the skeleton line section, and correcting them by using a bidirectional sniffing repair algorithm; judging whether there is a topological abnormality at the road section connection of the skeleton line after the elevation correction, and in response to the existence of the topological abnormality, selecting a corresponding correction method according to the topological abnormality condition for correction; judging whether there is an elevation abnormality at the road section connection of the skeleton line after the topological correction, and in response to the existence of the elevation abnormality, correcting the abnormal elevation according to the skeleton line node distribution condition at the road section connection; constructing a three-dimensional road curve model of the road element based on the corrected skeleton line; correcting the elevation data of the elevation abnormal nodes by using a bidirectional sniffing repair algorithm, which comprises the following steps: taking the elevation abnormal node as a reference point, searching for the nearest elevation normal nodes in two directions of the road element respectively; calculating the slope between the two elevation normal nodes, and comparing the slope with a slope threshold value; in response to the slope being less than the slope threshold value, generating the elevation data of all nodes between the two elevation normal nodes by using an interpolation algorithm; in response to the slope exceeding the slope threshold value, taking the former elevation normal node as a fixed point, searching for an elevation normal node with a slope less than the slope threshold value between the fixed point and the latter elevation normal node in the direction from the fixed point to the latter elevation normal node, and calculating the planar distance between the searched elevation normal node and the fixed point; comparing the planar distance with a distance threshold value, and in response to the planar distance being less than the distance threshold value, generating the elevation data of all nodes between the former elevation normal node and the searched node by using an interpolation algorithm; in response to the planar distance exceeding the distance threshold value, manually intervening.

2. The three-dimensional road curve modeling method of claim 1, wherein, extracting the skeleton line of the road element, which comprises the following steps: according to the element form type of the road element, selecting a suitable extraction tool to extract the skeleton line of the road element from the two-dimensional stock surveying and mapping data.

3. The three-dimensional road curve modeling method of claim 1, wherein, assigning the skeleton line with the elevation information of the road element according to the attribute information, which comprises the following steps: performing a densification operation on the skeleton line of the road element, and assigning the densified skeleton line with the elevation information.

4. The three-dimensional road curve modeling method of claim 1, wherein, selecting a corresponding correction method according to the topological abnormality condition for correction, which comprises the following steps: when the skeleton line node of the low-grade road does not reach the coverage range of the high-grade road, generating a buffer zone with the skeleton line node as the origin and a set topological distance as the radius; judging whether there is an intersection between the buffer zone and the high-grade road; in response to the existence of the intersection, selecting a suitable intersection point from all the intersection points as a new node of the low-grade road skeleton line at the road section connection; in response to the non-existence of the intersection, manually checking the surveying and mapping data of the road element and the high-grade road; when the skeleton line node is located in the coverage range of the high-grade road, reversely cutting the skeleton line node to the boundary of the coverage range; judging whether there is a skeleton line node at the boundary of the skeleton line of the high-grade road; in response to the non-existence, adding a new node at the boundary and assigning the new node with elevation data; in response to the existence, taking the skeleton line node as a new node of the low-grade road skeleton line.

5. The three-dimensional road curve modeling method of claim 1, wherein, Judging whether there is an elevation anomaly of the skeleton line after the topology correction at the road segment connection, comprising: Judging the consistency of the elevation data of the skeleton line of each connected road at the road segment connection, and if the elevation data of different connected roads is inconsistent, whether there is an elevation anomaly of the skeleton line at the road segment connection.

6. The three-dimensional road curve modeling method of claim 1, wherein, Correcting the abnormal elevation according to the distribution of the skeleton line nodes at the road segment connection, comprising: When each connected road has a skeleton line node at the road segment connection, calculating whether each skeleton line node meets the slope condition; In response to not meeting the slope condition, calculating the repair elevation by using a bidirectional sniffing repair algorithm; Assigning the repair elevation to the skeleton line node of the low-level road and re-judging whether the skeleton line node meets the slope condition; In response to meeting the slope condition, completing the correction; In response to not meeting the slope condition, obtaining the elevation value range corresponding to each skeleton line node under the requirement of road slope; From the intersection of the elevation value ranges of different skeleton line nodes, selecting the elevation value closest to the elevation of the high-level road node and assigning it to the low-level road node; When only one connected road or no connected road has a skeleton line node at the road segment connection, adding a node at the road segment connection and assigning the elevation data at the location to the added node; Judging whether the elevation of the added node meets the slope condition, and in response to not meeting the slope condition, calculating the repair elevation by using a bidirectional sniffing repair algorithm; Assigning the repair elevation to the skeleton line node of the low-level road and re-judging whether the skeleton line node meets the slope condition; In response to meeting the slope condition, completing the correction; In response to not meeting the slope condition, obtaining the elevation value range corresponding to each skeleton line node under the requirement of road slope; From the intersection of the elevation value ranges of different skeleton line nodes, selecting the elevation value closest to the elevation of the high-level road node and assigning it to the low-level road node.

7. A three-dimensional road curve model construction method, characterized by, Comprising: Using the three-dimensional road curve model construction method of any one of claims 1-6 to construct a three-dimensional road curve model of a road feature; Combining the three-dimensional road curve model and attribute information of the road feature to respectively construct a road segment surface model and an intersection surface model; Merging all the road segment surface models and intersection surface models to construct the three-dimensional road surface model.

8. The three-dimensional road curve model construction method according to claim 7, characterized by, Further comprising: Using the elevation information of the three-dimensional road surface model to repair the elevation data of a digital elevation model.

9. The three-dimensional road curve model construction method according to claim 7, characterized by, Further comprising: Dividing the three-dimensional road curve model into complete route entities, road segment units, and intersection units, and respectively assigning corresponding attribute information to the complete route entities, road segment units, and intersection units; Encoding the complete route entities, road segment units, and intersection units respectively, and constructing a road semantic relationship network according to the encoding results; Based on the road semantic relationship network, constructing a corresponding knowledge graph with semantic units as nodes, relationship types as contexts, and correlation strength and constraint threshold as attributes.

Images