CN117633137A - Map data analysis and extraction method and system based on deep learning - Google Patents
Map data analysis and extraction method and system based on deep learning Download PDFInfo
- Publication number
- CN117633137A CN117633137A CN202311702423.3A CN202311702423A CN117633137A CN 117633137 A CN117633137 A CN 117633137A CN 202311702423 A CN202311702423 A CN 202311702423A CN 117633137 A CN117633137 A CN 117633137A
- Authority
- CN
- China
- Prior art keywords
- data
- key information
- extraction
- confidence
- building
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000605 extraction Methods 0.000 title claims abstract description 102
- 238000007405 data analysis Methods 0.000 title claims abstract description 37
- 238000013135 deep learning Methods 0.000 title claims abstract description 31
- 238000012544 monitoring process Methods 0.000 claims abstract description 98
- 238000012545 processing Methods 0.000 claims abstract description 67
- 238000004458 analytical method Methods 0.000 claims abstract description 18
- 238000013075 data extraction Methods 0.000 claims abstract description 11
- 230000010354 integration Effects 0.000 claims description 89
- 230000008859 change Effects 0.000 claims description 53
- 238000000034 method Methods 0.000 claims description 38
- 230000008569 process Effects 0.000 claims description 28
- 238000007781 pre-processing Methods 0.000 claims description 15
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000012937 correction Methods 0.000 claims description 5
- 238000004140 cleaning Methods 0.000 claims description 3
- 238000012552 review Methods 0.000 claims 1
- 230000006870 function Effects 0.000 abstract description 5
- 238000013480 data collection Methods 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 230000000052 comparative effect Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Landscapes
- Traffic Control Systems (AREA)
Abstract
The invention discloses a map data analysis and extraction method and system based on deep learning, and relates to the technical field of map data analysis. The invention is used for solving the technical problems that the flexibility of map data processing and the accuracy of updating the data set can not be further improved by carrying out differentiated analysis and extraction on map data with different confidence or updating conditions. Multiple signals about update or confidence conditions are obtained by performing multiple analysis processing on the monitoring set, confidence related labels are given, differentiated data extraction is conveniently performed on the monitoring set corresponding to the confidence related labels, flexibility of map data processing and accuracy of updating the data set are improved, and functions of timely replacing and newly adding accurate map information to the database are achieved.
Description
Technical Field
The invention belongs to the technical field of map data analysis, and particularly relates to a map data analysis and extraction method and system based on deep learning.
Background
Map data analysis and extraction is widely used in many fields including urban planning, traffic navigation, environmental monitoring, etc. Especially in the extraction of high-precision map elements, the method for manufacturing the map elements based on semi-automatic elements is long in time consumption, high in manual interaction and low in efficiency. The traditional pure laser point cloud method can provide less information and has high computational complexity.
In the map data analysis and extraction method in the prior art, firstly, a map point data set to be processed is obtained, the set comprises M map point data, and a distribution area corresponding to the set is determined according to the M map point data. Dividing the distribution area into n sub-areas by using an equal-area dividing method, and acquiring the number of map point data included in each sub-area. And meanwhile, determining the number m of sub-areas with non-zero map point data, judging whether the m meets the preset condition, if not, updating n by using the target parameter, and re-dividing the distribution area until the m meets the preset condition. And when m meets the preset condition, respectively determining one map point data from the m sub-areas to obtain m map point data. And sorting the m map point data according to the weight value and/or the density value corresponding to each map point data, and extracting N map point data according to the sorting result. However, the research finds that the method for sorting and extracting based on the weight values or the density values is too complicated to perform differential analysis and extraction on the map data with different confidence or update conditions, so that the flexibility of map data processing and the accuracy of updating the data set are further improved.
Disclosure of Invention
The invention aims to provide a map data analysis and extraction method and system based on deep learning, which are used for solving the technical problems that map data with different confidence or update conditions cannot be differentially analyzed and extracted in the prior art so as to further improve the flexibility of map data processing and the accuracy of updating a data set.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
the invention provides a map data analysis and extraction method based on deep learning, which comprises the following steps:
s100, collecting map data: collecting road key information, building key information and traffic facility key information in different dividing areas, and preprocessing the road key information, the building key information and the traffic facility key information to generate a road monitoring set, a building monitoring set and a traffic facility monitoring set;
s200, map data analysis: carrying out multiple analysis processing on the road monitoring set, the building monitoring set and the traffic facility monitoring set to obtain an integration set, generating an update signal, a confidence signal or an untrusted signal, giving confidence labels to the integration sets corresponding to the update signal and the confidence signal, and giving non-confidence labels to the integration sets corresponding to the untrusted signal;
s300, deep learning extraction: performing rechecking extraction processing on the integrated set corresponding to the non-confidence label to obtain a rechecking extraction data set; the method comprises the steps of comparing and extracting an integrated set corresponding to an opposite beacon to obtain an updated extracted data set and a reliable extracted data set;
s400, updating map data: transmitting the rechecked data in the rechecked extraction data set to a database, and carrying out replacement operation on the rechecked data and corresponding data in the original road key information, the original building key information and the original traffic facility key information in the database; transmitting the updated extracted data set as newly added information to a database for newly adding operation; and transmitting the reliable data in the reliable extraction data set to a database, and carrying out replacement operation with the corresponding data in the original road key information, the original building key information and the original traffic facility key information in the database.
Further, the road key information includes road name data, road type data and lane data, the building key information includes building name data, building boundary data and building position data, and the traffic facility information includes traffic facility name data and traffic facility position data;
the process of preprocessing the road key information, the building key information and the traffic facility key information is as follows: matching road key information, building key information and traffic facility key information in the dividing area with information judgment rules stored in a database in advance, judging that the rules are not met if the matching is unsuccessful, and cleaning and removing the information which is not met; and if the matching is successful, judging that the road key information, the building key information and the traffic facility key information which meet the rule are integrated into a road monitoring set, a building monitoring set and a traffic facility monitoring set respectively.
Further, the road monitoring set, the building monitoring set and the traffic facility monitoring set are subjected to multiple analysis processing to obtain an integrated set, and the process of generating the update signal, the confidence signal or the non-confidence signal is as follows:
s11, repeatedly collecting information which is unsuccessfully matched in a preprocessing process until the repeated collecting information is matched with an information judgment rule stored in advance in a database, automatically updating the repeated collecting information matched with the information judgment rule to a road monitoring set, a building monitoring set and a traffic facility monitoring set, and integrating to generate a primary processing integration set;
s12, extracting road name data, building name data and traffic facility name data in the first-level processing integration set, and matching the road name data, the building name data and the traffic facility name data with the road name data, the building name data and the traffic facility name data of corresponding divided areas stored in the database in advance; if the matching is successful, entering the next step; if the matching is unsuccessful, generating an update signal and giving a confidence label to the primary processing integration set;
s13, converting the successfully matched primary processing integration set into a secondary processing integration set, matching the secondary processing integration set with original road key information, original building key information and original traffic facility key information of corresponding divided areas stored in a database in advance, counting the data change proportions in a road monitoring set, a building monitoring set and a traffic facility monitoring set in the secondary processing integration set, marking the data change proportions as DB, JB and TB respectively, and acquiring corresponding confidence coefficients delta through a calculation formula; the confidence coefficient delta is calculated by the following formula:
;
wherein ZB is a confidence correction factor, and w1, w2 and w3 are different preset proportion coefficients;
s14, comparing the confidence coefficient delta with a confidence coefficient threshold, if the confidence coefficient is smaller than the confidence coefficient threshold, generating a confidence signal and giving a secondary processing integration set corresponding to the confidence signal a confidence label; if the confidence coefficient is greater than or equal to the confidence coefficient threshold, generating an untrusted signal and giving an untrusted label to the secondary processing integration set corresponding to the untrusted signal.
Further, the process of rechecking and extracting is as follows: extracting change data in an integration set corresponding to the non-confidence label, repeatedly collecting the change data, judging as rechecking data if the content of the repeatedly collected change data is unchanged, and judging the integration set corresponding to the non-confidence label as rechecking the extracted data set; if the content of the change data is changed again, repeatedly collecting the change data for a plurality of times, selecting the data with the same collection process and the maximum content change data ratio as the rechecking data, and integrating the rechecking data into an integration set corresponding to the non-confidence label to generate a rechecking extraction data set.
Further, the process of the comparative extraction treatment is as follows: for an integration set corresponding to the update signal, directly judging the integration set as an update extraction data set; extracting change data in an integration set corresponding to the confidence signal, repeatedly collecting the change data, judging the integration set as reliable data if the content of the repeatedly collected change data is unchanged, and judging the integration set as reliable extraction data set; if the content of the change data is changed again, the change data is repeatedly collected for a plurality of times, the data with the largest content change data ratio in the same collection process is selected as reliable data, and the reliable data are integrated into an integration set to generate a reliable extraction data set.
The invention also provides a map data analysis and extraction system based on the deep learning, which comprises a map data acquisition module, a map data analysis module, a deep learning and extraction module, a map data updating module and a database;
the map data acquisition module is used for acquiring road key information, building key information and traffic facility key information in different partitioned areas, preprocessing the road key information, the building key information and the traffic facility key information to generate a road monitoring set, a building monitoring set and a traffic facility monitoring set, and transmitting the road monitoring set, the building monitoring set and the traffic facility monitoring set to the map data analysis module;
the map data analysis module is used for carrying out multiple analysis processing on the road monitoring set, the building monitoring set and the traffic facility monitoring set to obtain an integration set, generating an update signal, a confidence signal or an untrusted signal, giving confidence labels to the integration sets corresponding to the update signal and the confidence signal, and giving non-confidence labels to the integration sets corresponding to the untrusted signal;
the deep learning extraction module is used for carrying out rechecking extraction processing on the integration set corresponding to the non-confidence label to obtain a rechecking extraction data set and sending the rechecking extraction data set to the map data updating module; the map data updating module is used for comparing and extracting the integration sets corresponding to the opposite beacon labels to obtain an updated extraction data set and a reliable extraction data set and sending the updated extraction data set and the reliable extraction data set to the map data updating module;
the map data updating module is used for transmitting the rechecking data in the rechecking extraction data set to the database, and carrying out replacement operation on the rechecking data in the rechecking extraction data set and the corresponding data in the original road key information, the original building key information and the original traffic facility key information in the database; transmitting the updated extracted data set as newly added information to a database for newly adding operation; and transmitting the reliable data in the reliable extraction data set to a database, and performing replacement operation with the corresponding data in the original road key information, the original building key information and the original traffic facility key information of the database.
Further, the database is used for storing information judgment rules, original road key information, original building key information and original traffic facility key information in advance, and is also used for replacing and adding map information.
In summary, due to the adoption of the technical scheme, the beneficial effects of the invention are as follows:
1. according to the invention, map data in different partitioned areas are acquired from three dimensions of road key information, building key information and traffic facility key information, and the information which does not accord with the rule can be cleaned and removed in advance in a preprocessing mode of matching the information judgment rule, so that the map data can be conveniently integrated into a monitoring set which accords with the judgment rule, and the reliability of the map data set is improved; multiple signals about update or confidence condition are obtained by performing multiple analysis processing on the monitoring set, confidence related labels are given, differentiated data extraction is conveniently performed on the monitoring set corresponding to the confidence related labels, flexibility of map data processing and accuracy of updating the data set are improved, and functions of timely replacing and newly adding accurate map information in a database are achieved.
2. The multiple analysis processing of the invention is aimed at a monitoring set which does not accord with the information judgment rule in the preprocessing process, repeated collection is carried out until the information judgment rule is matched, and then the monitoring set is integrated to generate a first-stage processing integration set; and then the name data in the primary processing integration set is matched with a database, an update signal is generated after unsuccessful matching, the subsequent new-addition operation on the newly-added name data and other related data is facilitated, the newly-added name data is converted into a secondary processing integration set after successful matching, a confidence coefficient reflecting the comprehensive change condition of information is obtained through a mode of marking the data change proportion, confidence correction and formula calculation, a confidence signal or an untrusted signal is generated after threshold comparison, and the subsequent data extraction operation is facilitated in a differentiated mode.
3. The rechecking and extracting process ensures the accuracy of rechecking and extracting the data set in a repeated acquisition mode; and in the process of the comparison extraction processing, the integration set corresponding to the update signal is directly judged to be an update extraction data set, and the data with the same maximum duty ratio content is used as reliable data and integrated into the reliable extraction data set in a repeated acquisition mode, so that the accuracy of the reliable extraction data set is ensured.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 shows a flow chart of the map data analysis and extraction method based on deep learning of the present invention;
FIG. 2 shows a flow chart of the multiple analysis process of the present invention;
fig. 3 shows a block diagram of the deep learning-based map data analysis extraction system of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1
Referring to fig. 3, the present embodiment provides a map data analysis and extraction system based on deep learning, which includes a map data acquisition module, a map data analysis module, a deep learning and extraction module, a map data update module, and a database.
The map data acquisition module is used for acquiring road key information, building key information and traffic facility key information in different partitioned areas, preprocessing the road key information, the building key information and the traffic facility key information to generate a road monitoring set, a building monitoring set and a traffic facility monitoring set, and transmitting the road monitoring set, the building monitoring set and the traffic facility monitoring set to the map data analysis module.
The road key information includes road name data, road type data, and lane data, the building key information includes building name data, building boundary data, and building position data, and the traffic facility information includes traffic facility name data and traffic facility position data. Of course, the information collected in the embodiment is not limited to the above information, and may be adjusted and classified according to the actual needs and functions of the map data.
The regional division can be based on common administrative division, geographic feature division, functional division, economic division or manually defined division. The road type data is determined according to the purpose and function of the road, including expressways, primary roads, secondary roads, tertiary roads, urban expressways, urban arterial roads, urban secondary arterial roads, urban branches and urban loops. Building boundary data are represented by longitude and latitude coordinates of each corner point of a building; building position data is represented by latitude and longitude coordinates of a building geometric center; the transportation means location data is represented by longitude and latitude coordinates of the transportation means geometric center.
The preprocessing process of the road key information, the building key information and the traffic facility key information is as follows: matching road key information, building key information and traffic facility key information in the dividing area with information judgment rules stored in a database in advance, judging that the rules are not met if the matching is unsuccessful, and cleaning and removing the information which is not met; and if the matching is successful, judging that the road key information, the building key information and the traffic facility key information which meet the rule are integrated into a road monitoring set, a building monitoring set and a traffic facility monitoring set respectively.
The information judgment rules stored in the database in advance are defined by people and can be adjusted, for example, the latitude value range of the coordinates of the building position data is not between-180 degrees and +180 degrees, the latitude value range is not between-90 degrees and 90 degrees, the matching is unsuccessful, and the judgment is not in accordance with the rules; for example, if the road type data is a highway and the lane data is one, the matching is unsuccessful, and the rule is not satisfied.
The map data analysis module is used for carrying out multiple analysis processing on the road monitoring set, the building monitoring set and the traffic facility monitoring set to obtain an integration set, generating an update signal, a confidence signal or an untrusted signal, giving confidence labels to the integration sets corresponding to the update signal and the confidence signal, and giving non-confidence labels to the integration sets corresponding to the untrusted signal.
The deep learning extraction module is used for carrying out rechecking extraction processing on the integration set corresponding to the non-confidence label to obtain a rechecking extraction data set and sending the rechecking extraction data set to the map data updating module; and the integrated set corresponding to the opposite beacon is subjected to comparison extraction processing to obtain an updated extraction data set and a reliable extraction data set, and the updated extraction data set and the reliable extraction data set are sent to a map data updating module.
The map data updating module is used for transmitting the rechecking data in the rechecking extraction data set to the database, and carrying out replacement operation on the rechecking data in the rechecking extraction data set and the corresponding data in the original road key information, the original building key information and the original traffic facility key information in the database; transmitting the updated extracted data set as newly added information to a database for newly adding operation; and transmitting the reliable data in the reliable extraction data set to a database, and performing replacement operation with the corresponding data in the original road key information, the original building key information and the original traffic facility key information of the database.
The database is used for storing information judgment rules, original road key information, original building key information and original traffic facility key information in advance and is also used for replacing and adding map information.
According to the map data collection method and device, map data in different partitioned areas are collected from three dimensions of road key information, building key information and traffic facility key information, and information which does not accord with rules can be cleaned and removed in advance in a preprocessing mode of information judgment rule matching, so that the map data collection method and device are convenient to integrate into a monitoring set which accords with the judgment rules, and the reliability of the map data collection is improved; multiple signals about update or confidence condition are obtained by performing multiple analysis processing on the monitoring set, confidence related labels are given, differentiated data extraction is conveniently performed on the monitoring set corresponding to the confidence related labels, flexibility of map data processing and accuracy of updating the data set are improved, and functions of timely replacing and newly adding accurate map information in a database are achieved.
Example 2
Referring to fig. 2, the present embodiment is an improvement on the basis of embodiment 1, and the process of performing multiple analysis processing on the road monitoring set, the building monitoring set and the traffic facility monitoring set to obtain an integrated set, and generating the update signal, the confidence signal or the non-confidence signal is as follows:
s11, repeatedly collecting information which is unsuccessfully matched in a preprocessing process until the repeated collecting information is matched with an information judgment rule stored in advance in a database, automatically updating the repeated collecting information matched with the information judgment rule to a road monitoring set, a building monitoring set and a traffic facility monitoring set, and integrating to generate a primary processing integration set;
s12, extracting road name data, building name data and traffic facility name data in the first-level processing integration set, and matching the road name data, the building name data and the traffic facility name data with the road name data, the building name data and the traffic facility name data of corresponding divided areas stored in the database in advance; if the matching is successful, entering the next step; if the matching is unsuccessful, generating an update signal and giving a confidence label to the primary processing integration set;
s13, converting the successfully matched primary processing integration set into a secondary processing integration set, matching the secondary processing integration set with original road key information, original building key information and original traffic facility key information of corresponding divided areas stored in a database in advance, counting the data change proportions in a road monitoring set, a building monitoring set and a traffic facility monitoring set in the secondary processing integration set, marking the data change proportions as DB, JB and TB respectively, and acquiring corresponding confidence coefficients delta through a calculation formula; the confidence coefficient delta is calculated by the following formula:
;
wherein ZB is a confidence correction factor and can take the values of 0.856, w1, w2 and w3 are different preset proportional coefficients, w2 is more than 0.5 and less than w3 and less than 1.6, and w1, w2 and w3 can take the values of 1.528, 0.637 and 1.126 respectively;
the data change ratio refers to a ratio of the change of the contents of the data except the road name data, the building name data and the traffic facility name data in the road monitoring set, the building monitoring set and the traffic facility monitoring set; the confidence coefficient is a numerical value for evaluating the comprehensive change condition of each piece of information in the secondary processing integration set which does not generate an update signal; the larger the confidence coefficient is, the larger the content change of the corresponding secondary processing integration set relative to the original data in the database is, and the worse the reliability is;
s14, comparing the confidence coefficient delta with a confidence coefficient threshold, if the confidence coefficient is smaller than the confidence coefficient threshold, generating a confidence signal and giving a secondary processing integration set corresponding to the confidence signal a confidence label; if the confidence coefficient is greater than or equal to the confidence coefficient threshold, generating an untrusted signal and giving an untrusted label to the secondary processing integration set corresponding to the untrusted signal.
In the multiple analysis processing process, firstly, repeatedly collecting a monitoring set which does not accord with the information judgment rule in the pretreatment process until the monitoring set is matched with the information judgment rule, and integrating the monitoring set to generate a first-stage processing integration set; and then the name data in the primary processing integration set is matched with a database, an update signal is generated after unsuccessful matching, the subsequent new-addition operation on the newly-added name data and other related data is facilitated, the newly-added name data is converted into a secondary processing integration set after successful matching, a confidence coefficient reflecting the comprehensive change condition of information is obtained through a mode of marking the data change proportion, confidence correction and formula calculation, a confidence signal or an untrusted signal is generated after threshold comparison, and the subsequent data extraction operation is facilitated in a differentiated mode.
The rechecking and extracting process is as follows: extracting change data in an integration set corresponding to the non-confidence label, repeatedly collecting the change data, judging as rechecking data if the content of the repeatedly collected change data is unchanged, and judging the integration set corresponding to the non-confidence label as rechecking the extracted data set; if the content of the change data is changed again, continuously and repeatedly collecting the change data for a plurality of times, selecting the data with the same content change data ratio in the collecting process and the largest ratio as the rechecking data, and integrating the rechecking data into an integration set corresponding to the non-confidence label to generate a rechecking extraction data set; the number of times is preferably 3 to 5.
The method comprises the steps of repeatedly acquiring change data in an integration set corresponding to an untrusted label, obtaining consistent check data in a repeated acquisition mode, judging the integration set to which the change data belongs as a check extraction data set, repeatedly acquiring the change data with the content changed again for a plurality of times, taking the data with the largest content of the same proportion as the check data, and integrating the check extraction data set to generate a check extraction data set; the accuracy of rechecking the extracted data set is ensured in a repeated acquisition mode.
The comparative extraction process is as follows: for an integration set corresponding to the update signal, directly judging the integration set as an update extraction data set; extracting change data in an integration set corresponding to the confidence signal, repeatedly collecting the change data, judging the integration set as reliable data if the content of the repeatedly collected change data is unchanged, and judging the integration set as reliable extraction data set; if the content of the change data is changed again, continuously and repeatedly collecting the change data for a plurality of times, selecting the data with the same content change data with the largest ratio in the collecting process as reliable data, and integrating the reliable data into an integration set to generate a reliable extraction data set; the number of times is preferably 3 to 5.
And in the process of the comparison extraction processing, the integration set corresponding to the update signal is directly judged to be an update extraction data set, and the data with the same maximum duty ratio content is used as reliable data and integrated into the reliable extraction data set in a repeated acquisition mode, so that the accuracy of the reliable extraction data set is ensured.
Example 3
Referring to fig. 1, the present embodiment provides a map data analysis and extraction method based on deep learning, which is applicable to the map data analysis and extraction system based on deep learning of embodiment 1, and includes the following steps:
s100, collecting map data: collecting road key information, building key information and traffic facility key information in different dividing areas, and preprocessing the road key information, the building key information and the traffic facility key information to generate a road monitoring set, a building monitoring set and a traffic facility monitoring set;
s200, map data analysis: carrying out multiple analysis processing on the road monitoring set, the building monitoring set and the traffic facility monitoring set to obtain an integration set, generating an update signal, a confidence signal or an untrusted signal, giving confidence labels to the integration sets corresponding to the update signal and the confidence signal, and giving non-confidence labels to the integration sets corresponding to the untrusted signal;
s300, deep learning extraction: performing rechecking extraction processing on the integrated set corresponding to the non-confidence label to obtain a rechecking extraction data set; the method comprises the steps of comparing and extracting an integrated set corresponding to an opposite beacon to obtain an updated extracted data set and a reliable extracted data set;
s400, updating map data: transmitting the rechecked data in the rechecked extraction data set to a database, and carrying out replacement operation on the rechecked data and corresponding data in the original road key information, the original building key information and the original traffic facility key information in the database; transmitting the updated extracted data set as newly added information to a database for newly adding operation; and transmitting the reliable data in the reliable extraction data set to a database, and carrying out replacement operation with the corresponding data in the original road key information, the original building key information and the original traffic facility key information in the database.
The map data analysis and extraction method based on deep learning comprises the steps of map data acquisition, map data analysis, deep learning extraction and map data updating, wherein road key information, building key information and traffic facility key information of different divided areas are acquired from the beginning, multiple signals about updating and confidence conditions are obtained in a multiple analysis processing mode, and corresponding confidence or non-confidence labels are given; the consistency and accuracy of the newly added and replaced map information can be guaranteed to the greatest extent by combining the rechecking extraction processing with the deep learning extraction mode of the comparison extraction processing, and the reliability of the user on software applying the map data is improved.
The preset weight coefficient is used for balancing the duty ratio weight of each item of data in formula calculation, so that the accuracy of a calculation result is promoted; the size of the coefficient is a specific numerical value obtained by quantizing each parameter, so that the subsequent comparison is convenient, and the size of the coefficient depends on the number of sample data and the corresponding weight factor coefficient is preliminarily set for each group of sample data by a person skilled in the art; as long as the proportional relation between the parameter and the quantized value is not affected. The related formulas are all formulas obtained by software simulation after a large amount of data are collected for dimension removal, and a formula close to a true value is selected, and coefficients in the formulas are set by a person skilled in the art according to actual conditions. The threshold value and the preset range value mentioned in the application are obtained and selected by software simulation by a software communication technician collecting a large amount of data.
Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application. The functional units in the embodiments of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit.
The foregoing is merely specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily think about changes or substitutions within the technical scope of the present application, and the changes and substitutions are intended to be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (7)
1. The map data analysis and extraction method based on deep learning is characterized by comprising the following steps of:
s100, collecting map data: collecting road key information, building key information and traffic facility key information in different dividing areas, and preprocessing the road key information, the building key information and the traffic facility key information to generate a road monitoring set, a building monitoring set and a traffic facility monitoring set;
s200, map data analysis: carrying out multiple analysis processing on the road monitoring set, the building monitoring set and the traffic facility monitoring set to obtain an integration set, generating an update signal, a confidence signal or an untrusted signal, giving confidence labels to the integration sets corresponding to the update signal and the confidence signal, and giving non-confidence labels to the integration sets corresponding to the untrusted signal;
s300, deep learning extraction: performing rechecking extraction processing on the integrated set corresponding to the non-confidence label to obtain a rechecking extraction data set; the method comprises the steps of comparing and extracting an integrated set corresponding to an opposite beacon to obtain an updated extracted data set and a reliable extracted data set;
s400, updating map data: transmitting the rechecked data in the rechecked extraction data set to a database, and carrying out replacement operation on the rechecked data and corresponding data in the original road key information, the original building key information and the original traffic facility key information in the database; transmitting the updated extracted data set as newly added information to a database for newly adding operation; and transmitting the reliable data in the reliable extraction data set to a database, and carrying out replacement operation with the corresponding data in the original road key information, the original building key information and the original traffic facility key information in the database.
2. The deep learning-based map data analysis and extraction method of claim 1, wherein the road key information includes road name data, road type data, and lane data, the building key information includes building name data, building boundary data, and building position data, and the traffic facility information includes traffic facility name data and traffic facility position data;
the process of preprocessing the road key information, the building key information and the traffic facility key information is as follows: matching road key information, building key information and traffic facility key information in the dividing area with information judgment rules stored in a database in advance, judging that the rules are not met if the matching is unsuccessful, and cleaning and removing the information which is not met; and if the matching is successful, judging that the road key information, the building key information and the traffic facility key information which meet the rule are integrated into a road monitoring set, a building monitoring set and a traffic facility monitoring set respectively.
3. The deep learning-based map data analysis and extraction method of claim 1, wherein the process of performing multiple analysis processes on the road monitoring set, the building monitoring set, and the traffic facility monitoring set to obtain an integrated set, and generating the update signal, the confidence signal, or the non-confidence signal is as follows:
s11, repeatedly collecting information which is unsuccessfully matched in a preprocessing process until the repeated collecting information is matched with an information judgment rule stored in advance in a database, automatically updating the repeated collecting information matched with the information judgment rule to a road monitoring set, a building monitoring set and a traffic facility monitoring set, and integrating to generate a primary processing integration set;
s12, extracting road name data, building name data and traffic facility name data in the first-level processing integration set, and matching the road name data, the building name data and the traffic facility name data with the road name data, the building name data and the traffic facility name data of corresponding divided areas stored in the database in advance; if the matching is successful, entering the next step; if the matching is unsuccessful, generating an update signal and giving a confidence label to the primary processing integration set;
s13, converting the successfully matched primary processing integration set into a secondary processing integration set, matching the secondary processing integration set with original road key information, original building key information and original traffic facility key information of corresponding divided areas stored in a database in advance, counting the data change proportions in a road monitoring set, a building monitoring set and a traffic facility monitoring set in the secondary processing integration set, marking the data change proportions as DB, JB and TB respectively, and acquiring corresponding confidence coefficients delta through a calculation formula; the confidence coefficient delta is calculated by the following formula:
;
wherein ZB is a confidence correction factor, and w1, w2 and w3 are different preset proportion coefficients;
s14, comparing the confidence coefficient delta with a confidence coefficient threshold, if the confidence coefficient is smaller than the confidence coefficient threshold, generating a confidence signal and giving a secondary processing integration set corresponding to the confidence signal a confidence label; if the confidence coefficient is greater than or equal to the confidence coefficient threshold, generating an untrusted signal and giving an untrusted label to the secondary processing integration set corresponding to the untrusted signal.
4. The deep learning-based map data analysis and extraction method according to claim 1, wherein the process of the review extraction process is as follows: extracting change data in an integration set corresponding to the non-confidence label, repeatedly collecting the change data, judging as rechecking data if the content of the repeatedly collected change data is unchanged, and judging the integration set corresponding to the non-confidence label as rechecking the extracted data set; if the content of the change data is changed again, repeatedly collecting the change data for a plurality of times, selecting the data with the same collection process and the maximum content change data ratio as the rechecking data, and integrating the rechecking data into an integration set corresponding to the non-confidence label to generate a rechecking extraction data set.
5. The deep learning-based map data analysis and extraction method according to claim 1, wherein the process of the contrast extraction process is as follows: for an integration set corresponding to the update signal, directly judging the integration set as an update extraction data set; extracting change data in an integration set corresponding to the confidence signal, repeatedly collecting the change data, judging the integration set as reliable data if the content of the repeatedly collected change data is unchanged, and judging the integration set as reliable extraction data set; if the content of the change data is changed again, the change data is repeatedly collected for a plurality of times, the data with the largest content change data ratio in the same collection process is selected as reliable data, and the reliable data are integrated into an integration set to generate a reliable extraction data set.
6. The map data analysis and extraction system based on deep learning is characterized by comprising a map data acquisition module, a map data analysis module, a deep learning and extraction module, a map data updating module and a database;
the map data acquisition module is used for acquiring road key information, building key information and traffic facility key information in different partitioned areas, preprocessing the road key information, the building key information and the traffic facility key information to generate a road monitoring set, a building monitoring set and a traffic facility monitoring set, and transmitting the road monitoring set, the building monitoring set and the traffic facility monitoring set to the map data analysis module;
the map data analysis module is used for carrying out multiple analysis processing on the road monitoring set, the building monitoring set and the traffic facility monitoring set to obtain an integration set, generating an update signal, a confidence signal or an untrusted signal, giving confidence labels to the integration sets corresponding to the update signal and the confidence signal, and giving non-confidence labels to the integration sets corresponding to the untrusted signal;
the deep learning extraction module is used for carrying out rechecking extraction processing on the integration set corresponding to the non-confidence label to obtain a rechecking extraction data set and sending the rechecking extraction data set to the map data updating module; the map data updating module is used for comparing and extracting the integration sets corresponding to the opposite beacon labels to obtain an updated extraction data set and a reliable extraction data set and sending the updated extraction data set and the reliable extraction data set to the map data updating module;
the map data updating module is used for transmitting the rechecking data in the rechecking extraction data set to the database, and carrying out replacement operation on the rechecking data in the rechecking extraction data set and the corresponding data in the original road key information, the original building key information and the original traffic facility key information in the database; transmitting the updated extracted data set as newly added information to a database for newly adding operation; and transmitting the reliable data in the reliable extraction data set to a database, and performing replacement operation with the corresponding data in the original road key information, the original building key information and the original traffic facility key information of the database.
7. The deep learning based map data analysis and extraction system of claim 6, wherein the database is used for storing information decision rules, original road key information, original building key information, and original traffic facility key information in advance, and is also used for replacing and adding map information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311702423.3A CN117633137A (en) | 2023-12-12 | 2023-12-12 | Map data analysis and extraction method and system based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311702423.3A CN117633137A (en) | 2023-12-12 | 2023-12-12 | Map data analysis and extraction method and system based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117633137A true CN117633137A (en) | 2024-03-01 |
Family
ID=90025230
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311702423.3A Pending CN117633137A (en) | 2023-12-12 | 2023-12-12 | Map data analysis and extraction method and system based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117633137A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117870651A (en) * | 2024-03-11 | 2024-04-12 | 北京理工大学前沿技术研究院 | Map high-precision acquisition method, memory and storage medium based on RTK-SLAM technology |
-
2023
- 2023-12-12 CN CN202311702423.3A patent/CN117633137A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117870651A (en) * | 2024-03-11 | 2024-04-12 | 北京理工大学前沿技术研究院 | Map high-precision acquisition method, memory and storage medium based on RTK-SLAM technology |
| CN117870651B (en) * | 2024-03-11 | 2024-05-07 | 北京理工大学前沿技术研究院 | Map high-precision acquisition method, memory and storage medium based on RTK-SLAM technology |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109815993B (en) | GPS track-based regional feature extraction, database establishment and intersection identification method | |
| CN107292417B (en) | Regional heavy pollution discrimination and forecast method and device based on heavy pollution sequence case library | |
| CN110413905B (en) | Method, device and equipment for acquiring road alignment and storage medium | |
| CN117633137A (en) | Map data analysis and extraction method and system based on deep learning | |
| CN106323301A (en) | Road information obtaining method and device | |
| CN110020224B (en) | Correlation method, device, equipment and medium for map interest point data | |
| CN110798805B (en) | Data processing method and device based on GPS track and storage medium | |
| CN111814596B (en) | Automatic city function partitioning method for fusing remote sensing image and taxi track | |
| CN110619258B (en) | Road track checking method based on high-resolution remote sensing image | |
| CN107145991B (en) | Time-varying random network dynamic path searching method considering road section correlation | |
| CN112861972A (en) | Site selection method and device for exhibition area, computer equipment and medium | |
| CN111797188B (en) | Urban functional area quantitative identification method based on open source geospatial vector data | |
| CN114428828A (en) | Method and device for digging new road based on driving track and electronic equipment | |
| CN111524353A (en) | Method for traffic text data for speed prediction and travel planning | |
| CN112365595A (en) | Analysis method for identifying traffic accident multi-point based on alarm data point element | |
| CN113888867B (en) | Parking space recommendation method and system based on LSTM (least squares) position prediction | |
| CN106931976A (en) | A kind of method and device of navigation information updating | |
| CN112883236B (en) | Map updating method and device, electronic equipment and storage medium | |
| CN116824868B (en) | Method, device, equipment and medium for identifying illegal parking points and predicting congestion of vehicles | |
| CN117128950A (en) | Point cloud map construction method and device, electronic equipment and storage medium | |
| CN108053646A (en) | Traffic characteristic acquisition methods, Forecasting Methodology and system based on time-sensitive feature | |
| CN116013084A (en) | Traffic management and control scene determining method and device, electronic equipment and storage medium | |
| CN115984603A (en) | Fine classification method and system for urban green land based on GF-2 and open map data | |
| CN115563096A (en) | Closed park deduplication method based on longitude and latitude | |
| CN115169466A (en) | Method and device for drawing image of land, electronic equipment and computer readable medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |