CN113392804A - Multi-angle-based traffic police target data set scene construction method and system - Google Patents
Multi-angle-based traffic police target data set scene construction method and system Download PDFInfo
- Publication number
- CN113392804A CN113392804A CN202110749503.9A CN202110749503A CN113392804A CN 113392804 A CN113392804 A CN 113392804A CN 202110749503 A CN202110749503 A CN 202110749503A CN 113392804 A CN113392804 A CN 113392804A
- Authority
- CN
- China
- Prior art keywords
- data set
- traffic police
- weather
- classification
- scene
- 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
- 238000010276 construction Methods 0.000 title claims abstract description 39
- 238000002372 labelling Methods 0.000 claims abstract description 89
- 238000000034 method Methods 0.000 claims abstract description 32
- 238000012545 processing Methods 0.000 claims description 28
- 238000012549 training Methods 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 9
- 238000012544 monitoring process Methods 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000005286 illumination Methods 0.000 claims description 6
- 239000012535 impurity Substances 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 4
- 230000002708 enhancing effect Effects 0.000 claims description 3
- 230000003313 weakening effect Effects 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 11
- 230000000694 effects Effects 0.000 description 7
- 230000009286 beneficial effect Effects 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 5
- 238000001514 detection method Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 238000012986 modification Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 238000011160 research Methods 0.000 description 5
- 230000009471 action Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 241000283070 Equus zebra Species 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 238000009826 distribution Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 239000000779 smoke Substances 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Tourism & Hospitality (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a multi-angle-based scene construction method and a system for a traffic police target data set, wherein the method comprises the following steps: collecting a first weather condition data set, classifying the first weather condition data set to generate a first weather condition data subset and a second weather condition data subset; respectively obtaining first and second weather visibility; constructing a first angle traffic police data set; obtaining a first user image set; the first user image set is classified according to distance to generate a second angle traffic police data set; classifying the gender difference of the first user image set to generate a third angle traffic police data set; and constructing a multi-scene duty data set of the first user, and labeling the characteristics. The method solves the technical problems that in the prior art, the construction of a data set aiming at a traffic police target alone is not perfect enough, and the construction can not be carried out based on a real road traffic scene, so that the traffic police target data set can not be fully applied to accurately judge the command gesture of a traffic police.
Description
Technical Field
The invention relates to the technical field of intelligent driving, in particular to a scene construction method and a scene construction system of a traffic police target data set based on multiple angles.
Background
The biggest challenge facing current intelligent driving technology is safety. For example, the vehicle may keep a safe distance from the vehicle ahead, may pass through a crossing under traffic light indication, may pass through a zebra crossing, a school, or the like at a reduced speed, and the like. To solve these problems, an intelligent driving system is required to accurately sense various environments. A mature intelligent driving system not only needs to accurately identify fixed signals such as traffic lights, pavement markers and the like, but also needs to detect a traffic police target and can identify the traffic police command action correspondingly.
However, in the process of implementing the technical solution of the invention in the embodiments of the present application, the inventors of the present application find that the above-mentioned technology has at least the following technical problems:
in the prior art, the data set for the traffic police target alone is not well constructed, and the data set cannot be constructed based on a real road traffic scene, so that the traffic police target data set cannot be fully applied to accurately judge the command gesture of the traffic police.
Disclosure of Invention
Aiming at the defects in the prior art, the embodiment of the application aims to solve the technical problems that the data set construction aiming at the traffic police target alone is not perfect, the data set construction cannot be carried out based on a real road traffic scene, and the command gesture of the traffic police cannot be accurately judged by fully applying the traffic police target data set in the prior art by providing the scene construction method and the scene construction system of the traffic police target data set based on multiple angles. The first angle traffic police data set is constructed based on different weather visibility, the second angle traffic police data set is constructed based on different distance, the third angle traffic police data set is constructed based on gender difference, based on which the multi-scene on duty data set of the first user can be constructed, wherein, different weather visibility, different distance, sex difference, etc. are included, and different characteristics are marked, the constructed traffic police target data set is clearly classified, a large amount of data and standard labels are beneficial to improving the detection performance of the algorithm in specific scenes, and the multi-scene duty data set is ensured to cover various conditions of the research target as much as possible, so that the generalization capability of the trained model is stronger, and then more closely to real road traffic scene for the command gesture of traffic police carries out the technological effect of accurate judgement under various traffic scenes.
In one aspect, an embodiment of the present application provides a scenarization construction method for a traffic police target data set based on multiple angles, where the method includes: acquiring a first set of weather condition data for the first area based on the big data; classifying the first set of weather condition data based on a first classification logic, generating a first subset of weather condition data and a second subset of weather condition data; obtaining first weather visibility according to the first weather condition data subset, and obtaining second weather visibility according to the second weather condition data subset, wherein the first weather visibility is greater than a preset visibility threshold value, and the second weather visibility is smaller than the preset visibility threshold value; constructing a first angle traffic police data set based on the first weather visibility and the second weather visibility; acquiring a first user image set in the second weather visibility based on a monitoring camera, wherein the first user is a road on-duty traffic police; based on second classification logic, performing distance classification on the first user image set to generate a second angle traffic police data set; based on a third classification logic, classifying gender differences of the first user image set to generate a third angle traffic police data set; and constructing the multi-scene on-duty data set of the first user according to the first angle traffic police data set, the second angle traffic police data set and the third angle traffic police data set, and carrying out feature labeling.
On the other hand, the application also provides a scenarized construction system of a traffic police target data set based on multiple angles, wherein the system comprises: a first acquisition unit: the first acquisition unit is used for acquiring a first weather condition data set of a first area based on the big data; a first classification unit: the first classification unit is used for classifying the first weather condition data set based on first classification logic to generate a first weather condition data subset and a second weather condition data subset; a first obtaining unit: the first obtaining unit is configured to obtain a first visibility in weather according to the first subset of weather condition data, and obtain a second visibility in weather according to the second subset of weather condition data, where the first visibility in weather is greater than a preset visibility threshold, and the second visibility in weather is less than the preset visibility threshold; a first building unit: the first construction unit is used for constructing a first angle traffic police data set based on the first weather visibility and the second weather visibility; a second obtaining unit: the second obtaining unit is used for obtaining a first user image set in the second weather visibility state based on a monitoring camera, wherein the first user is a road traffic police; a second classification unit: the second classification unit is used for classifying the first user image set according to the distance based on a second classification logic to generate a second angle traffic police data set; a third classification unit: the third classification unit is used for classifying gender differences of the first user image set based on a third classification logic to generate a third angle traffic police data set; a second building element: the second construction unit is used for constructing the multi-scene on-duty data set of the first user according to the first angle traffic police data set, the second angle traffic police data set and the third angle traffic police data set, and performing feature labeling.
One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:
it is known that the first angle traffic police data set is constructed based on different weather visibility, the second angle traffic police data set is constructed based on different distance, the third angle traffic police data set is constructed based on gender difference, based on which the multi-scene duty data set of the first user can be constructed, wherein, different weather visibility, different distance, sex difference, etc. are included, and different characteristics are marked, the constructed traffic police target data set is clearly classified, a large amount of data and standard labels are beneficial to improving the detection performance of the algorithm in specific scenes, and the multi-scene duty data set is ensured to cover various conditions of the research target as much as possible, so that the generalization capability of the trained model is stronger, and then more closely to real road traffic scene for the command gesture of traffic police carries out the technological effect of accurate judgement under various traffic scenes.
The foregoing description is only an overview of the technical solutions of the present application, and the present application can be implemented according to the content of the description in order to make the technical means of the present application more clearly understood, and the following detailed description of the present application is given in order to make the above and other objects, features, and advantages of the present application more clearly understandable.
Drawings
Other features, objects and advantages of the invention will become more apparent upon reading of the detailed description of non-limiting embodiments with reference to the following drawings:
fig. 1 is a schematic flowchart of a method for constructing a scene based on a multi-angle traffic police target data set according to an embodiment of the present application;
fig. 2 is a schematic flowchart of a traversal analysis performed on the second multi-scene on-duty image set by a multi-angle-based scenarized construction method for a traffic police target data set according to an embodiment of the present application;
fig. 3 is a schematic flowchart of incremental learning on the first feature set in a multi-angle-based scenarized construction method for a traffic police target data set according to an embodiment of the present application;
fig. 4 is a schematic flow chart of a method for constructing a weather condition classification coordinate system based on a multi-angle scenarized construction method of a traffic police target data set according to an embodiment of the present application;
fig. 5 is a schematic flowchart of a convolution operation for traversing the first surplus image set according to a multi-angle-based scenarized construction method for a traffic police target data set in the embodiment of the present application;
fig. 6 is a schematic flowchart illustrating a process of enhancing the primary feature factor set in a multi-angle-based scenarized construction method for a traffic police target data set according to an embodiment of the present application;
fig. 7 is a schematic flowchart of a method for constructing a scene based on a multi-angle traffic police target data set according to an embodiment of the present application, for performing quantitative processing on the information index information;
FIG. 8 is a schematic structural diagram of a system for constructing a scene based on a multi-angle traffic police target data set according to an embodiment of the present application;
fig. 9 is a schematic structural diagram of an exemplary electronic device according to an embodiment of the present application.
Detailed Description
The embodiment of the application provides a multi-angle-based scene construction method and system for a traffic police target data set, and solves the technical problems that in the prior art, the construction of a data set aiming at a traffic police target alone is not perfect enough, and the construction cannot be carried out based on a real road traffic scene, so that the command gesture of a traffic police cannot be accurately judged by fully applying the traffic police target data set. The first angle traffic police data set is constructed based on different weather visibility, the second angle traffic police data set is constructed based on different distance, the third angle traffic police data set is constructed based on gender difference, based on which the multi-scene on duty data set of the first user can be constructed, wherein, different weather visibility, different distance, sex difference, etc. are included, and different characteristics are marked, the constructed traffic police target data set is clearly classified, a large amount of data and standard labels are beneficial to improving the detection performance of the algorithm in specific scenes, and the multi-scene duty data set is ensured to cover various conditions of the research target as much as possible, so that the generalization capability of the trained model is stronger, and then more closely to real road traffic scene for the command gesture of traffic police carries out the technological effect of accurate judgement under various traffic scenes.
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are merely some embodiments of the present application and not all embodiments of the present application, and it should be understood that the present application is not limited to the example embodiments described herein.
Summary of the application
The biggest challenge facing current intelligent driving technology is safety. For example, the vehicle may keep a safe distance from the vehicle ahead, may pass through a crossing under traffic light indication, may pass through a zebra crossing, a school, or the like at a reduced speed, and the like. To solve these problems, an intelligent driving system is required to accurately sense various environments. A mature intelligent driving system not only needs to accurately identify fixed signals such as traffic lights, pavement markers and the like, but also needs to detect a traffic police target and can identify the traffic police command action correspondingly. In the prior art, the data set for the traffic police target alone is not well constructed, and the data set cannot be constructed based on a real road traffic scene, so that the traffic police target data set cannot be fully applied to accurately judge the command gesture of the traffic police.
In view of the above technical problems, the technical solution provided by the present application has the following general idea:
the embodiment of the application provides a scene construction method of a traffic police target data set based on multiple angles, wherein the method comprises the following steps: acquiring a first set of weather condition data for the first area based on the big data; classifying the first set of weather condition data based on a first classification logic, generating a first subset of weather condition data and a second subset of weather condition data; obtaining first weather visibility according to the first weather condition data subset, and obtaining second weather visibility according to the second weather condition data subset, wherein the first weather visibility is greater than a preset visibility threshold value, and the second weather visibility is smaller than the preset visibility threshold value; constructing a first angle traffic police data set based on the first weather visibility and the second weather visibility; acquiring a first user image set in the second weather visibility based on a monitoring camera, wherein the first user is a road on-duty traffic police; based on second classification logic, performing distance classification on the first user image set to generate a second angle traffic police data set; based on a third classification logic, classifying gender differences of the first user image set to generate a third angle traffic police data set; and constructing the multi-scene on-duty data set of the first user according to the first angle traffic police data set, the second angle traffic police data set and the third angle traffic police data set, and carrying out feature labeling.
For better understanding of the above technical solutions, the following detailed descriptions will be provided in conjunction with the drawings and the detailed description of the embodiments.
Example one
As shown in fig. 1, an embodiment of the present application provides a method for constructing a scene of a traffic police target data set based on multiple angles, where the method includes:
step S100: acquiring a first set of weather condition data for the first area based on the big data;
in particular, the greatest challenge facing current smart driving technology is safety. For example, the vehicle may keep a safe distance from the vehicle ahead, may pass through a crossing under traffic light indication, may pass through a zebra crossing, a school, or the like at a reduced speed, and the like. To solve these problems, an intelligent driving system is required to accurately sense various environments. A mature intelligent driving system not only needs to accurately identify fixed signals such as traffic lights, pavement markers and the like, but also needs to detect a traffic police target and can identify the traffic police command action correspondingly. In this application embodiment, in order to found the traffic police target data set as comprehensive as possible, be convenient for more accurate make corresponding discernment to traffic police's command action, can cut in from the multiangle, and then found the traffic police target data set of scene ization, first region is the target area in the intelligent driving, the region that the vehicle passed through promptly, first weather condition data set is for passing through big data acquisition, the historical weather conditions of first region, including situations such as sunny day, rainy day, haze day, through gathering the weather conditions, can further found the traffic police target data set of scene ization.
Step S200: classifying the first set of weather condition data based on a first classification logic, generating a first subset of weather condition data and a second subset of weather condition data;
step S300: obtaining first weather visibility according to the first weather condition data subset, and obtaining second weather visibility according to the second weather condition data subset, wherein the first weather visibility is greater than a preset visibility threshold value, and the second weather visibility is smaller than the preset visibility threshold value;
specifically, given the first weather condition data set, in order to construct a traffic police target data set based on the first weather condition data set, the traffic police target data set may be classified according to the first classification logic, that is, classified according to the high or low visibility of the air, which affects the judgment of the traffic police gesture, when the traffic police encounters a fog-haze weather, the visibility of the air is reduced, and the command gesture of the traffic police cannot be further judged, where the first weather condition data subset and the second weather condition data subset are the classification results, where the first visibility of the weather is corresponding to the first weather condition data subset, the second visibility of the weather is corresponding to the second weather condition data subset, the preset visibility threshold is a preset clear weather visibility, and further, the first visibility of the weather is greater than the preset visibility threshold, the first weather condition is better, the command gesture of the traffic police can be clearly judged, the second weather visibility is smaller than the preset visibility threshold value, namely, the second weather condition is poorer, the command gesture of the traffic police cannot be clearly judged, and the scenes of the traffic police target data set can be further enriched by classifying the collected weather conditions.
Step S400: constructing a first angle traffic police data set based on the first weather visibility and the second weather visibility;
specifically, a first angle traffic police data set can be constructed based on the first weather visibility and the second weather visibility, that is, a first angle traffic police data set is constructed according to different weather visibility, and the first angle traffic police data set comprises traffic police target data sets with different weather visibility.
Step S500: acquiring a first user image set in the second weather visibility based on a monitoring camera, wherein the first user is a road on-duty traffic police;
step S600: based on second classification logic, performing distance classification on the first user image set to generate a second angle traffic police data set;
specifically, the common traffic police targets are mostly present at the crossing intersections, so that the traffic guidance pictures of the traffic police can be collected based on the monitoring camera, the first user image set is the image set in the second weather visibility, that is, the traffic guidance pictures of the traffic police are collected when the weather visibility is low, and further, the first user image set is classified in a far and near distance based on the second classification logic, which can be further understood that the second weather visibility is low, so that the command gestures of the traffic police cannot be timely and effectively judged, and the long-distance traffic police image set and the short-distance traffic police image set can be respectively collected to generate a second angle traffic police data set, wherein the second angle traffic police data set comprises traffic police target data sets in different distances.
Step S700: based on a third classification logic, classifying gender differences of the first user image set to generate a third angle traffic police data set;
specifically, in order to make the scene construction of the traffic police target data set richer, the first user image set may be classified according to gender differences based on a third classification logic, traffic guidance may slightly differ due to the gender differences, and the traffic police may be distinguished based on a hat wearing type of a male traffic police and a female traffic police, so as to clearly determine the guidance gesture of the traffic police, where the third angular traffic police data set includes the traffic police target data set under the gender differences.
Step S800: and constructing the multi-scene on-duty data set of the first user according to the first angle traffic police data set, the second angle traffic police data set and the third angle traffic police data set, and carrying out feature labeling.
Specifically, it is known that the first angle traffic police data set is constructed based on different weather visibility, the second angle traffic police data set is constructed based on different distance, the third angle traffic police data set is constructed based on gender difference, based on which the multi-scene duty data set of the first user can be constructed, wherein, different weather visibility, different distance, sex difference, etc. are included, and different characteristics are marked, the constructed traffic police target data set is clearly classified, a large amount of data and standard labels are beneficial to improving the detection performance of the algorithm in specific scenes, and the multi-scene duty data set is ensured to cover various conditions of the research target as much as possible, so that the generalization capability of the trained model is stronger, and then more closely to real road traffic scene for the command gesture of traffic police carries out the technological effect of accurate judgement under various traffic scenes.
Preferably, as shown in fig. 2, the embodiment of the present application further includes:
step S910: obtaining a matched multi-scene on-duty image set according to the multi-scene on-duty data set;
step S920: performing enhancement processing on the multi-scene on-duty image set to generate a second multi-scene on-duty image set of the first user;
step S930: inputting the second multi-scene duty image set into a weather feature labeling model for training to obtain a first labeling result, wherein the first labeling result is used for labeling the weather features in the second multi-scene duty image set;
step S940: inputting the second multi-scene on-duty image set into a distance feature labeling model for training to obtain a second labeling result, wherein the second labeling result is used for labeling distance features in the second multi-scene on-duty image set;
step S950: inputting the second multi-scene duty image set into a gender difference labeling model for training to obtain a third labeling result, wherein the third labeling result is to label gender difference characteristics in the second multi-scene duty image set;
step S960: and traversing and analyzing the second multi-scene on-duty image set according to the first labeling result, the second labeling result and the third labeling result to obtain a first surplus image set, and screening the images.
Specifically, in order to label the features of the images in the multi-scene duty data set, further, a matched multi-scene duty image set may be obtained according to the multi-scene duty data set, in order to label the multi-scene duty image set based on the image information, the multi-scene duty image set may be enhanced, that is, processed in a manner of rotating, blocking, zooming, adding an angle, adding noise, mirroring, etc., the second multi-scene duty image set is an enhanced image set, further, the second multi-scene duty image set is trained by sequentially inputting a weather feature labeling model, a distance feature labeling model, and a gender difference labeling model into the second multi-scene duty image set, the weather feature labeling model may label the weather features in the second multi-scene duty image set, the distance feature labeling model may label the distance features in the second multi-scene duty image set, the gender difference annotation model may annotate gender difference features in the second set of multi-scene duty images, thereby, complete feature labeling can be carried out on any duty image in the second multi-scene duty image set, and then, according to the first labeling result, the second labeling result and the third labeling result, performing traversal analysis on the second multi-scene duty image set to obtain a first surplus image set, wherein the first surplus image set can be understood as a set of other duty images without labeling weather features, distance features and gender difference features, and then the first surplus image set is screened, so that an image set containing additional features is screened out, the accurate characteristic marking of the images in the multi-scene duty data set is realized through a weather characteristic marking model, a distance characteristic marking model and a gender difference marking model.
Preferably, as shown in fig. 3, the embodiment of the present application further includes:
step S961: performing image recognition and information extraction on the first surplus image set to obtain a first recognition result;
step S962: judging whether a first feature set exists in the first surplus image set or not according to the first identification result, wherein the first feature set is different from the first labeling result, the second labeling result and the third labeling result;
step S963: and if the first surplus image set has the first feature set, performing incremental learning on the first feature set to obtain a fourth labeling result.
Specifically, in order to perform image screening on the first surplus image set, further, image recognition and information extraction may be performed on the first surplus image set to obtain a first recognition result, based on an image recognition technology, a computer may be used to process, analyze and understand an image to recognize targets and objects in various different modes, and further effectively extract the targets and objects in the different modes, where the first recognition result is a result after processing, and includes a feature set of each image in the first surplus image set, and further, according to the first recognition result, it is determined whether the first surplus image set has a first feature set, where the first feature set is different from the first labeling result, the second labeling result, and the third labeling result, that is, it is determined whether the weather feature set has a surplus image set or not And for example, the feature sets except the distance feature and the gender difference feature comprise whether pedestrians and other factors appear in the on-duty image, if so, the first feature set is subjected to incremental learning to obtain a fourth labeling result, the fourth labeling result is a result for labeling the first feature set, and the scene construction of the traffic police target data set is ensured to be richer by carrying out image screening and labeling on the first surplus image set.
Preferably, as shown in fig. 4, the classifying the first antenna condition data set based on the first classification logic, and the step S200 further includes:
step S210: constructing a weather condition classification coordinate system, taking air impurity information as a horizontal coordinate and taking illumination intensity information as a vertical coordinate;
step S220: performing regional labeling classification on the weather condition classification coordinate system to obtain a first label classification result;
step S230: inputting the first weather condition data set into the weather condition classification coordinate system to obtain a weather condition classification vector;
step S240: performing distance calculation on the weather condition classification vector to obtain an Euclidean distance data set;
step S250: obtaining a weather condition classification data set according to the Euclidean distance data set, wherein the weather condition classification data set is the shortest k distances in the Euclidean distance data set;
step S260: mapping and matching are carried out according to the weather condition classification data set and the first label classification result, and a first classification result is obtained;
step S270: classifying the first weather condition dataset according to the first classification result.
Specifically, a weather condition classification coordinate system is constructed, wherein the weather condition classification includes conditions such as sunny days, rainy days, haze days and the like, and the weather condition classification coordinate system is constructed by taking air impurity information as an abscissa and taking illumination intensity information as an ordinate. The air impurity information comprises so-called dust, smoke, haze and the like, when impurities in the air are more, the atmospheric visibility is reduced, meanwhile, the illumination intensity information also influences the atmospheric visibility, namely, when the weather condition is cloudy, the atmospheric visibility is also lower due to the fact that the illumination intensity is weaker, and then the weather condition classification coordinate system is subjected to regional labeling classification, and different regions correspond to different label classification results, namely different regions correspond to different classifications of the weather condition. And inputting the first weather condition data set into the weather condition classification coordinate system to obtain a weather condition classification vector, and performing mapping matching according to the weather condition classification data set and the first label classification result to obtain a first classification result.
And the Euclidean distance data set is an Euclidean metric distance data set, namely the straight-line distance between two points in a coordinate system, and the distance calculation is carried out on the weather condition classification vector to obtain the Euclidean distance data set between the vector and other weather condition classifications. The weather condition classification data set is the shortest k distances in the Euclidean distance data set, and the k value is a part of the Euclidean distance data set and can be set by self. And performing mapping matching according to the weather condition classification data set and the first label classification result to obtain a classification label corresponding to the vector, and determining the weather condition classification corresponding to the vector according to the classification result. And classifying the first weather condition data set according to the corresponding weather condition classification, so that a method for performing vector mapping by constructing a weather condition classification coordinate system is achieved, the scheme classification result is more accurate, and the technical effect of more accurate weather condition classification is ensured.
Preferably, as shown in fig. 5, the step S963 of incrementally learning the first feature set further includes:
step S9631: obtaining a target volume-based feature of the first feature set;
step S9632: performing traversal convolution operation on the first surplus image set according to the target volume base characteristic to obtain a first convolution result;
step S9633: and obtaining the fourth labeling result according to the first convolution result, and performing incremental learning.
Specifically, in order to perform incremental learning on the first feature set more scientifically and accurately, the first surplus image set has an overall distribution convolution feature, the first feature set has the target volume-based feature, the target convolution feature is a distribution feature that needs to be obtained through training, then, based on the target volume-based feature, traversal convolution operation is performed on the first surplus image set, the first convolution result is a maximum value in the convolution operation, and then, based on the first convolution result, the fourth labeling result is obtained and is subjected to incremental learning, so that the first feature set is subjected to incremental learning more scientifically and accurately.
Preferably, as shown in fig. 6, the step S920 of performing enhancement processing on the multi-scene duty image set further includes:
step S921: acquiring a main characteristic factor set and a secondary characteristic factor set in each scene on-duty image according to the multi-scene on-duty image set;
step S922: judging whether the secondary characteristic factor set influences the primary characteristic factor set or not;
step S923: and if the secondary characteristic factor set influences the primary characteristic factor set, performing enhancement processing on the primary characteristic factor set, and performing weakening processing on the secondary characteristic factor set.
Specifically, in order to perform enhancement processing on the multi-scene duty image set, further, a main characteristic factor set and a sub-characteristic factor set in each scene duty image may be obtained according to the multi-scene duty image set, where the main characteristic factor set may be an image set of a traffic police target, the sub-characteristic factor set may be an image set of pedestrians on roads, other than the traffic police target, and further, it is determined whether the sub-characteristic factor set affects the main characteristic factor set, that is, it is determined whether pedestrians on roads affect locking and determining of the traffic police target, if so, the main characteristic factor set may be enhanced, and the sub-characteristic factor set is weakened, that is, irrelevant factors other than the traffic police target are weakened, so that locking and determining of the traffic police target are not affected, and meanwhile, the traffic police target image is enhanced, and the enhancement processing of the multi-scene on-duty image set is realized by generating visual contrast.
Preferably, as shown in fig. 7, the step S923 is further configured to perform an enhancement process on the set of main characteristic factors:
step S9231: obtaining information index information of the main characteristic factor set;
step S9232: presetting target image index information of the first user;
step S9233: and carrying out quantitative processing on the information index information according to the target image index information.
Specifically, in order to perform enhancement processing on the main feature factor set, information index information of the main feature factor set may be further obtained, that is, after a traffic police target is locked, the area, the perimeter, the shape, and the like of a traffic police target image may be obtained, and target image index information of the first user may be preset, where the target image index information may be understood as index information that a preset traffic police target image needs to reach, and then, according to the target image index information, the information index information is quantitatively processed, so that the information index information meets the requirement of the target image index information, and further, the enhancement processing on the main feature factor set is realized.
Compared with the prior art, the invention has the following beneficial effects:
1. the first angle traffic police data set is constructed based on different weather visibility, the second angle traffic police data set is constructed based on different distance, the third angle traffic police data set is constructed based on gender difference, based on which the multi-scene on duty data set of the first user can be constructed, wherein, different weather visibility, different distance, sex difference, etc. are included, and different characteristics are marked, the constructed traffic police target data set is clearly classified, a large amount of data and standard labels are beneficial to improving the detection performance of the algorithm in specific scenes, and the multi-scene duty data set is ensured to cover various conditions of the research target as much as possible, so that the generalization capability of the trained model is stronger, and then more closely to real road traffic scene for the command gesture of traffic police carries out the technological effect of accurate judgement under various traffic scenes.
Example two
Based on the same inventive concept as the scenarized construction method of a multi-angle-based traffic police target data set in the foregoing embodiment, the present invention further provides a scenarized construction system of a multi-angle-based traffic police target data set, as shown in fig. 8, the system includes:
the first acquisition unit 11: the first acquisition unit 11 is used for acquiring a first antenna condition data set of a first area based on the big data;
first classification unit 12: the first classification unit 12 is configured to classify the first weather condition data set based on a first classification logic, and generate a first weather condition data subset and a second weather condition data subset;
the first obtaining unit 13: the first obtaining unit 13 is configured to obtain a first visibility in weather according to the first subset of weather condition data, and obtain a second visibility in weather according to the second subset of weather condition data, where the first visibility in weather is greater than a preset visibility threshold, and the second visibility in weather is less than the preset visibility threshold;
the first building element 14: the first construction unit 14 is configured to construct a first angle traffic police data set based on the first weather visibility and the second weather visibility;
the second obtaining unit 15: the second obtaining unit 15 is configured to obtain, based on a monitoring camera, a first user image set in the second weather visibility, where the first user is a road traffic police;
the second classification unit 16: the second classification unit 16 is configured to perform distance classification on the first user image set based on a second classification logic, so as to generate a second angle traffic police data set;
third classification unit 17: the third classification unit 17 is configured to classify the first user image set according to gender differences based on a third classification logic, so as to generate a third angle traffic police data set;
second building element 18: the second constructing unit 18 is configured to construct the multi-scene on-duty data set of the first user according to the first angle traffic police data set, the second angle traffic police data set, and the third angle traffic police data set, and perform feature labeling.
Further, the system further comprises:
a third obtaining unit: the third obtaining unit is used for obtaining a matched multi-scene on-duty image set according to the multi-scene on-duty data set;
a first generation unit: the first generating unit is used for performing enhancement processing on the multi-scene duty image set to generate a second multi-scene duty image set of the first user;
a first input unit: the first input unit is used for inputting the second multi-scene duty image set into a weather feature labeling model for training to obtain a first labeling result, and the first labeling result is used for labeling the weather features in the second multi-scene duty image set;
a second input unit: the second input unit is used for inputting a distance feature labeling model into the second multi-scene on-duty image set for training to obtain a second labeling result, and the second labeling result is used for labeling distance features in the second multi-scene on-duty image set;
a third input unit: the third input unit is used for inputting the second multi-scene on-duty image set into a gender difference labeling model for training to obtain a third labeling result, and the third labeling result is used for labeling gender difference characteristics in the second multi-scene on-duty image set;
a first analysis unit: the first analysis unit is used for performing traversal analysis on the second multi-scene on-duty image set according to the first labeling result, the second labeling result and the third labeling result to obtain a first surplus image set, and performing image screening.
Further, the system further comprises:
a fourth obtaining unit: the fourth obtaining unit is used for carrying out image recognition and information extraction on the first surplus image set to obtain a first recognition result;
a first judgment unit: the first judging unit is configured to judge whether a first feature set exists in the first surplus image set according to the first recognition result, where the first feature set is different from the first labeling result, the second labeling result, and the third labeling result;
a fifth obtaining unit: the fifth obtaining unit is configured to, if the first surplus image set has the first feature set, perform incremental learning on the first feature set to obtain a fourth labeling result.
Further, the system further comprises:
a third building element: the third construction unit is used for constructing a weather condition classification coordinate system, and the air impurity information is used as a horizontal coordinate, and the illumination intensity information is used as a vertical coordinate;
a fourth classification unit: the fourth classification unit is used for performing regional labeling classification on the weather condition classification coordinate system to obtain a first label classification result;
a fourth input unit: the fourth input unit is used for inputting the first weather condition data set into the weather condition classification coordinate system to obtain a weather condition classification vector;
the first calculation unit: the first calculation unit is used for performing distance calculation on the weather condition classification vector to obtain a Euclidean distance data set;
a sixth obtaining unit: the sixth obtaining unit is configured to obtain a weather condition classification dataset according to the euclidean distance dataset, where the weather condition classification dataset is shortest k distances in the euclidean distance dataset;
a first matching unit: the first matching unit is used for carrying out mapping matching according to the weather condition classification data set and the first label classification result to obtain a first classification result;
a fifth classification unit: the fifth classification unit is used for classifying the first weather condition data set according to the first classification result.
Further, the system further comprises:
a seventh obtaining unit: the seventh obtaining unit is configured to obtain a target volume base characteristic of the first characteristic set;
a first arithmetic unit: the first operation unit is used for performing traversal convolution operation on the first surplus image set according to the target volume base characteristic to obtain a first convolution result;
an eighth obtaining unit: the eighth obtaining unit is configured to obtain the fourth labeling result according to the first convolution result, and perform incremental learning.
Further, the system further comprises:
a ninth obtaining unit: the ninth obtaining unit is configured to obtain a main feature factor set and a sub feature factor set in each scene duty image according to the multi-scene duty image set;
a second judgment unit: the second judging unit is configured to judge whether the set of sub-feature factors affects the set of main feature factors;
a first processing unit: the first processing unit is used for performing enhancement processing on the primary characteristic factor set and performing weakening processing on the secondary characteristic factor set if the secondary characteristic factor set affects the primary characteristic factor set.
Further, the system further comprises:
a tenth obtaining unit: the tenth obtaining unit is configured to obtain information index information of the main feature factor set;
a first preset unit: the first preset unit is used for presetting target image index information of the first user;
a second processing unit: the second processing unit is used for carrying out quantitative processing on the information index information according to the target image index information.
Various changes and specific examples of the method for constructing a multi-angle-based traffic police target data set in the first embodiment of fig. 1 are also applicable to the system for constructing a multi-angle-based traffic police target data set in the present embodiment, and through the foregoing detailed description of the method for constructing a multi-angle-based traffic police target data set, those skilled in the art can clearly know the method for constructing a multi-angle-based traffic police target data set in the present embodiment, so that for the brevity of the description, detailed description is not repeated again.
EXAMPLE III
The electronic apparatus of the embodiment of the present application is described below with reference to fig. 9.
Fig. 9 illustrates a schematic structural diagram of an electronic device according to an embodiment of the present application.
Based on the inventive concept of the multi-angle-based scenarized construction method for the traffic police target data set in the embodiment, the invention further provides a multi-angle-based scenarized construction system for the traffic police target data set, on which a computer program is stored, which when executed by a processor implements the steps of any one of the methods of the multi-angle-based scenarized construction system for the traffic police target data set.
Where in fig. 9 a bus architecture (represented by bus 300), bus 300 may include any number of interconnected buses and bridges, bus 300 linking together various circuits including one or more processors, represented by processor 302, and memory, represented by memory 304. The bus 300 may also link together various other circuits such as peripherals, voltage regulators, power management circuits, and the like, which are well known in the art, and therefore, will not be described any further herein. A bus interface 305 provides an interface between the bus 300 and the receiver 301 and transmitter 303. The receiver 301 and the transmitter 303 may be the same element, i.e., a transceiver, providing a means for communicating with various other systems over a transmission medium. The processor 302 is responsible for managing the bus 300 and general processing, and the memory 304 may be used for storing data used by the processor 302 in performing operations.
The embodiment of the application provides a scene construction method of a traffic police target data set based on multiple angles, wherein the method comprises the following steps: acquiring a first set of weather condition data for the first area based on the big data; classifying the first set of weather condition data based on a first classification logic, generating a first subset of weather condition data and a second subset of weather condition data; obtaining first weather visibility according to the first weather condition data subset, and obtaining second weather visibility according to the second weather condition data subset, wherein the first weather visibility is greater than a preset visibility threshold value, and the second weather visibility is smaller than the preset visibility threshold value; constructing a first angle traffic police data set based on the first weather visibility and the second weather visibility; acquiring a first user image set in the second weather visibility based on a monitoring camera, wherein the first user is a road on-duty traffic police; based on second classification logic, performing distance classification on the first user image set to generate a second angle traffic police data set; based on a third classification logic, classifying gender differences of the first user image set to generate a third angle traffic police data set; and constructing the multi-scene on-duty data set of the first user according to the first angle traffic police data set, the second angle traffic police data set and the third angle traffic police data set, and carrying out feature labeling.
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 a system for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction system which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps 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 changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations 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 include such modifications and variations.
Claims (9)
1. A scenarized construction method of a multi-angle-based traffic police target data set is disclosed, wherein the method comprises the following steps:
acquiring a first set of weather condition data for the first area based on the big data;
classifying the first set of weather condition data based on a first classification logic, generating a first subset of weather condition data and a second subset of weather condition data;
obtaining first weather visibility according to the first weather condition data subset, and obtaining second weather visibility according to the second weather condition data subset, wherein the first weather visibility is greater than a preset visibility threshold value, and the second weather visibility is smaller than the preset visibility threshold value;
constructing a first angle traffic police data set based on the first weather visibility and the second weather visibility;
acquiring a first user image set in the second weather visibility based on a monitoring camera, wherein the first user is a road on-duty traffic police;
based on second classification logic, performing distance classification on the first user image set to generate a second angle traffic police data set;
based on a third classification logic, classifying gender differences of the first user image set to generate a third angle traffic police data set;
and constructing the multi-scene on-duty data set of the first user according to the first angle traffic police data set, the second angle traffic police data set and the third angle traffic police data set, and carrying out feature labeling.
2. The method of claim 1, wherein the method further comprises:
obtaining a matched multi-scene on-duty image set according to the multi-scene on-duty data set;
performing enhancement processing on the multi-scene on-duty image set to generate a second multi-scene on-duty image set of the first user;
inputting the second multi-scene duty image set into a weather feature labeling model for training to obtain a first labeling result, wherein the first labeling result is used for labeling the weather features in the second multi-scene duty image set;
inputting the second multi-scene on-duty image set into a distance feature labeling model for training to obtain a second labeling result, wherein the second labeling result is used for labeling distance features in the second multi-scene on-duty image set;
inputting the second multi-scene duty image set into a gender difference labeling model for training to obtain a third labeling result, wherein the third labeling result is to label gender difference characteristics in the second multi-scene duty image set;
and traversing and analyzing the second multi-scene on-duty image set according to the first labeling result, the second labeling result and the third labeling result to obtain a first surplus image set, and screening the images.
3. The method of claim 2, wherein the method further comprises:
performing image recognition and information extraction on the first surplus image set to obtain a first recognition result;
judging whether a first feature set exists in the first surplus image set or not according to the first identification result, wherein the first feature set is different from the first labeling result, the second labeling result and the third labeling result;
and if the first surplus image set has the first feature set, performing incremental learning on the first feature set to obtain a fourth labeling result.
4. The method of claim 1, wherein said classifying the first set of antenna condition data based on a first classification logic further comprises:
constructing a weather condition classification coordinate system, taking air impurity information as a horizontal coordinate and taking illumination intensity information as a vertical coordinate;
performing regional labeling classification on the weather condition classification coordinate system to obtain a first label classification result;
inputting the first weather condition data set into the weather condition classification coordinate system to obtain a weather condition classification vector;
performing distance calculation on the weather condition classification vector to obtain an Euclidean distance data set;
obtaining a weather condition classification data set according to the Euclidean distance data set, wherein the weather condition classification data set is the shortest k distances in the Euclidean distance data set;
mapping and matching are carried out according to the weather condition classification data set and the first label classification result, and a first classification result is obtained;
classifying the first weather condition dataset according to the first classification result.
5. The method of claim 3, wherein the incrementally learning the first set of features further comprises:
obtaining a target volume-based feature of the first feature set;
performing traversal convolution operation on the first surplus image set according to the target volume base characteristic to obtain a first convolution result;
and obtaining the fourth labeling result according to the first convolution result, and performing incremental learning.
6. The method of claim 2, wherein said enhancing said multi-scene duty image set further comprises:
acquiring a main characteristic factor set and a secondary characteristic factor set in each scene on-duty image according to the multi-scene on-duty image set;
judging whether the secondary characteristic factor set influences the primary characteristic factor set or not;
and if the secondary characteristic factor set influences the primary characteristic factor set, performing enhancement processing on the primary characteristic factor set, and performing weakening processing on the secondary characteristic factor set.
7. The method of claim 6, wherein the enhancing the set of dominant eigen factors further comprises:
obtaining information index information of the main characteristic factor set;
presetting target image index information of the first user;
and carrying out quantitative processing on the information index information according to the target image index information.
8. A system for scenarized construction of a multi-angle based traffic police target data set, wherein the system comprises:
a first acquisition unit: the first acquisition unit is used for acquiring a first weather condition data set of a first area based on the big data;
a first classification unit: the first classification unit is used for classifying the first weather condition data set based on first classification logic to generate a first weather condition data subset and a second weather condition data subset;
a first obtaining unit: the first obtaining unit is configured to obtain a first visibility in weather according to the first subset of weather condition data, and obtain a second visibility in weather according to the second subset of weather condition data, where the first visibility in weather is greater than a preset visibility threshold, and the second visibility in weather is less than the preset visibility threshold;
a first building unit: the first construction unit is used for constructing a first angle traffic police data set based on the first weather visibility and the second weather visibility;
a second obtaining unit: the second obtaining unit is used for obtaining a first user image set in the second weather visibility state based on a monitoring camera, wherein the first user is a road traffic police;
a second classification unit: the second classification unit is used for classifying the first user image set according to the distance based on a second classification logic to generate a second angle traffic police data set;
a third classification unit: the third classification unit is used for classifying gender differences of the first user image set based on a third classification logic to generate a third angle traffic police data set;
a second building element: the second construction unit is used for constructing the multi-scene on-duty data set of the first user according to the first angle traffic police data set, the second angle traffic police data set and the third angle traffic police data set, and performing feature labeling.
9. A system for scenarized construction of a multi-angle based traffic police target data set comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the method according to any one of claims 1 to 7 when executing the program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110749503.9A CN113392804B (en) | 2021-07-02 | 2021-07-02 | Multi-angle-based traffic police target data set scene construction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110749503.9A CN113392804B (en) | 2021-07-02 | 2021-07-02 | Multi-angle-based traffic police target data set scene construction method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113392804A true CN113392804A (en) | 2021-09-14 |
| CN113392804B CN113392804B (en) | 2022-08-16 |
Family
ID=77625030
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110749503.9A Active CN113392804B (en) | 2021-07-02 | 2021-07-02 | Multi-angle-based traffic police target data set scene construction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113392804B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114648672A (en) * | 2022-02-25 | 2022-06-21 | 北京百度网讯科技有限公司 | Method and device for constructing sample image set, electronic equipment and readable storage medium |
Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108830188A (en) * | 2018-05-30 | 2018-11-16 | 西安理工大学 | Vehicle checking method based on deep learning |
| CN108932503A (en) * | 2018-07-20 | 2018-12-04 | 上海小蚁科技有限公司 | The recognition methods of Chinese herbaceous peony obstacle and device, storage medium, terminal under bad weather |
| CN109446371A (en) * | 2018-11-09 | 2019-03-08 | 苏州清研精准汽车科技有限公司 | A kind of intelligent automobile emulation testing scene library generating method and test macro and method |
| CN110136079A (en) * | 2019-05-05 | 2019-08-16 | 长安大学 | Image defogging method based on scene depth segmentation |
| CN110210354A (en) * | 2019-05-23 | 2019-09-06 | 南京邮电大学 | A kind of detection of haze weather traffic mark with know method for distinguishing |
| CN110399688A (en) * | 2019-07-30 | 2019-11-01 | 奇瑞汽车股份有限公司 | The environmental working condition of automatic Pilot determines method, apparatus and storage medium |
| CN110494715A (en) * | 2017-02-06 | 2019-11-22 | 瓦亚视觉感知有限公司 | Computer aided pilot |
| CN110753953A (en) * | 2017-06-06 | 2020-02-04 | 智加科技公司 | Method and system for object-centric stereo vision in autonomous vehicles via cross-modality verification |
| US20200086879A1 (en) * | 2018-09-14 | 2020-03-19 | Honda Motor Co., Ltd. | Scene classification prediction |
| CN111125442A (en) * | 2019-12-11 | 2020-05-08 | 苏州智加科技有限公司 | Data labeling method and device |
| CN111340151A (en) * | 2020-05-25 | 2020-06-26 | 南京信息工程大学 | Weather phenomenon recognition system and method for assisting automatic driving of vehicle |
| CN112074790A (en) * | 2018-03-18 | 2020-12-11 | 德利沃科技有限公司 | Apparatus, system and method for autonomous driving and remotely operated vehicle |
| CN112395951A (en) * | 2020-10-23 | 2021-02-23 | 中国地质大学（武汉） | Complex scene-oriented domain-adaptive traffic target detection and identification method |
| CN112434586A (en) * | 2020-11-16 | 2021-03-02 | 中山大学 | Multi-complex scene target detection method based on domain adaptive learning |
| CN112560925A (en) * | 2020-12-10 | 2021-03-26 | 中国科学院深圳先进技术研究院 | Complex scene target detection data set construction method and system |
| CN112581379A (en) * | 2019-09-30 | 2021-03-30 | 华为技术有限公司 | Image enhancement method and device |
| CN112818910A (en) * | 2021-02-23 | 2021-05-18 | 腾讯科技（深圳）有限公司 | Vehicle gear control method and device, computer equipment and storage medium |
-
2021
- 2021-07-02 CN CN202110749503.9A patent/CN113392804B/en active Active
Patent Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110494715A (en) * | 2017-02-06 | 2019-11-22 | 瓦亚视觉感知有限公司 | Computer aided pilot |
| CN110753953A (en) * | 2017-06-06 | 2020-02-04 | 智加科技公司 | Method and system for object-centric stereo vision in autonomous vehicles via cross-modality verification |
| CN112074790A (en) * | 2018-03-18 | 2020-12-11 | 德利沃科技有限公司 | Apparatus, system and method for autonomous driving and remotely operated vehicle |
| CN108830188A (en) * | 2018-05-30 | 2018-11-16 | 西安理工大学 | Vehicle checking method based on deep learning |
| CN108932503A (en) * | 2018-07-20 | 2018-12-04 | 上海小蚁科技有限公司 | The recognition methods of Chinese herbaceous peony obstacle and device, storage medium, terminal under bad weather |
| US20200086879A1 (en) * | 2018-09-14 | 2020-03-19 | Honda Motor Co., Ltd. | Scene classification prediction |
| CN109446371A (en) * | 2018-11-09 | 2019-03-08 | 苏州清研精准汽车科技有限公司 | A kind of intelligent automobile emulation testing scene library generating method and test macro and method |
| CN110136079A (en) * | 2019-05-05 | 2019-08-16 | 长安大学 | Image defogging method based on scene depth segmentation |
| CN110210354A (en) * | 2019-05-23 | 2019-09-06 | 南京邮电大学 | A kind of detection of haze weather traffic mark with know method for distinguishing |
| CN110399688A (en) * | 2019-07-30 | 2019-11-01 | 奇瑞汽车股份有限公司 | The environmental working condition of automatic Pilot determines method, apparatus and storage medium |
| CN112581379A (en) * | 2019-09-30 | 2021-03-30 | 华为技术有限公司 | Image enhancement method and device |
| CN111125442A (en) * | 2019-12-11 | 2020-05-08 | 苏州智加科技有限公司 | Data labeling method and device |
| CN111340151A (en) * | 2020-05-25 | 2020-06-26 | 南京信息工程大学 | Weather phenomenon recognition system and method for assisting automatic driving of vehicle |
| CN112395951A (en) * | 2020-10-23 | 2021-02-23 | 中国地质大学（武汉） | Complex scene-oriented domain-adaptive traffic target detection and identification method |
| CN112434586A (en) * | 2020-11-16 | 2021-03-02 | 中山大学 | Multi-complex scene target detection method based on domain adaptive learning |
| CN112560925A (en) * | 2020-12-10 | 2021-03-26 | 中国科学院深圳先进技术研究院 | Complex scene target detection data set construction method and system |
| CN112818910A (en) * | 2021-02-23 | 2021-05-18 | 腾讯科技（深圳）有限公司 | Vehicle gear control method and device, computer equipment and storage medium |
Non-Patent Citations (5)
| Title |
|---|
| GUANGNAN ZHANG 等: "Risk factors associated with traffic violations and accident severity in China", 《ACCIDENT ANALYSIS & PREVENTION》 * |
| WEIHUA ZHANG 等: "Investigating factors influence drivers" speed selection behavior under reduce visibility conditions", 《TRAFFIC INJURY PREVENTION》 * |
| 刘志涛 等: "基于YOLO的雾天行人实时检测方法", 《软件》 * |
| 田竞: "驾驶风险信息分析及交通冲突预警研究", 《中国优秀博硕士学位论文全文数据库（硕士）》 * |
| 郭璠: "图像去雾方法和评价及其应用研究", 《中国优秀博硕士学位论文全文数据库（博士） 信息科技辑》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114648672A (en) * | 2022-02-25 | 2022-06-21 | 北京百度网讯科技有限公司 | Method and device for constructing sample image set, electronic equipment and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113392804B (en) | 2022-08-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Jensen et al. | Vision for looking at traffic lights: Issues, survey, and perspectives | |
| CN110069986B (en) | Traffic signal lamp identification method and system based on hybrid model | |
| CN103034836B (en) | Road sign detection method and road sign checkout equipment | |
| Romdhane et al. | An improved traffic signs recognition and tracking method for driver assistance system | |
| KR102177880B1 (en) | Class labeling apparatus for autonomous driving | |
| JP2021530048A (en) | Multi-layered target classification method and device, traffic sign detection method and device, device and medium | |
| CN111274926B (en) | Image data screening method, device, computer equipment and storage medium | |
| CN104050447A (en) | Traffic light identification method and device | |
| CN106951898B (en) | Vehicle candidate area recommendation method and system and electronic equipment | |
| CN110909656B (en) | Pedestrian detection method and system integrating radar and camera | |
| Bi et al. | A new method of target detection based on autonomous radar and camera data fusion | |
| Sallah et al. | Road sign detection and recognition system for real-time embedded applications | |
| Chen et al. | A computer vision algorithm for locating and recognizing traffic signal control light status and countdown time | |
| Chen et al. | Vehicle detection based on visual attention mechanism and adaboost cascade classifier in intelligent transportation systems | |
| Mijić et al. | Traffic sign detection using YOLOv3 | |
| CN113392804B (en) | Multi-angle-based traffic police target data set scene construction method and system | |
| Shang et al. | A novel method for vehicle headlights detection using salient region segmentation and PHOG feature | |
| Aldoski et al. | Impact of traffic sign diversity on autonomous vehicles: a literature review | |
| Ali et al. | IRUVD: a new still-image based dataset for automatic vehicle detection | |
| Coronado et al. | Detection and classification of road signs for automatic inventory systems using computer vision | |
| Liu et al. | Real-time traffic light recognition based on smartphone platforms | |
| Barshooi et al. | Nighttime Driver Behavior Prediction Using Taillight Signal Recognition via CNN-SVM Classifier | |
| Yu et al. | Research on detection and recognition algorithm of road traffic signs | |
| Ding et al. | A comprehensive approach for road marking detection and recognition | |
| CN115187886A (en) | Vehicle violation detection method and device and electronic equipment |
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 |