CN112101299A - Automatic traffic sign extraction method and system based on binocular CCD camera - Google Patents
Automatic traffic sign extraction method and system based on binocular CCD camera Download PDFInfo
- Publication number
- CN112101299A CN112101299A CN202011200276.6A CN202011200276A CN112101299A CN 112101299 A CN112101299 A CN 112101299A CN 202011200276 A CN202011200276 A CN 202011200276A CN 112101299 A CN112101299 A CN 112101299A
- Authority
- CN
- China
- Prior art keywords
- traffic sign
- image
- sample
- image data
- extracted
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/582—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of traffic signs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/51—Indexing; Data structures therefor; Storage structures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
Abstract
The invention provides a binocular CCD camera-based automatic traffic sign extraction method and system, wherein the method comprises the following steps: image data processing: the method comprises the steps of image geographic reference processing, image data checking and modification; collecting, training and managing a traffic sign sample library: the method comprises the steps of collecting a traffic sign prior sample, training the sample, carrying out classification management, checking the sample, importing and exporting; extracting traffic sign information: automatically extracting coordinates and signboard information of the target image; and (3) traffic sign inspection: verifying and modifying the extracted result of the traffic sign, and storing the verified and modified data into a corresponding database; traffic sign result management: the method comprises the steps of importing and exporting the extraction result, carrying out classification management, viewing the result and the like. The invention greatly reduces the workload of manually browsing images and extracting the traffic signs, and can efficiently and normatively manage the sample information and the traffic sign results.
Description
Technical Field
The invention relates to the technical field of map plotting and surveying, in particular to a method and a system for automatically extracting a traffic sign based on a binocular CCD camera.
Background
In recent years, intelligent driving is rapidly developed in the intelligent transportation industry as a new field, and a high-precision map is an indispensable important component for intelligent driving. The intelligent driving mainly depends on various technologies such as induction sensors and positioning systems as supports to realize unmanned driving, but the conditions of road traffic are complicated, the perception capability of the sensors is greatly influenced by severe weather, information such as traffic signs cannot be identified and quickly responded by hardware equipment, the content and the position of the traffic signs are critical to safe driving of vehicles, and if the information cannot be identified quickly and accurately, the safety of unmanned vehicles is reduced.
In view of the above situation, the high-precision map makes up the deficiency, and as data pre-built in the navigation device, the high-precision map integrates the existing traffic sign information into the map, so that the high-precision map can guide the driving of the vehicle without being affected by the external environment, predict the traffic sign content in front of the road in advance, provide a decision basis for the intelligent driving device, and create safe, comfortable and stable riding experience.
The inventor of the present application finds that the method of the prior art has at least the following technical problems in the process of implementing the present invention:
in the prior art, generally, the type and the coordinates of the traffic sign are identified and extracted manually by reading image data, and because the traffic signs are of a plurality of types and a large number, the existing method needs to consume a large amount of manpower to extract the traffic sign, so that the extraction efficiency of the traffic sign is low.
Disclosure of Invention
The invention provides a binocular CCD camera-based automatic traffic sign extraction method and system, which are used for solving or at least partially solving the technical problem of low traffic sign extraction efficiency in the conventional method.
In order to solve the above technical problem, a first aspect of the present invention provides a method for automatically extracting a traffic sign based on a binocular CCD camera, including:
s1: processing the image geographic reference, matching coordinate information for the image data, checking the image data and repairing the problem image data, wherein the checked and repaired image data are stored in an image database;
s2: reading image data of a preset proportion from an image database, collecting traffic sign samples from the read image data, and storing the collected traffic sign samples into a traffic sign sample library, wherein the traffic sign sample library comprises the traffic sign samples and historical traffic sign samples collected from the read image data, and performing sample training on all the traffic sign samples stored in the traffic sign sample library, wherein the trained traffic sign samples comprise traffic sign images, corresponding traffic sign types and position information, and the position information is picture frame pixel coordinates of a traffic sign external rectangular frame;
s3: managing the collected and trained traffic sign samples;
s4: reading image data from an image database one by one to serve as image data to be extracted at the time, matching pixel colors and the shapes of the patterns formed by the pixels of the images to be extracted and the samples in a traffic sign sample library, and extracting the type of the traffic sign sample to serve as the type of a traffic sign target object in the images to be extracted when the matching degree of the pixel colors of the images to be extracted and the samples in a certain sample and the matching degree of the shapes of the patterns formed by the pixels meet requirements; calculating pixel coordinates of a circumscribed rectangular frame of the pixels matched with the traffic sign target object; acquiring longitude and latitude coordinate information of a central point of a traffic sign target object by shooting longitude and latitude coordinate information recorded by the binocular CCD camera of the image to be extracted and a central point coordinate calculated by picture frame pixel coordinates matched with the traffic sign target object, and storing the extracted traffic sign image, traffic sign type, picture frame pixel coordinates and the longitude and latitude coordinate information of the central point of the traffic sign corresponding to the traffic sign target object into a traffic sign operation library;
s5: reading the traffic sign image in the traffic sign operation library, reversely projecting a picture frame of an external rectangular frame of the traffic sign in the extracted image of the traffic sign target object, displaying the type of the traffic sign image, and judging whether the position and the type of the picture frame of the external rectangular frame are consistent with the actual position and the actual type of the traffic sign in the image data; if the position and the type are consistent, no modification is made, otherwise, the traffic sign image read from the traffic sign operation library is modified to be consistent with the actual position and the actual type.
In one embodiment, S1 includes:
matching the GPS receiving time according to the image shooting time to obtain the camera coordinate when the current image is shot;
matching the acquired camera coordinates with the image data;
and checking the problems of exposure, color cast, light spots and light columns in the image data, and performing image processing and repairing on the problem image data.
In one embodiment, S3 includes:
and inquiring, browsing, importing and exporting the collected and trained traffic sign sample.
In one embodiment, before calculating the coordinates of the center point through the coordinates of the frame pixels matched with the traffic sign object in step S4, the method further comprises:
and judging whether the coordinates of the central point of the traffic sign target object can be accurately calculated, if the number of pixels of the traffic sign on the corresponding image is less than the preset number of pixels, or the traffic sign is incompletely displayed in the image, or the traffic sign is worn and shielded, the coordinates of the central point of the target cannot be accurately calculated.
In one embodiment, the method further comprises:
when the coordinate of the central point of the traffic sign target object can be accurately calculated, the extracted traffic sign image, the traffic sign type and the coordinate information corresponding to the traffic sign target object are used as extraction result data, wherein the coordinate information comprises the coordinate of the central point of the target object and the picture frame pixel coordinate of the circumscribed rectangular frame;
and when the coordinate of the central point of the traffic sign target object cannot be accurately calculated, taking the extracted traffic sign image, traffic sign type and coordinate information corresponding to the traffic sign target object as identification result data, wherein the coordinate information comprises picture frame pixel coordinates of an external rectangular frame.
In one embodiment, the method further comprises:
if the extraction result is the extraction result data, the extraction result data is verified and modified and then submitted to a traffic sign result library together with the corresponding traffic sign image;
and if the extraction result is the identification result data, submitting the data to the traffic sign sample library together with the corresponding traffic sign image after verification and modification.
In one embodiment, the method further comprises S6: the traffic sign extraction result data are subjected to classified management, and the method specifically comprises the following steps:
inquiring, browsing, information statistics, importing and exporting the extracted result data of the traffic sign, wherein the inquired information comprises the following steps: type of sign, city of collection, time of collection, type of camera.
Based on the same inventive concept, the second aspect of the present invention provides an automatic traffic sign extraction system based on a binocular CCD camera, comprising:
the image data processing submodule is used for processing image geographical reference, matching coordinate information for the image data, checking the image data and repairing problem image data, wherein the checked and repaired image data are stored in the image database;
the traffic sign sample acquisition submodule is used for reading image data with a preset proportion from an image database, acquiring a traffic sign sample from the read image data, and storing the acquired traffic sign sample into a traffic sign sample library, wherein the traffic sign sample library comprises the traffic sign sample and historical traffic sign samples acquired from the read image data, and performing sample training on all traffic sign samples stored in the traffic sign sample library, wherein the trained traffic sign sample comprises a traffic sign image, a corresponding traffic sign type and position information, and the position information is picture frame pixel coordinates of a traffic sign external rectangular frame;
the traffic sign sample management submodule is used for managing the collected and trained traffic sign samples;
the traffic sign information extraction submodule is used for reading image data from an image database one by one to serve as image data to be extracted at the time, matching pixel colors and the figure shapes formed by pixels of the images to be extracted and samples in a traffic sign sample library, and extracting the type of the traffic sign sample to serve as the type of a traffic sign target object in the images to be extracted when the pixel color matching degree of the images to be extracted and a certain sample and the figure shape matching degree formed by pixels meet requirements; calculating pixel coordinates of a circumscribed rectangular frame of the pixels matched with the traffic sign target object; acquiring longitude and latitude coordinate information of a central point of a traffic sign target object by shooting longitude and latitude coordinate information recorded by the binocular CCD camera of the image to be extracted and calculating the central point coordinate through picture frame pixel coordinates matched with the traffic sign target object, and storing the extracted traffic sign image, the traffic sign type, the picture frame pixel coordinates and the longitude and latitude coordinate information of the central point of the traffic sign corresponding to the traffic sign target object into a traffic sign operation library;
the traffic sign checking submodule is used for reading the traffic sign images in the traffic sign operation library, inversely projecting the picture frames of the external rectangular frames of the traffic signs in the extracted images of the traffic sign target objects, displaying the types of the traffic sign images and judging whether the picture frame positions and the types of the external rectangular frames are consistent with the actual positions and the actual types of the traffic signs in the image data; if the position and the type are consistent, no modification is made, otherwise, the traffic sign image read from the traffic sign operation library is modified to be consistent with the actual position and the actual type.
In one embodiment, the system further comprises a traffic sign result management submodule for performing classification management on the traffic sign extraction result data.
One or more technical solutions in the embodiments of the present application have at least one or more of the following technical effects:
the invention discloses a traffic sign automatic extraction method based on a binocular CCD camera, which comprises the steps of collecting traffic sign samples from read image data, storing the collected traffic sign samples into a traffic sign sample library, wherein the traffic sign sample library comprises traffic sign samples and historical traffic sign samples collected from read preset proportion image data, taking all traffic sign samples in the traffic sign sample library as historical prior information, then reading image data in the image data one by one as an image to be extracted, matching the image to be extracted with samples in the traffic sign sample library so as to identify traffic sign target objects contained in the image to be extracted, if the matching is successful, the type of the traffic sign corresponding to the traffic sign image is the type of the traffic sign target objects, and calculating external picture frame pixel coordinates through the matched image pixel coordinates, and calculating the longitude and latitude coordinates of the central point through the camera and the picture frame pixel coordinates, thereby extracting the position information of the traffic sign target object, specifically comprising the picture frame pixel coordinates and the longitude and latitude coordinate information of the central point of the traffic sign, thereby realizing the automatic extraction of the traffic sign information in the image data.
Furthermore, the invention also further performs verification and check on the extracted result of the traffic sign so as to improve the accuracy of extraction and efficiently and normatively manage the sample information and the traffic sign result.
The embodiment also provides a traffic sign automatic extraction system based on the binocular CCD camera, and the system also manages image data, sample data, operation process data and result data in a database form, so that the problems of low efficiency and redundancy of a file data management form are avoided, and the problem of file loss caused in a data circulation process is also avoided; the mode of database management can be used for more conveniently inquiring and displaying data, the safety of the database can be ensured by using authority management, and the subsequent use of the traffic sign extraction result is more flexible and easy to operate; historical version control can be performed by using database management, the sample data volume is increased, and the accuracy of the automatic traffic sign extraction method is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
Fig. 1 is a flowchart of a binocular CCD camera-based automatic traffic sign extraction method according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a traffic sign automatic extraction system based on a binocular CCD camera according to an embodiment of the present invention.
Detailed Description
The invention provides a binocular CCD camera-based automatic traffic sign extraction method and system, which are used for solving the technical problem of low traffic sign extraction efficiency in the existing method, so that the purpose of improving the traffic sign information extraction efficiency is achieved.
In order to achieve the above object, the present invention is generally conceived as follows:
the method for automatically extracting the traffic sign based on the binocular CCD camera comprises the following steps: image data processing: the method comprises the steps of image geographic reference processing, image data checking and modification; collecting, training and managing a traffic sign sample library: the method comprises the steps of collecting a traffic sign prior sample, training the sample, carrying out classification management, checking the sample, importing and exporting; extracting traffic sign information: automatically extracting coordinates and signboard information of the target image; and (3) traffic sign inspection: verifying and modifying the extracted result of the traffic sign, and storing the verified and modified data into a corresponding database; traffic sign result management: the method comprises the steps of importing and exporting the extraction result, carrying out classification management, viewing the result and the like.
According to the binocular CCD camera-based automatic traffic sign extraction method and system, the historical prior traffic sign information provided by the traffic sign sample library is utilized to extract the traffic sign information (type and position) in the image, the extraction efficiency can be improved, the extracted position information is back projected into the image for inspection, the accuracy of the extraction result can be ensured, and the sample information and the traffic sign result are efficiently and normatively managed.
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example one
The embodiment provides an automatic extraction method of a traffic sign based on a binocular CCD camera, please refer to fig. 1, and the method includes:
s1: processing the image geographic reference, matching coordinate information for the image data, checking the image data and repairing the problem image data, wherein the checked and repaired image data are stored in an image database;
s2: reading image data of a preset proportion from an image database, collecting traffic sign samples from the read image data, and storing the collected traffic sign samples into a traffic sign sample library, wherein the traffic sign sample library comprises the traffic sign samples and historical traffic sign samples collected from the read image data, and performing sample training on all the traffic sign samples stored in the traffic sign sample library, wherein the trained traffic sign samples comprise traffic sign images, corresponding traffic sign types and position information, and the position information is picture frame pixel coordinates of a traffic sign external rectangular frame;
s3: managing the collected and trained traffic sign samples;
s4: reading image data from an image database one by one to serve as image data to be extracted at the time, matching pixel colors and the shapes of the patterns formed by the pixels of the images to be extracted and the samples in a traffic sign sample library, and extracting the type of the traffic sign sample to serve as the type of a traffic sign target object in the images to be extracted when the matching degree of the pixel colors of the images to be extracted and the samples in a certain sample and the matching degree of the shapes of the patterns formed by the pixels meet requirements; calculating pixel coordinates of a circumscribed rectangular frame of the pixels matched with the traffic sign target object; acquiring longitude and latitude coordinate information of a central point of a traffic sign target object by shooting longitude and latitude coordinate information recorded by the binocular CCD camera of the image to be extracted and calculating the central point coordinate through picture frame pixel coordinates matched with the traffic sign target object, and storing the extracted traffic sign image, the traffic sign type, the picture frame pixel coordinates and the longitude and latitude coordinate information of the central point of the traffic sign corresponding to the traffic sign target object into a traffic sign operation library;
s5: reading a traffic sign image in a traffic sign operation library, reversely projecting a picture frame of an external rectangular frame of the traffic sign in the extracted image of the traffic sign target object, displaying the type of the traffic sign image, and judging whether the position and the type of the picture frame of the external rectangular frame are consistent with the actual position and the actual type in the image data; if the two are consistent, the modification is not carried out, and if the two are not consistent, the modification is carried out.
Specifically, the image database constructed in S1 is the basis for constructing the traffic sign sample library in step S2 and extracting the traffic signs in step S4.
The image data of the preset ratio read from the image database in S2 can be set according to practical situations, for example, 5%, 10%, 20%, and so on. The historical traffic sign sample refers to the historical known traffic sign image and the corresponding traffic sign type. The traffic sign sample collection can provide an image visualization interface and a labeling tool. And carrying out sample training by taking the traffic sign sample data in the traffic sign sample library as prior data, and providing the sample data to the automatic traffic sign extraction step S4 as matching reference data. That is, step S2 provides a priori traffic sign information for the traffic sign extraction of step S4.
In S4, when matching the image to be extracted with the traffic sign sample library, the image to be extracted is matched with the samples in the traffic sign sample library in terms of pixel color and graphic shape formed by pixels, so as to identify the traffic sign target object included in the image to be extracted.
In the specific implementation process, pixel color matching, namely RGB value matching, is performed firstly, if the difference values of three values in the RGB value of the pixel in the image to be extracted and the RGB value of the image in the pixel coordinate range of the traffic sign sample frame are all within the threshold range, the pixel color matching degree is indicated to meet the requirement, for example, the difference values of the RGB values are all smaller than 3, the pixel color matching is considered to be successful, then, the matching of the graph shape formed by the pixels is further performed, namely, the graph shape formed by all pixel points of which the pixel color matching degree in the image to be extracted meets the requirement is matched with the graph shape formed by all pixel points of which the pixel color in the sample corresponds to the matching, and if the graph shape types formed by the two pixel points are the same, the graph shape matching degree is considered to meet the requirement, namely, the graph shape matching is successful. For example, the shape of the graph formed by all the pixels of which the pixel color matching degree meets the requirement in the image to be extracted is circular, the shape of the graph formed by all the pixels correspondingly matched in the traffic sign sample is also circular, and the shapes of the two graphs are the same in type, so that the matching degree of the graph shape meets the requirement, otherwise, the shapes of the graphs are different, and if one is triangular and the other is circular, the matching degree of the graph shape does not meet the requirement. When the pixel color matching degree and the graph shape matching degree meet the requirements, matching is successful, namely the traffic sign which is consistent with the traffic sign sample type exists in the image to be extracted, the traffic sign type corresponding to the traffic sign image stored in the traffic sign sample library is used as the type of a traffic sign target object, and then the pixel coordinate of the circumscribed rectangular frame of the pixel which is successfully matched with the traffic sign target object is calculated; the binocular CCD camera for shooting the image to be extracted records longitude and latitude coordinate information, and the longitude and latitude coordinate information of the central point of the traffic sign target object can be obtained according to the calculated pixel coordinates of the external rectangular frame matched with the traffic sign target object and the recorded longitude and latitude coordinate information.
In the specific implementation process, the whole image is read from the image database constructed in the step S1, then the pixel matching is performed on the read whole image and the traffic sign image stored in the traffic sign sample library, if the matching is unsuccessful, the process is ended, and if the matching is successful, the corresponding traffic sign type is extracted, that is, the traffic sign type corresponding to the traffic sign image stored in the traffic sign sample library matched with the image is extracted, and then the position information, that is, the picture frame pixel coordinates of the external rectangular frame is extracted. Further, the extraction result may be divided into extraction result data (the coordinates of the central point may be calculated) and extraction result data (the coordinates of the central point may not be calculated) according to whether the coordinates of the central point of the traffic sign target object may be accurately calculated.
Wherein, when calculating the longitude and latitude coordinates of the central point: the method comprises the steps of firstly calculating picture frame pixel coordinates of an external rectangular frame of pixels matched with a traffic sign target object, then calculating central pixel coordinates through the picture frame pixel coordinates of the external rectangular frame, and converting the central pixel coordinates and longitude and latitude coordinate information recorded by a CCD camera into longitude and latitude coordinates of a central point. The traffic sign types comprise forbidden signs, indicating signs and subclasses thereof, the road indicating signs are of a large class and the like, and the coordinate information comprises longitude and latitude coordinates and elevation information. The information stored in the traffic sign operation library comprises traffic sign data obtained when matching can be successfully carried out in one operation, wherein the one operation refers to a process of automatically extracting the traffic signs after image data are read once. Through step S4, the information provided by the traffic sign sample library may be used as historical prior information, so as to achieve automatic extraction of the traffic sign, mainly extracting the type of the traffic sign and the location information in the image.
Specifically, the steps S1-S4 can realize automatic extraction of the traffic sign. In order to further improve the accuracy of extraction, the extraction result of step S4 is further checked and verified by step S5. The back projection is a process of displaying the picture frame pixel coordinates of the traffic sign circumscribed rectangle frame obtained in S4 on the image. Since the frame pixel coordinates of the bounding rectangle are themselves obtained from the image, they are now displayed in reverse on the image, hence the name backprojection.
And judging whether the picture frame position and the type of the external rectangular frame are consistent with the traffic sign type in the image data, comparing the picture frame position of the external rectangular frame of the traffic sign image read from the traffic sign operation library with the actual position of the image data, comparing the type of the traffic sign image read from the traffic sign operation library with the actual type of the image data, and if the picture frame position and the type of the external rectangular frame are the same as the actual position of the image data, not modifying the image data information of the traffic sign operation library. Otherwise (there is a discrepancy between the position and the type, or both of them), the information (position and type) of the traffic sign image read from the traffic sign operation library is modified to be consistent with the actual information (actual position and actual type) of the image data.
For example, if the frame position of the circumscribed rectangular frame of the traffic sign image read from the traffic sign operation library is not at the actual traffic sign (actual position) corresponding to the image, the frame position of the circumscribed rectangular frame of the read traffic sign image needs to be moved to the actual traffic sign; if the type of the traffic sign image read from the traffic sign operation library is a speed limit sign, but the actual type of the image is a height limit sign, the speed limit type needs to be modified into the height limit type.
In S5, the pixel coordinates of the image are recorded during the extraction process, and the extracted circumscribed rectangle is displayed on the image.
In one embodiment, S1 includes:
matching the GPS receiving time according to the image shooting time to obtain the camera coordinate when the current image is shot;
matching the acquired camera coordinates with the image data;
and checking the problems of exposure, color cast, light spots and light columns in the image data, and performing image processing and repairing on the problem image data.
In one embodiment, S3 includes:
and inquiring, browsing, importing and exporting the collected and trained traffic sign sample.
Specifically, the traffic sign sample data query supports various modes of classified query, including sign types, collection cities, collection time, camera models and the like.
In one embodiment, before calculating the coordinates of the center point by the coordinates of the frame pixels matched by the traffic sign object, the method further comprises:
and judging whether the coordinates of the central point of the traffic sign target object can be accurately calculated, if the number of pixels of the traffic sign on the corresponding image is less than the preset number of pixels, or the traffic sign is incompletely displayed in the image, the coordinates of the central point of the target cannot be accurately calculated.
In one embodiment, the method further comprises:
when the coordinate of the central point of the traffic sign target object can be accurately calculated, the extracted traffic sign image, traffic sign type and coordinate information corresponding to the traffic sign target object are used as extraction result data, wherein the coordinate information comprises the coordinate of the central point of the target object and the coordinate of a picture frame pixel;
and when the coordinate of the central point of the traffic sign target object cannot be accurately calculated, taking the extracted traffic sign image, traffic sign type and coordinate information corresponding to the traffic sign target object as identification result data, wherein the coordinate information comprises picture frame pixel coordinates.
Specifically, longitude and latitude coordinate information of a central point of the traffic sign target object is obtained through longitude and latitude coordinates of a binocular CCD camera corresponding to the image and central pixel coordinates calculated through picture frame pixel coordinates of the traffic sign target object, and in addition, elevation information of the image can be obtained. The traffic sign image in the extraction result data and the identification result data may include an image number.
When the coordinate of the central point can be accurately calculated, information such as the coordinate, the type, the picture frame pixel coordinate, the corresponding image number and the like of the target object is generated into extraction result data; and when the longitude and latitude coordinates of the central point of the accurate target object cannot be calculated, generating information such as the type of the target object, the picture frame pixel coordinates, the corresponding image number and the like into identification result data.
In one embodiment, the method further comprises:
if the extraction result is the extraction result data, the extraction result data is verified and modified and then submitted to a traffic sign result library together with the corresponding traffic sign image;
and if the extraction result is the identification result data, submitting the data to the traffic sign sample library together with the corresponding traffic sign image after verification and modification.
Specifically, when the data is the extracted result data, the data verified and modified in step S5 is stored in the traffic sign result library, where the information stored in the traffic sign result library includes traffic sign information in which the coordinates of the central point can be accurately calculated in the extracted result, specifically, the information includes a traffic sign image, a traffic sign type, and coordinate information, where the coordinate information includes the coordinates of the central point and the coordinates of frame pixels of the circumscribed rectangular frame.
The traffic sign sample library stores traffic sign samples collected from the read part of the current image data, historical traffic sign samples and traffic sign information of which the coordinates of the central point cannot be accurately calculated in the extraction result, wherein the traffic sign samples comprise traffic sign images, corresponding traffic sign types and position information, and the position information is picture frame pixel coordinates of a rectangular frame externally connected with the traffic signs; the traffic sign information which cannot accurately calculate the coordinates of the central point in the extraction result comprises traffic sign images, traffic sign types and coordinate information, wherein the coordinate information comprises picture frame pixel coordinates of an external rectangular frame.
The information stored in the traffic sign operation library comprises traffic sign data obtained in one-time operation and when the matching is successful, the traffic sign data specifically comprises traffic sign images, traffic sign types and coordinate information corresponding to traffic sign objects, and the coordinate information comprises picture frame pixel coordinates of an external rectangular frame.
In one embodiment, the method further includes step S6:
and carrying out classified management on the extracted result data of the traffic signs.
In one embodiment, the S6 performs classification management on the traffic sign extraction result data, including:
inquiring, browsing, information statistics, importing and exporting the extracted result data of the traffic sign, wherein the inquired information comprises the following steps: type of sign, city of collection, time of collection, type of camera.
Example two
Based on the same inventive concept, the present embodiment provides an automatic traffic sign extracting system based on a binocular CCD camera, please refer to fig. 2, the system includes:
the image data processing submodule is used for processing image geographical reference, matching coordinate information for the image data, checking the image data and repairing problem image data, wherein the checked and repaired image data are stored in the image database;
the traffic sign sample acquisition submodule is used for reading image data with a preset proportion from an image database, acquiring a traffic sign sample from the read image data, and storing the acquired traffic sign sample into a traffic sign sample library, wherein the traffic sign sample library comprises the traffic sign sample and historical traffic sign samples acquired from the read image data, and performing sample training on all traffic sign samples stored in the traffic sign sample library, wherein the trained traffic sign sample comprises a traffic sign image, a corresponding traffic sign type and position information, and the position information is picture frame pixel coordinates of a traffic sign external rectangular frame;
the traffic sign sample management submodule is used for managing the collected and trained traffic sign samples;
the traffic sign information extraction submodule is used for reading image data from an image database one by one to serve as image data to be extracted at the time, matching pixel colors and the figure shapes formed by pixels of the images to be extracted and samples in a traffic sign sample library, and extracting the type of the traffic sign sample to serve as the type of a traffic sign target object in the images to be extracted when the pixel color matching degree of the images to be extracted and a certain sample and the figure shape matching degree formed by pixels meet requirements; calculating pixel coordinates of a circumscribed rectangular frame of the pixels matched with the traffic sign target object; acquiring longitude and latitude coordinate information of a central point of a traffic sign target object by shooting longitude and latitude coordinate information recorded by the binocular CCD camera of the image to be extracted and calculating the central point coordinate through picture frame pixel coordinates matched with the traffic sign target object, and storing the extracted traffic sign image, the traffic sign type, the picture frame pixel coordinates and the longitude and latitude coordinate information of the central point of the traffic sign corresponding to the traffic sign target object into a traffic sign operation library;
the traffic sign checking submodule is used for reading the traffic sign images in the traffic sign operation library, inversely projecting the picture frames of the external rectangular frames of the traffic signs in the extracted images of the traffic sign target objects, displaying the types of the traffic sign images and judging whether the picture frame positions and the types of the external rectangular frames are consistent with the actual positions and the actual types of the traffic signs in the image data; if the position and the type are consistent, no modification is made, otherwise, the traffic sign image read from the traffic sign operation library is modified to be consistent with the actual position and the actual type.
In one embodiment, the system further comprises a traffic sign result management submodule for performing classification management on the traffic sign extraction result data.
Specifically, the image data processing submodule is used for image geographic reference processing, matching coordinate information for image data, and is also used for image data inspection and repair, and providing an image data visual inspection interface and an image repair function.
The traffic sign sample collection submodule is used for collecting and training a traffic sign sample, the module provides an image visual interface and a labeling function, and after the traffic sign sample collection submodule collects and trains, data management can be performed on the traffic sign sample management submodule, and the traffic sign sample collection submodule comprises: sample query, browse, import and export.
The traffic sign information extraction submodule takes the traffic sign sample library data as prior data, automatically identifies potential traffic sign target objects in the images according to the data after image processing by the traffic sign information extraction submodule, matches the identified traffic sign target objects in the prior data (the traffic sign sample library), acquires the traffic sign type of the traffic sign target objects and the picture frame pixel coordinates of the target external connection rectangular frame, and then calculates the central point coordinates of the target objects:
if the coordinate of the central point of the target object can be accurately calculated, generating an extraction result datum by using the information such as the coordinate, the type, the picture frame pixel coordinate, the corresponding image number and the like of the target object; if the coordinates of the center point of the target object cannot be accurately calculated, the type of the target object, the picture frame pixel coordinates, the corresponding image number and other information are generated into identification result data, and both the extraction result data and the identification result data are stored in a traffic sign operation library.
For example: the traffic sign is too far from the camera, resulting in too few pixels (less than 36 pixels) on the image for the sign to compute; or the sign is missing a portion of the image but is recognized, such calculated coordinates are also inaccurate.
The traffic sign checking sub-module is used for providing checking and modifying functions for data in the traffic sign operation library, displaying an external picture frame of the traffic sign by reversely projecting in the extracted image of the target object through browsing the image, and displaying the type information of the traffic sign, and checking whether the position and the type of the picture frame are consistent with the ground object (actual position and type) in the image by an operator; if the two are consistent, the modification is not made, if the two are not consistent, the modification is required to be correct,
after the extracted result data is verified to be correct, submitting the result data and the corresponding image to a traffic sign result base;
and after the identification result data is verified to be correct, submitting the identification result data and the corresponding image to a traffic sign sample library.
The traffic sign sample library is suitable for storing traffic sign images and corresponding types and is used as a priori knowledge base for machine learning, so that the identification accuracy is improved; and the traffic sign result library is used for storing the traffic sign image, the type, the longitude and latitude coordinates of the central point and the coordinates of the external picture frame pixels as results. The traffic sign result library supports the functions of inquiry, browsing, information statistics, import and export.
The following describes the traffic sign automatic extraction system based on the binocular CCD camera according to the present invention in detail by using a specific example.
As shown in fig. 2, the traffic sign automatic extraction task management platform is used for task state flow management between data processing modules.
After the traffic sign automatically extracts an operation task management platform to newly establish an image processing task, starting an image data processing submodule;
the image data processing submodule can automatically process image geographic reference, then starts image data inspection, and an operator checks whether the image has the problems of exposure, color cast, light spots, light beams and the like, if the problems exist, the image processing function is used for repairing the problem image, the processed image data enters an image database, and meanwhile, the task state of 'image data processing is finished' is sent to the automatic operation task extraction management platform of the traffic sign.
At the moment, the automatic traffic sign extraction task management platform can select to collect traffic sign samples or extract traffic sign information;
if the traffic sign sample collection is selected, creating a traffic sign sample collection task in the automatic traffic sign extraction task management platform, and starting a traffic sign sample collection submodule;
the traffic sign sample collection submodule provides an image visual interface and can carry out manual browsing, a label is added on an image with a traffic sign, the label information comprises the type of the traffic sign and a picture frame of a rectangular frame externally connected with the traffic sign, the traffic sign image and the added label information form sample information of the traffic sign, the collected sample is trained into a prior traffic sign sample through a sample training function provided by the traffic sign sample collection submodule, and the sample information is sent into a traffic sign sample library and sent to a traffic sign automatic extraction task management platform with a task state of 'sample collection completion';
after the access authority is verified in the traffic sign automatic extraction operation task management platform, a traffic sign sample management submodule is started;
the traffic sign sample management submodule provides operations of inquiring, browsing, importing, exporting and the like of a traffic sign sample library: the query operation provides query of information such as the type of a label, a collection city, collection time, the type of a camera, an operator and the like; the browsing operation provides visual display of the query result; importing external sample data into a traffic sign sample library by an importing operation; the export operation provides data for exporting the query content of the traffic sign sample library into a plurality of selectable formats, and all operations in the module are recorded into the automatic traffic sign extraction task management platform.
If the traffic sign information extraction is selected, the task state is converted from 'image data processing is finished' into 'information to be extracted', and a traffic sign information extraction submodule is started;
the traffic sign information extraction submodule automatically extracts the type of a traffic sign target object by reading an image database and matching image data with samples in the traffic sign sample database, calculates to obtain an external rectangular picture frame pixel coordinate, then automatically calculates the longitude and latitude coordinate of the central point of the traffic sign target object through the longitude and latitude coordinate of a binocular CCD camera contained in an image and externally-imported camera parameters, generates extraction result data when the central coordinate of the target object is accurate, generates identification result data when the central coordinate of the target object is inaccurate, stores the two data in a traffic sign operation library, and simultaneously sends a task state of 'operation data to be checked' to an automatic operation task extraction management platform of the traffic sign.
The traffic sign automatic extraction operation task management platform can perform a traffic sign inspection task and start a traffic sign inspection submodule;
the traffic sign inspection sub-module reads a traffic sign operation library, an operator verifies the type of the traffic sign and position information inversely projected into the image by browsing the image in a visual interface, wrong data is manually modified, the verified extraction result data is input into a traffic sign result library, the identification result data is input into a traffic sign sample library, and meanwhile, the task state of 'operation data inspection completion' is sent to an automatic traffic sign extraction operation task management platform;
after the access authority is verified in the traffic sign automatic extraction operation task management platform, a traffic sign result management submodule is started;
the traffic sign result management submodule provides operations of querying, counting, browsing, importing, exporting and the like of a traffic sign result library: the query operation provides query of information such as the type of a label, a collection city, collection time, the type of a camera, an operator and the like; the statistical operation may show the query results through selectable chart styles; the browsing operation provides visual display of the query result; external sample data provided by the import operation is imported into a traffic sign result library; the export operation provides data for exporting the query content of the traffic sign result base into a plurality of selectable formats, and all operations in the module are recorded into the automatic traffic sign extraction task management platform.
The production operation process of the system for automatically extracting the traffic sign comprises the following steps:
firstly, image geo-referencing processing is carried out on image data to obtain coordinates of a corresponding camera, then a problem image is checked and repaired, and the checked image is stored in an image database.
And then automatically extracting the traffic signs, automatically reading the data of the image database, matching the data with the samples of the traffic sign sample library, generating a traffic sign identification result and a traffic sign extraction result, and storing the results in a traffic sign operation library.
The traffic sign sample library stores traffic sign samples comprising historical traffic sign sample data and traffic sign samples made of image data read from the image database in a certain proportion.
The data of the traffic sign operation library are checked and modified by operators, the checked traffic sign identification result is stored in a traffic sign sample library, and the traffic sign extraction result is stored in a traffic sign result library.
Effects and effects of the embodiments
The method for automatically extracting the traffic sign based on the binocular CCD camera, which is adopted by the embodiment, takes the fact that the imaging result of the CCD camera is regular on the pixel level of the traffic sign, so that a traffic sign sample library is constructed, wherein the traffic sign sample library comprises a traffic sign sample and a historical traffic sign sample which are collected from read image data, and the traffic sign sample and the historical traffic sign sample are used as prior traffic sign information to automatically identify and extract a traffic sign target object in an image.
And the operator only needs to verify and modify the extraction result, the automatic extraction process does not need manual participation, the manual participation degree is low, and the occurrence probability of operation errors is reduced.
According to the method and the system for automatically extracting the traffic signs based on the binocular CCD camera, the traffic sign automatic extraction task management platform is used for creating data tasks and transferring data, the working procedures are distributed and integrated in the operation process of automatically extracting the traffic signs, and operation flow supervision and standardized management can be effectively carried out.
The system of the embodiment also manages the image data, the sample data, the operation process data and the result data in the form of a database, so that the problems of low efficiency and redundancy of a file data management form are avoided, and the problem of file loss caused in a data circulation process can be avoided; the mode of database management can be used for more conveniently inquiring and displaying data, the safety of the database can be ensured by using authority management, and the subsequent use of the traffic sign extraction result is more flexible and easy to operate; historical version control can be performed by using database management, the sample data volume is increased, and the accuracy of the automatic traffic sign extraction method is improved.
After the operation system of the embodiment is applied, the standardized management of the operation process is realized from the aspect of flow management; from the aspect of data management, the safety of data is ensured, and the error occurrence rate of manual operation is reduced; from the aspect of automatic extraction of the traffic sign, the operation time is shortened, the manual operation amount is reduced, and the cost in the manual aspect is saved.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various modifications and variations can be made in the embodiments of the present invention without departing from the spirit or scope of the embodiments of the invention. Thus, if such modifications and variations of the embodiments of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to encompass such modifications and variations.
Claims (9)
1. A traffic sign automatic extraction method based on a binocular CCD camera is characterized by comprising the following steps:
s1: processing the image geographic reference, matching coordinate information for the image data, checking the image data and repairing the problem image data, wherein the checked and repaired image data are stored in an image database;
s2: reading image data of a preset proportion from an image database, collecting traffic sign samples from the read image data, and storing the collected traffic sign samples into a traffic sign sample library, wherein the traffic sign sample library comprises the traffic sign samples and historical traffic sign samples collected from the read image data, and performing sample training on all the traffic sign samples stored in the traffic sign sample library, wherein the trained traffic sign samples comprise traffic sign images, corresponding traffic sign types and position information, and the position information is picture frame pixel coordinates of a traffic sign external rectangular frame;
s3: managing the collected and trained traffic sign samples;
s4: reading image data from an image database one by one to serve as image data to be extracted at the time, matching pixel colors and the shapes of the patterns formed by the pixels of the images to be extracted and the samples in a traffic sign sample library, and extracting the type of the traffic sign sample to serve as the type of a traffic sign target object in the images to be extracted when the matching degree of the pixel colors of the images to be extracted and the samples in a certain sample and the matching degree of the shapes of the patterns formed by the pixels meet requirements; calculating pixel coordinates of a circumscribed rectangular frame of the pixels matched with the traffic sign target object; acquiring longitude and latitude coordinate information of a central point of a traffic sign target object by shooting longitude and latitude coordinate information recorded by the binocular CCD camera of the image to be extracted and calculating the central point coordinate through picture frame pixel coordinates matched with the traffic sign target object, and storing the extracted traffic sign image, the traffic sign type, the picture frame pixel coordinates and the longitude and latitude coordinate information of the central point of the traffic sign corresponding to the traffic sign target object into a traffic sign operation library;
s5: reading the traffic sign image in the traffic sign operation library, reversely projecting a picture frame of an external rectangular frame of the traffic sign in the extracted image of the traffic sign target object, displaying the type of the traffic sign image, and judging whether the position and the type of the picture frame of the external rectangular frame are consistent with the actual position and the actual type of the traffic sign in the image data; if the position and the type are consistent, no modification is made, otherwise, the traffic sign image read from the traffic sign operation library is modified to be consistent with the actual position and the actual type.
2. The method of claim 1, wherein S1 includes:
matching the GPS receiving time according to the image shooting time to obtain the camera coordinate when the current image is shot;
matching the acquired camera coordinates with the image data;
and checking the problems of exposure, color cast, light spots and light columns in the image data, and performing image processing and repairing on the problem image data.
3. The method of claim 1, wherein S3 includes:
and inquiring, browsing, importing and exporting the collected and trained traffic sign sample.
4. The method of claim 2, wherein before calculating the coordinates of the center point through the coordinates of the frame pixels matched with the traffic sign object in step S4, the method further comprises:
and judging whether the coordinate of the central point of the traffic sign target object can be accurately calculated, if the number of pixels of the traffic sign on the corresponding image is less than the preset number of pixels, or the traffic sign is incompletely displayed in the image, or the traffic sign is worn and shielded, the coordinate of the central point of the traffic sign target object cannot be accurately calculated.
5. The method of claim 4, wherein the method further comprises:
when the coordinate of the central point of the traffic sign target object can be accurately calculated, the extracted traffic sign image, the traffic sign type and the coordinate information corresponding to the traffic sign target object are used as extraction result data, wherein the coordinate information comprises the coordinate of the central point of the target object and the picture frame pixel coordinate of the circumscribed rectangular frame;
and when the coordinate of the central point of the traffic sign target object cannot be accurately calculated, taking the extracted traffic sign image, traffic sign type and coordinate information corresponding to the traffic sign target object as identification result data, wherein the coordinate information comprises picture frame pixel coordinates of an external rectangular frame.
6. The method of claim 5, wherein the method further comprises:
if the extraction result is the extraction result data, the extraction result data is verified and modified and then submitted to a traffic sign result library together with the corresponding traffic sign image;
and if the extraction result is the identification result data, submitting the data to the traffic sign sample library together with the corresponding traffic sign image after verification and modification.
7. The method according to claim 1, further comprising S6: the traffic sign extraction result data are subjected to classified management, and the classified management comprises the following steps:
inquiring, browsing, information statistics, importing and exporting the extracted result data of the traffic sign, wherein the inquired information comprises the following steps: type of sign, city of collection, time of collection, type of camera.
8. The utility model provides a traffic sign automatic extraction system based on binocular CCD camera which characterized in that includes:
the image data processing submodule is used for processing image geographical reference, matching coordinate information for the image data, checking the image data and repairing problem image data, wherein the checked and repaired image data are stored in the image database;
the traffic sign sample acquisition submodule is used for reading image data with a preset proportion from an image database, acquiring a traffic sign sample from the read image data, and storing the acquired traffic sign sample into a traffic sign sample library, wherein the traffic sign sample library comprises the traffic sign sample and historical traffic sign samples acquired from the read image data, and performing sample training on all traffic sign samples stored in the traffic sign sample library, wherein the trained traffic sign sample comprises a traffic sign image, a corresponding traffic sign type and position information, and the position information is picture frame pixel coordinates of a traffic sign external rectangular frame;
the traffic sign sample management submodule is used for managing the collected and trained traffic sign samples;
the traffic sign information extraction submodule is used for reading image data from an image database one by one to serve as image data to be extracted at the time, matching pixel colors and the figure shapes formed by pixels of the images to be extracted and samples in a traffic sign sample library, and extracting the type of the traffic sign sample to serve as the type of a traffic sign target object in the images to be extracted when the pixel color matching degree of the images to be extracted and a certain sample and the figure shape matching degree formed by pixels meet requirements; calculating pixel coordinates of a circumscribed rectangular frame of the pixels matched with the traffic sign target object; acquiring longitude and latitude coordinate information of a central point of a traffic sign target object by shooting longitude and latitude coordinate information recorded by the binocular CCD camera of the image to be extracted and calculating the central point coordinate through picture frame pixel coordinates matched with the traffic sign target object, and storing the extracted traffic sign image, the traffic sign type, the picture frame pixel coordinates and the longitude and latitude coordinate information of the central point of the traffic sign corresponding to the traffic sign target object into a traffic sign operation library;
the traffic sign checking submodule is used for reading the traffic sign images in the traffic sign operation library, inversely projecting the picture frames of the external rectangular frames of the traffic signs in the extracted images of the traffic sign target objects, displaying the types of the traffic sign images and judging whether the picture frame positions and the types of the external rectangular frames are consistent with the actual positions and the actual types of the traffic signs in the image data; if the position and the type are consistent, no modification is made, otherwise, the traffic sign image read from the traffic sign operation library is modified to be consistent with the actual position and the actual type.
9. The system of claim 8, further comprising a traffic sign result management sub-module for performing classification management on the traffic sign extraction result data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011200276.6A CN112101299B (en) | 2020-11-02 | 2020-11-02 | Automatic traffic sign extraction method and system based on binocular CCD camera |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011200276.6A CN112101299B (en) | 2020-11-02 | 2020-11-02 | Automatic traffic sign extraction method and system based on binocular CCD camera |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112101299A true CN112101299A (en) | 2020-12-18 |
| CN112101299B CN112101299B (en) | 2021-02-19 |
Family
ID=73785397
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011200276.6A Active CN112101299B (en) | 2020-11-02 | 2020-11-02 | Automatic traffic sign extraction method and system based on binocular CCD camera |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112101299B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110109476A1 (en) * | 2009-03-31 | 2011-05-12 | Porikli Fatih M | Method for Recognizing Traffic Signs |
| CN103544484A (en) * | 2013-10-30 | 2014-01-29 | 广东工业大学 | Traffic sign identification method and system based on SURF |
| KR20170058512A (en) * | 2015-11-18 | 2017-05-29 | 주식회사 만도 | Apparatus for traffic signs recognition and method thereof |
| CN110501018A (en) * | 2019-08-13 | 2019-11-26 | 广东星舆科技有限公司 | A kind of traffic mark board information collecting method for serving high-precision map producing |
-
2020
- 2020-11-02 CN CN202011200276.6A patent/CN112101299B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110109476A1 (en) * | 2009-03-31 | 2011-05-12 | Porikli Fatih M | Method for Recognizing Traffic Signs |
| CN103544484A (en) * | 2013-10-30 | 2014-01-29 | 广东工业大学 | Traffic sign identification method and system based on SURF |
| KR20170058512A (en) * | 2015-11-18 | 2017-05-29 | 주식회사 만도 | Apparatus for traffic signs recognition and method thereof |
| CN110501018A (en) * | 2019-08-13 | 2019-11-26 | 广东星舆科技有限公司 | A kind of traffic mark board information collecting method for serving high-precision map producing |
Non-Patent Citations (3)
| Title |
|---|
| ALTAF ALAM，ET AL: "《Indian Traffic Sign Detection and Recognition》", 《INTERNATIONAL JOURNAL OF INTELLIGENT TRANSPORTATION SYSTEMS RESEARCH》 * |
| 张卡 等: "《针对车载移动测量系统的交通标志自动检测》", 《武汉大学学报 信息科学版》 * |
| 袁家明: "《基于车载近景摄影测量的道路交通标线信息自动提取方法研究》", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112101299B (en) | 2021-02-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110148196B (en) | Image processing method and device and related equipment | |
| KR102094341B1 (en) | System for analyzing pot hole data of road pavement using AI and for the same | |
| CN110969592B (en) | Image fusion method, automatic driving control method, device and equipment | |
| CN109782364B (en) | Traffic sign board missing detection method based on machine vision | |
| CN107430815A (en) | Method and system for automatic identification parking area | |
| CN111221808A (en) | Unattended high-precision map quality inspection method and device | |
| CN109271861B (en) | Multi-scale fusion point cloud traffic signboard automatic extraction method | |
| CN111444810A (en) | Traffic light information identification method, device, equipment and storage medium | |
| CN113252053A (en) | High-precision map generation method and device and electronic equipment | |
| CN112633661A (en) | BIM-based emergency dispatching command method, system, computer equipment and readable medium | |
| CN115164918A (en) | Semantic point cloud map construction method and device and electronic equipment | |
| CN115424265A (en) | Point cloud semantic segmentation and labeling method and system | |
| CN113255578B (en) | Traffic identification recognition method and device, electronic equipment and storage medium | |
| CN111316324A (en) | Automatic driving simulation system, method, equipment and storage medium | |
| CN112101299B (en) | Automatic traffic sign extraction method and system based on binocular CCD camera | |
| CN110415299B (en) | Vehicle position estimation method based on set guideboard under motion constraint | |
| CN108427935B (en) | Street view comparison image generation method and device | |
| CN113902927B (en) | Comprehensive information processing method fusing image and point cloud information | |
| CN114359859A (en) | Method and device for processing target object with shielding and storage medium | |
| CN114677458A (en) | Road mark generation method and device for high-precision map, electronic equipment and storage medium | |
| CN113048988B (en) | Method and device for detecting change elements of scene corresponding to navigation map | |
| CN112632198A (en) | Map data display method and device and electronic equipment | |
| CN113222070A (en) | Automatic labeling method and system for simulation image data | |
| CN115238119B (en) | Method and device for analyzing position information of oblique photography photo | |
| CN112507887B (en) | Intersection sign extracting and associating method and device |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |