CN108764042A - A kind of exception traffic information recognition methods, device and terminal device - Google Patents
A kind of exception traffic information recognition methods, device and terminal device Download PDFInfo
- Publication number
- CN108764042A CN108764042A CN201810381142.5A CN201810381142A CN108764042A CN 108764042 A CN108764042 A CN 108764042A CN 201810381142 A CN201810381142 A CN 201810381142A CN 108764042 A CN108764042 A CN 108764042A
- Authority
- CN
- China
- Prior art keywords
- information
- traffic information
- recognition result
- harvester
- video
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/27—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique
- G10L25/30—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique using neural networks
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
- G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
Abstract
The present invention is suitable for car networking technology field, provides a kind of abnormal traffic information recognition methods, device and terminal device, the method includes:Obtain the traffic information that at least one harvester is sent;Traffic information is inputted into deep learning model, obtains the preliminary recognition result of traffic information;Control signal is sent to the first harvester according to preset rules and preliminary recognition result, the control signal is used to indicate the reference information of the first harvester acquisition road conditions;First harvester is the corresponding harvester of preliminary recognition result;Final anomalous identification result is determined according to reference information and preliminary recognition result.The present invention realizes the comprehensive identification to traffic information, the recognition effect to traffic information is improved, to further enhance road safety management by making full use of traffic information.
Description
Technical field
The invention belongs to a kind of car networking technology field more particularly to abnormal traffic information recognition methods, device and terminals
Equipment.
Background technology
Car networking refers to GPS (Global Positioning System, the global positioning system by being integrated on automobile
System) positioning, the electronic building bricks such as sensor, camera and image procossing, acquisition of the completion vehicle to itself environment and status information,
All vehicles carry out information exchange according to the communication protocol and Data Exchange Standard of agreement.Wherein, V2X (vehicle to X, vehicle
To the information exchange of arbitrary objects) refer to vehicle to extraneous information exchange, it is the key technology of the following intelligent transport system.It
So that vehicle can be communicated with vehicle, vehicle between base station, base station and base station, have safely for car networking, automatic Pilot particularly significant
Effect.
It is identified or the detection to vehicle flowrate currently, the application of car networking is generally only road pavement breakage, but on
It is relatively low to the utilization rate of road information to state application process, it is bad to the recognition effect of target object.
Invention content
In view of this, an embodiment of the present invention provides a kind of abnormal traffic information recognition methods, device and terminal device, with
It is low to solve road information utilization rate in the prior art, the bad problem of the recognition effect of target object.
The first aspect of the embodiment of the present invention provides a kind of abnormal traffic information recognition methods, including:
Obtain the traffic information that at least one harvester is sent;
Traffic information is inputted into deep learning model, obtains the preliminary recognition result of traffic information;
Control signal is sent to the first harvester according to preset rules and preliminary recognition result, control signal is used to indicate
First harvester acquires the reference information of road conditions;First harvester is the corresponding harvester of preliminary recognition result;
Final anomalous identification result is determined according to reference information and preliminary recognition result.
The second aspect of the embodiment of the present invention provides a kind of abnormal traffic information identification device, including:
Traffic information acquisition module, the traffic information sent for obtaining at least one harvester;
Preliminary recognition result acquisition module obtains the first of traffic information for traffic information to be inputted deep learning model
Walk recognition result;
Signal transmitting module is controlled, is controlled for being sent to the first harvester according to preset rules and preliminary recognition result
Signal, control signal are used to indicate the reference information of the first harvester acquisition road conditions;First harvester is that preliminary identification is tied
The corresponding harvester of fruit;
Final anomalous identification result acquisition module, for determining that final exception is known according to reference information and preliminary recognition result
Other result.
The third aspect of the embodiment of the present invention provides a kind of terminal device, including memory, processor and is stored in
In memory and the computer program that can run on a processor, processor are realized as described above different when executing computer program
The step of normal traffic information recognition methods.
The fourth aspect of the embodiment of the present invention provides a kind of computer readable storage medium, computer readable storage medium
It is stored with computer program, abnormal traffic information recognition methods as described above is realized when computer program is executed by processor
Step.
Existing advantageous effect is the embodiment of the present invention compared with prior art:The embodiment of the present invention obtains at least one adopt
The traffic information that acquisition means are sent;Traffic information is inputted into deep learning model, obtains the preliminary recognition result of traffic information;Root
Control signal is sent to the first harvester according to preset rules and preliminary recognition result, control signal is used to indicate the first acquisition dress
Set the reference information of acquisition road conditions;First harvester is the corresponding harvester of preliminary recognition result;According to reference information and
Preliminary recognition result determines final anomalous identification result.The embodiment of the present invention is realized and is satisfied the need by making full use of traffic information
The comprehensive identification of condition information, improves the recognition effect to traffic information, to further enhance road safety management.
Description of the drawings
It to describe the technical solutions in the embodiments of the present invention more clearly, below will be to embodiment or description of the prior art
Needed in attached drawing be briefly described, it should be apparent that, the accompanying drawings in the following description be only the present invention some
Embodiment for those of ordinary skill in the art without having to pay creative labor, can also be according to these
Attached drawing obtains other attached drawings.
Fig. 1 is the implementation process schematic diagram of abnormal traffic information recognition methods provided in an embodiment of the present invention;
Fig. 2 is the specific implementation flow chart of step S102 in Fig. 1 provided in an embodiment of the present invention;
Fig. 3 is the specific implementation flow chart of step S102 in Fig. 1 provided in an embodiment of the present invention;
Fig. 4 is the specific implementation flow chart of step S102 in Fig. 1 provided in an embodiment of the present invention;
Fig. 5 is the structural schematic diagram of abnormal traffic information identification device provided in an embodiment of the present invention;
Fig. 6 is the schematic diagram of terminal device provided in an embodiment of the present invention.
Specific implementation mode
In being described below, for illustration and not for limitation, it is proposed that such as tool of particular system structure, technology etc
Body details, to understand thoroughly the embodiment of the present invention.However, it will be clear to one skilled in the art that there is no these specific
The present invention can also be realized in the other embodiments of details.In other situations, it omits to well-known system, device, electricity
The detailed description of road and method, in case unnecessary details interferes description of the invention.
Term " comprising " in description and claims of this specification and above-mentioned attached drawing and their any deformations, meaning
Figure, which is to cover, non-exclusive includes.Such as process, method or system comprising series of steps or unit, product or equipment do not have
It is defined in the step of having listed or unit, but further includes the steps that optionally not listing or unit, or optionally also wrap
It includes for the intrinsic other steps of these processes, method, product or equipment or unit.In addition, term " first ", " second " and
" third " etc. is for distinguishing different objects, not for description particular order.
In order to illustrate technical scheme of the present invention, illustrated below by specific embodiment.
Embodiment 1:
Fig. 1 shows a kind of implementation process for abnormal traffic information recognition methods that one embodiment of the present of invention provides,
The flow executive agent of the present embodiment can be terminal device, and details are as follows for process:
In S101, the traffic information that at least one harvester is sent is obtained.
In the present embodiment, the overall architecture of abnormal traffic information recognition methods may include sensor layer, edge calculations
Layer and big data layer.
In the present embodiment, sensor layer includes a variety of harvesters.
In the present embodiment, in traffic route both sides multiple harvesters can be disposed, harvester is acquiring road conditions
Information.Traffic information be people on traffic route, vehicle and environment information.
In S102, traffic information is inputted into deep learning model, obtains the preliminary recognition result of traffic information.
In the present embodiment, in edge calculations layer, the traffic information that harvester acquires is input to deep learning model,
Obtain preliminary recognition result.
In the present embodiment, edge calculations layer includes multiple edge calculations gateways, and edge calculations gateway is adopted according to neighbouring
The traffic information of acquisition means acquisition obtains preliminary recognition result,.
In the present embodiment, using terminal device as flow main body, traffic information is inputted into deep learning model, deep learning
Model obtains preliminary recognition result to identify the feature of traffic information according to feature.Preliminary recognition result is to preliminary
Determine the recognition result of traffic information, preliminary recognition result can include but is not limited to track of vehicle information, ponding accumulated snow information,
Audio exception information, traffic accident relevant information, road surface breakage information or people/vehicle identification information it is one or more.
In the present embodiment, deep learning model is neural network model, and deep learning model can be to be learned based on offline
Practise the neural network model of training, or traffic information is inputted depth by the neural network model of unsupervised self study
It practises model to be identified, obtains preliminary recognition result.
In S103, control signal, control letter are sent to the first harvester according to preset rules and preliminary recognition result
Number it is used to indicate the reference information of the first harvester acquisition road conditions;First harvester is the corresponding acquisition of preliminary recognition result
Device.
In the present embodiment, the first harvester is determined by preliminary recognition result and preset rules, and control signal can wrap
The number information and work state information of the first harvester are included, terminal device is filled according to number information to corresponding first acquisition
Transmission control signal is set, work state information is used to indicate the first harvester adjustment working condition.First harvester passes through
Adjust working condition and further acquire the reference informations of road conditions, work state information may include harvester acquisition angles,
The information such as operating power and collection period.
In the present embodiment, preliminary recognition result can also include the more specific location information of traffic information, and terminal device can
To determine the number information and work state information of the first harvester according to more specific location information, and send a control signal to
One harvester.
In S104, final anomalous identification result is determined according to reference information and preliminary recognition result.
In the present embodiment, final anomalous identification result can be obtained according to reference information and preliminary recognition result.
In the present embodiment, reference information recognition result can be obtained according to reference information, identified further according to reference information
As a result with preliminary recognition result, final anomalous identification result is obtained.
In the present embodiment, in big data layer, the local identification that can also be obtained according to multiple edge calculations gateways is tied
Fruit carry out comprehensive descision, obtain final anomalous identification as a result, it is described local recognition result can be preliminary recognition result, can also
The local identification knot that local comprehensive identification obtains is carried out according to reference information and preliminary recognition result for an edge calculations gateway
Fruit.
By taking a specific application scenarios as an example, when the local recognition result obtained in an edge calculations gateway is certain
Ponding is 0.3 meter at the first of road, and the local recognition result obtained in another edge calculations gateway is this road
Ponding is 0.5 meter at second, then obtains final anomalous identification knot after can carrying out comprehensive descision by the local recognition result at two
Fruit is:Section at the first of this road and between second pays attention to evading there are ponding.
From above-described embodiment it is found that the traffic information that the embodiment of the present invention is sent by obtaining at least one harvester;
Traffic information is inputted into deep learning model, obtains the preliminary recognition result of traffic information;According to preset rules and preliminary identification
As a result control signal is sent to the first harvester, control signal is used to indicate the reference letter of the first harvester acquisition road conditions
Breath;First harvester is the corresponding harvester of preliminary recognition result;It is determined most according to reference information and preliminary recognition result
Whole anomalous identification result.The embodiment of the present invention realizes the comprehensive identification to traffic information, changes by making full use of traffic information
The recognition effect to traffic information has been apt to it, to further enhance road safety management.
As shown in Fig. 2, in one embodiment of the invention, Fig. 2 shows the specific implementation streams of step S102 in Fig. 1
Journey, details are as follows for process:
In S201, by video information input first nerves network model, detect video information in target object, and from
The structured features of target object are extracted in video information.
In the present embodiment, harvester includes video acquisition device, and traffic information includes video information, deep learning mould
Type includes first nerves network model and nervus opticus network model, and preliminary recognition result includes target object recognition result.
In the present embodiment, harvester includes video acquisition device, and video acquisition device can be camera herein, adopt
Acquisition means are for acquiring video information, and video information includes video image, after terminal device obtains video information, by video information
It is input to deep learning network.
In the present embodiment, deep learning network includes first nerves network model and nervus opticus network model, wherein
First nerves network model can be convolutional neural networks model, and first nerves network model is used to detect the mesh in video image
Object is marked, target object can be people or vehicle herein.After determining target object, identify target object structured features and
Unstructured feature, structured features be can identified feature, it is unstructured characterized by the feature that can not be described.
In S202, the structured features of target object are inputted into nervus opticus network model, obtain target object identification
As a result.
In the present embodiment, the structured features of target object are input to nervus opticus network model, herein the second god
Can be convolutional neural networks model through network model.
By taking a specific application scenarios as an example, the collected multiple video informations of multiple cameras are inputted into first nerves
In network model, first nerves network model detects the target object in video information.
In the present embodiment, target object can be personage, and object is further detected in nervus opticus network model
The structured features of body, herein structured features can include but is not limited to face, gait, hair style and clothes color of personage etc.
Feature;Target object can also be vehicle, and the structured features of vehicle can include but is not limited to license plate number, vehicle color and vehicle
Shape.
In the present embodiment, the structured features of target object can also include the location information of target object, position letter
Breath may include the target angle and target range of target object, and target angle is target object and corresponding video acquisition device
Angle between the straight line of formation and default zero degree line, target range is between target object and corresponding video acquisition device
Air line distance.
In the present embodiment, since there are multiple cameras while collected multiple video informations, and target object is simultaneously
It is not static, dynamic change, so a target object can obtain multigroup structured features, by above-mentioned multigroup structure
The structured features for changing feature as target object input nervus opticus network model.
In the present embodiment, the structured features for the target object that first nerves network model identifies are input to the second god
Through network model, nervus opticus network model is compared according to the structured features and Pre-existing structure information of target object
Identification, obtains target object recognition result, target object recognition result may include target object identity information and location information.
In one embodiment of the invention, after obtaining target object recognition result, target object can be identified and is tied
Fruit is sent to server, and server preserves target object recognition result, and feature recognition institute is carried out as nervus opticus network model
Pre-existing structure information.
In one embodiment, after obtaining target object recognition result, can also include:
1) control signal is sent to the first harvester according to target object recognition result and preset rules, control signal is used
Reference information is acquired in the first harvester of instruction;
2) according to reference information and target object recognition result, final anomalous identification result is obtained.
It in the present embodiment,, can be according to mesh in order to further determine recognition result after obtaining target object recognition result
The location information in object identification result is marked, control command is sent to the nearest harvester of distance objective object, makes the acquisition
Device adjusts working condition to obtain the reference information of the traffic information according to control command.
In the present embodiment, reference information can be inputted deep learning network model, obtains the identification knot of reference information
Fruit obtains final anomalous identification result according to the recognition result of reference information and target object recognition result.
From above-described embodiment it is found that the video information that different cameras are obtained inputs deep learning network model, to
Multigroup structured features of a target object are obtained, the accuracy of target object recognition result can be improved, improve target
The recognition effect of object.
As shown in figure 3, in one embodiment of the invention, Fig. 3 shows that the another of step S102 realizes stream in Fig. 1
Journey, details are as follows for process:
In S301, video information is inputted into third nerve network model, detection obtains the anomalous event in video information
Structured features and location information.
In the present embodiment, harvester includes video acquisition device, and traffic information includes video information, deep learning mould
Type includes third nerve network model and fourth nerve network model, and preliminary recognition result includes video anomalous identification result.
In the present embodiment, harvester obtains video information, and video information is inputted third nerve network model, the
Three neural network models can be convolutional neural networks model, and third nerve network model can be with first nerves network model phase
Together, or other neural network models do not limit herein.Third nerve network model is for identification in video information
The structured features and location information of anomalous event, structured features can include but is not limited to track of vehicle feature and traffic accident is special
Sign;Location information is the specific location that anomalous event occurs, and location information may include the video source angle of anomalous event and regard
Frequency source distance.Wherein, video source angle is the straight line and default zero degree line that anomalous event and corresponding video acquisition device are formed
Between angle, air line distance of the video source distance between anomalous event and corresponding video acquisition device.
In S302, the structured features of anomalous event and location information are inputted into fourth nerve network model, depending on
Frequency anomalous identification result.
In the present embodiment, fourth nerve network model can be two-way length Recognition with Recurrent Neural Network model in short-term, the 4th god
The traffic accident model trained based on off-line learning, the vehicle rail trained based on off-line learning are can include but is not limited to through network model
One in mark Exception Model, the traffic accident detection model based on self study and the track of vehicle abnormality detection model based on self study
Or it is multiple.The structured features of anomalous event and location information are inputted into fourth nerve network model, fourth nerve network model
It identifies that above-mentioned anomalous event is track of vehicle anomalous event, traffic accident anomalous event or other anomalous events, determines anomalous event
Precise information, obtain video anomalous identification as a result, video anomalous identification result includes the type and anomalous event of anomalous event
Location information.
From above-described embodiment it is found that video information input deep learning network is obtained video anomalous identification knot by above-mentioned
The method of fruit can be in time to obtain accurate video anomalous identification as a result, improving the identification accuracy of anomalous event
Other vehicles are alarmed and reminded, ensure that road safety.
As shown in figure 4, in one embodiment of the invention, Fig. 4 shows that another of step S102 realizes stream in Fig. 1
Journey, details are as follows for process:
In S401:The audio frequency characteristics and sound source information of audio-frequency information are extracted from audio-frequency information.
In the present embodiment, audio collecting device may include microphone array.
In the present embodiment, audio frequency characteristics and sound source information can be extracted by the method for Audio Signal Processing, it can also
Audio-frequency information is input to fifth nerve network model extraction audio frequency characteristics and sound source information.Fifth nerve network model can be with
First nerves network model is identical.Sound source information may include sound source angle and sound source distance, wherein sound source angle be sound source and
The angle for the straight line and default zero degree line that corresponding audio collecting device is formed, sound source distance are sound source and corresponding audio collection
Air line distance between device.
In S402, audio frequency characteristics and sound source information are inputted into deep learning model, obtain audio anomalous identification result.
In the present embodiment, deep learning model can be two-way length Recognition with Recurrent Neural Network model in short-term, may include base
In the anomalous audio detection model of self study or the anomalous audio detection based on off-line learning training and disaggregated model.Passing through will
Audio frequency characteristics and sound source information input deep learning model, are identified by the detection of deep learning model, obtain audio and know extremely
Other result.Audio anomalous identification result may include audio anomalous event specifically classification and audio anomalous event sound source information.
In the present embodiment, audio anomalous event can be divided into shot, scream, hit the classifications such as sound and explosive sound.It is logical
Cross depth learning model determine audio-frequency information whether be anomalous event and anomalous event type.
It, can be by voice recognition anomalous event, to further from above-described embodiment it is found that by obtaining audio-frequency information
Improve the recognition effect of road anomalous event.
In one embodiment of the invention, step S102 can also include another implementation process, detailed process in Fig. 1
Details are as follows:
In S501, environmental information is inputted into deep learning model, obtains environmental information recognition result.
In the present embodiment, deep learning model can be convolutional neural networks model, and environment data acquisition device can be with
Including but not limited in video acquisition device, audio collecting device, temperature sensor, Air Flow Rate Instrument, water-level gauge and snow depth meter
It is one or more.Environmental information recognition result may include road breakage information, precipitation information, snowfall information, temperature information
Wind speed information, ponding information and accumulated snow information etc..
In the present embodiment, video information is obtained by video acquisition device, video information is inputted into deep learning model,
To identify road breakage information, precipitation information or rainfall data.
In the present embodiment, temperature data can be obtained by temperature sensor, wind speed information is obtained by Air Flow Rate Instrument, is led to
It crosses water-level gauge and obtains ponding information, accumulated snow information is obtained by snow depth meter.
In the present embodiment, by the way that environmental information input deep learning network to be identified, environmental information identification is obtained
As a result, environmental information recognition result may include the type of environmental information and the specific data of the category information, such as water accumulating volume, product
The specific data such as snowfall, temperature and wind speed.
From above-described embodiment it is found that by obtain audio-frequency information and identify the audio anomalous identification in audio-frequency information as a result,
It can judge anomalous event from abnormal sound, improve the recognition effect of road conditions identification.
In one embodiment of the invention, the preset rules in Fig. 1 in step S103 specifically include:
In one embodiment, audio anomalous identification result is detected in the audio-frequency information acquired from audio collecting device
Afterwards, control signal is sent to the corresponding video acquisition device of audio anomalous identification result.
In the present embodiment, when collecting audio-frequency information using audio collecting device, and audio anomalous identification result is obtained
When, in order to improve the identification accuracy of anomalous event, on the basis of audio collecting device, in conjunction with video acquisition device, pass through
Video acquisition device obtains reference information, and by reference to information and audio anomalous identification as a result, comprehensive descision obtains accurately
Final anomalous identification result.
In the present embodiment, by the audio anomalous event sound source information in audio anomalous identification result, control life is sent
The video acquisition device nearest with sound source information is enabled, and adjusts the working condition of video acquisition device according to control command, is obtained
It is used as to the higher video information of clarity and refers to information.
In one embodiment of the invention, preset rules can also include:
After detecting audio anomalous identification result in the audio-frequency information acquired from audio collecting device, know extremely to audio
The corresponding audio collecting device of other result sends control signal.
In the present embodiment, since the audio-frequency information of first passage audio collecting device acquisition is not clear enough, obtained sound
Frequency abnormal results are possible will not be accurate, and in order to improve the accuracy of audio abnormal results, the present embodiment passes through audio and ties extremely
Audio anomalous event sound source information in fruit determines sound source angle and sound source distance, when sound source angle does not reach default optimal corner
When spending, the acquisition angles that control command adjusts audio collecting device to audio collecting device can be sent, control command includes adopting
Acquisition means angular adjustment information;When sound source distance is more than default optimal distance, can send control command to meet it is default most
The audio collecting device of excellent criterion distance, and adjust the acquisition angles of the audio collecting device.Control command may include acquisition
Device number information and harvester angular adjustment information.
From above-described embodiment it is found that collecting clarity most by the audio collecting device of optimal angle and optimal distance
High audio-frequency information, to improve the accuracy of final anomalous identification result.
In one embodiment, video anomalous identification result is detected in the video information acquired from video acquisition device
Afterwards, control signal is sent to the corresponding audio collecting device of video anomalous identification result.
In the present embodiment, when collecting video information using video acquisition device, and video anomalous identification result is obtained
When, in order to improve the identification accuracy of anomalous event, on the basis of video acquisition device, in conjunction with audio collecting device, pass through
Audio collecting device obtains reference information, and by reference to information and video anomalous identification as a result, comprehensive descision obtains accurately
Final anomalous identification result.
In the present embodiment, it by the location information in video anomalous identification result, sends control command and is regarded to abnormal
Frequency source is referred to apart from nearest video acquisition device, and according to the working condition of control command adjusting video acquisition device
Information.
From above-described embodiment it is found that determining immediate audio collection by the location information in video anomalous identification result
Device obtains the higher audio-frequency information of clarity and is used as with reference to information, and carrying out comprehensive identification to reference information and video information sentences
It is disconnected, to improve the accuracy of final anomalous identification result.
In one embodiment of the invention, preset rules can also include:
After detecting video anomalous identification result in the video information acquired from video acquisition device, know extremely to video
The corresponding video acquisition device of other result sends control signal.
In the present embodiment, since the video information of first passage video acquisition device acquisition is not clear enough, what is obtained regards
Frequency abnormal results are possible will not be accurate, and in order to improve the accuracy of video abnormal results, the present embodiment passes through video and ties extremely
Location information in fruit determines video source angle and video source distance, when video source angle does not reach default optimal angle,
Control command can be sent to video acquisition device, video acquisition device adjusts the acquisition of video acquisition device according to control command
Angle, control command include harvester angular adjustment information;When video source distance is more than default optimal distance, can send
Control command adjusts the acquisition angles of the video acquisition device to meeting the video acquisition device of default optimal distance standard.
Control command may include harvester number information and harvester angular adjustment information.
In one embodiment, video anomalous identification result is detected in the video information acquired from video acquisition device
Afterwards, control signal is sent to the corresponding environment data acquisition device of video anomalous identification result.
In the present embodiment, video anomalous identification result further includes environmental abnormality information, is adopted when using video acquisition device
When collecting video information, and identifying that video anomalous identification result is environmental abnormality information, in order to further determine environmental information
Specific data, on the basis of video acquisition device, combining environmental data acquisition device obtains environmental information, and by environment
Information, which is used as, refers to information, and by reference to information and video anomalous identification as a result, comprehensive descision obtains accurate final exception
Recognition result, final anomalous identification result include the specific category of environmental information and the specific data of the category.
By taking a specific application scenarios as an example, when the video anomalous identification result that the video information of detection identifies is vehicle
When misfortune, terminal device sends control command to neighbouring environment data acquisition device, and environment data acquisition device acquires environment letter
Breath, to obtain the concrete reason to cause a traffic accident according to environmental information and video anomalous identification result, to obtain accurate road
The recognition result of condition environment provides alert further directed to traffic accident reason, prevents the generation of more traffic accidents, ensure that road is pacified
Entirely.
In one embodiment of the invention, preset rules can also include:
After detecting environmental information recognition result in the environmental information acquired from environment data acquisition device, believe to environment
It ceases the corresponding environment data acquisition device of recognition result and sends control signal.
In the present embodiment, in order to further obtain environmental information, the accuracy of abnormal traffic information recognition result is improved,
The present embodiment sends control command to corresponding environment data acquisition device, environmental data is adopted by environmental information recognition result
Acquisition means obtain the reference information of environmental information, and root according to the working condition of control command adjusting ambient data acquisition device
According to environmental information and reference information, final anomalous identification result is obtained.
By taking a specific application scenarios as an example, when getting rainfall data by environment data acquisition device, then send out
Send control command to water-level gauge, water-level gauge acquires water level information according to control command, to obtain accurate environmental data most
Whole anomalous identification result.
In one embodiment of the invention, after step S104 in Fig. 1, abnormal traffic information recognition methods is also wrapped
It includes:
According to final anomalous identification as a result, sending warning message to corresponding terminal, warning message is used to indicate terminal root
It alarms according to warning message.
In the present embodiment, the terminal may include the end that vehicle termination, particular alarm terminal or other needs are alarmed
End.
In the present embodiment, when terminal is particular alarm terminal, according to final anomalous identification as a result, directly to specific report
Alert terminal sends warning message.
In the present embodiment, when terminal is vehicle termination, warning message can be sent to vehicle end in a broadcast manner
End.Final anomalous identification result can be sent to high in the clouds first, high in the clouds sends alarm broadcast according to final anomalous identification result
To vehicle termination, to inform the final anomalous identification of vehicle road conditions as a result, evading to make vehicle carry out early warning in advance.
In the present embodiment, when final anomalous identification result is traffic accident anomalous event, it is alert to correlation to send warning message
The terminal of side is alarmed, so that the police obtain information and timely processing alert in time.Alarm broadcast to traffic accident is sent simultaneously to send out
The front vehicle terminal of road where raw position, reminds the user that front has a car accident.
From above-described embodiment it is found that by corresponding terminal send warning message, can with timely processing anomalous event with
It avoids the jams, ensure that the coast is clear and safety.
For a further understanding of the embodiment of the present invention, by taking a specific scene as an example, details are as follows:
1) video acquisition device collects video information, video information show vehicle done it is snakelike evade, by identifying
It is track of vehicle anomalous event to video anomalous identification result;
2) video acquisition device acquires video information, and obtain video anomalous identification result by identification believes for road breakage
Breath.
3) according to track of vehicle anomalous event and road breakage information, determine that final anomalous identification result is " because road is broken
Damage causes track of vehicle abnormal ".
4) warning message is sent to the vehicle of this road rear pre-determined distance, so that the terminal of vehicle shows alarm signal
Breath.For example, warning message can be " road is damaged at 500 meters of front, please evades in advance ".
5) warning message is sent to server, and server preserves warning message, to make server be inquired in other-end
The warning message is sent when road conditions to other-end.
From above-described embodiment it is found that obtaining traffic information by different harvester, can comprehensive road condition information obtain
Accurate final recognition effect, improves the accuracy of final anomalous identification result, to according to accurate abnormal traffic information
Recognition result is prevented in advance, ensure that road safety.
It should be understood that the size of the serial number of each step is not meant that the order of the execution order in above-described embodiment, each process
Execution sequence should be determined by its function and internal logic, the implementation process without coping with the embodiment of the present invention constitutes any limit
It is fixed.
Embodiment 2:
As shown in figure 5, An embodiment provides a kind of abnormal traffic information identification devices 100, for holding
The method and step in embodiment corresponding to row Fig. 1 comprising:
Traffic information acquisition module 110, the traffic information sent for obtaining at least one harvester;
Preliminary recognition result acquisition module 120 obtains traffic information for traffic information to be inputted deep learning model
Preliminary recognition result;
Signal transmitting module 130 is controlled, for being sent to the first harvester according to preset rules and preliminary recognition result
Signal is controlled, control signal is used to indicate the reference information of the first harvester acquisition road conditions;First harvester is preliminary knowledge
The corresponding harvester of other result;
Final anomalous identification result acquisition module 140, it is final different for being determined according to reference information and preliminary recognition result
Normal recognition result.
From above-described embodiment it is found that the traffic information that the embodiment of the present invention is sent by obtaining at least one harvester;
Traffic information is inputted into deep learning model, obtains the preliminary recognition result of traffic information;According to preset rules and preliminary identification
As a result control signal is sent to the first harvester, control signal is used to indicate the reference letter of the first harvester acquisition road conditions
Breath;First harvester is the corresponding harvester of preliminary recognition result;It is determined most according to reference information and preliminary recognition result
Whole anomalous identification result.The embodiment of the present invention realizes the comprehensive identification to traffic information, changes by making full use of traffic information
The recognition effect to traffic information has been apt to it, to further enhance road safety management.
In one embodiment of the invention, the preliminary recognition result acquisition module 120 in the embodiment corresponding to Fig. 5 is gone back
Include the structure for executing the method and step in the embodiment corresponding to Fig. 2 comprising:
Target object structured features acquiring unit, for video information to be inputted first nerves network model, detection regards
Target object in frequency information, and from video information extract target object structured features.
Target object recognition result acquiring unit, for the structured features of target object to be inputted nervus opticus network mould
Type obtains target object recognition result.
From above-described embodiment it is found that the video information that different cameras are obtained inputs deep learning network model, to
Multigroup structured features of a target object are obtained, the accuracy of target object recognition result can be improved, improve target
The recognition effect of object.
In one embodiment of the invention, the preliminary recognition result acquisition module 120 in the embodiment corresponding to Fig. 5 is gone back
Include the structure for executing the method and step in the embodiment corresponding to Fig. 3 comprising:
Anomalous event structured features acquiring unit is detected for video information to be inputted third nerve network model
To the structured features and location information of the anomalous event in video information.
Video anomalous identification result acquiring unit, for the structured features of anomalous event and location information to be inputted the 4th
Neural network model obtains video anomalous identification result.
From above-described embodiment it is found that video information input deep learning network is obtained video anomalous identification knot by above-mentioned
The method of fruit, to obtain the accurate video anomalous identification of anomalous event as a result, improving the identification accuracy of anomalous event,
Other vehicles timely can be alarmed and be reminded, ensure that road safety.
In one embodiment of the invention, the preliminary recognition result acquisition module 120 in the embodiment corresponding to Fig. 5 is gone back
Include the structure for executing the method and step in the embodiment corresponding to Fig. 4 comprising:
Audio frequency characteristics acquiring unit, audio frequency characteristics and sound source information for extracting audio-frequency information from audio-frequency information.
Audio anomalous identification result acquiring unit is obtained for audio frequency characteristics and sound source information to be inputted deep learning model
To audio anomalous identification result.
It, can be by voice recognition anomalous event, to further from above-described embodiment it is found that by obtaining audio-frequency information
Improve the recognition effect of road anomalous event.
In one embodiment of the invention, the preliminary recognition result acquisition module 120 in the embodiment corresponding to Fig. 5 is gone back
Including:
Environmental information recognition result acquiring unit obtains environmental information for environmental information to be inputted deep learning model
Recognition result.
In one embodiment of the invention, the control signal transmitting module 130 in the embodiment corresponding to Fig. 5 is also wrapped
It includes:
First preset rules realize unit, for detecting that audio is different in the audio-frequency information acquired from audio collecting device
After normal recognition result, control signal is sent to the corresponding video acquisition device of audio anomalous identification result.
Alternatively, the second preset rules realize unit, for being detected in the video information acquired from video acquisition device
After video anomalous identification result, control signal is sent to the corresponding audio collecting device of video anomalous identification result.
Alternatively, third preset rules realize unit, for being detected in the video information acquired from video acquisition device
After video anomalous identification result, control signal is sent to the corresponding environment data acquisition device of video anomalous identification result.
In one embodiment of the invention, abnormal traffic information identification device 100 further includes:
Alarm module, for, as a result, sending warning message to corresponding terminal, warning message to be used according to final anomalous identification
It is alarmed according to warning message in instruction terminal.
From above-described embodiment it is found that by corresponding terminal send warning message, can with timely processing anomalous event with
It avoids the jams, ensure that the coast is clear and safety.
In one embodiment, abnormal traffic information identification device 100 further includes other function module/units, for real
Method and step in current embodiment 1 in each embodiment.
Embodiment 3:
The embodiment of the present invention additionally provides a kind of terminal device 6, including memory 61, processor 60 and is stored in storage
In device 61 and the computer program 62 that can run on processor 60, processor 60 are realized such as when executing the computer program 62
The step in each embodiment described in embodiment 1, such as step S101 shown in FIG. 1 to step S104.Alternatively, the processing
Device 60 realizes the function of each module in each device embodiment as described in example 2 above when executing the computer program 62,
Such as the function of module 110 to 140 shown in fig. 5.
The terminal device 6 can be that the calculating such as desktop PC, notebook, palm PC and cloud server are set
It is standby.The terminal device 6 may include, but be not limited only to, processor 60, memory 61.Such as the terminal device 6 can also wrap
Include input-output equipment, network access equipment, bus etc..
Alleged processor 60 can be central processing unit (Central Processing Unit, CPU), can also be
Other general processors, digital signal processor (Digital Signal Processor, DSP), application-specific integrated circuit
(Application Specific Integrated Circuit, ASIC), ready-made programmable gate array (Field-
Programmable Gate Array, FPGA) either other programmable logic device, discrete gate or transistor logic,
Discrete hardware components etc..General processor can be microprocessor or the processor 60 can also be any conventional processor
60 etc..
The memory 61 can be the internal storage unit of the terminal device 6, such as the hard disk of terminal device 6 or interior
It deposits.The memory 61 can also be to be equipped on the External memory equipment of the terminal device 6, such as the terminal device 6
Plug-in type hard disk, intelligent memory card (Smart Media Card, SMC), secure digital (Secure Digital, SD) card dodge
Deposit card (Flash Card) etc..Further, the memory 61 can also both include terminal device 6 internal storage unit or
Including External memory equipment.The memory 61 is for storing needed for the computer program 62 and the terminal device 6
Other programs and data.The memory 61 can be also used for temporarily storing the data that has exported or will export.
Embodiment 4:
The embodiment of the present invention additionally provides a kind of computer readable storage medium, and computer-readable recording medium storage has meter
Calculation machine program 62 realizes the step in each embodiment as described in example 1 above when computer program 62 is executed by processor 60,
Such as step S101 shown in FIG. 1 to step S104.Alternatively, being realized when the computer program 62 is executed by processor 60 strictly according to the facts
Apply the function of each module in each device embodiment described in example 2, such as the function of module 110 to 140 shown in fig. 5.
The computer program 62 can be stored in a computer readable storage medium, which is being located
It manages when device 60 executes, it can be achieved that the step of above-mentioned each embodiment of the method.Wherein, the computer program 62 includes computer journey
Sequence code, the computer program code can be source code form, object identification code form, executable file or certain intermediate shapes
Formula etc..The computer-readable medium may include:Any entity or device, note of the computer program code can be carried
Recording medium, USB flash disk, mobile hard disk, magnetic disc, CD, computer storage, read-only memory (ROM, Read-Only Memory),
Random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium
Deng.It should be noted that the content that the computer-readable medium includes can be real according to legislation in jurisdiction and patent
The requirement trampled carries out increase and decrease appropriate, such as in certain jurisdictions, according to legislation and patent practice, computer-readable medium
It is electric carrier signal and telecommunication signal not include.
The steps in the embodiment of the present invention can be sequentially adjusted, merged and deleted according to actual needs.
Module or unit in system of the embodiment of the present invention can be combined, divided and deleted according to actual needs.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
All any modification, equivalent and improvement etc., should all be included in the protection scope of the present invention made by within refreshing and principle.
Claims (10)
1. a kind of exception traffic information recognition methods, which is characterized in that including:
Obtain the traffic information that at least one harvester is sent;
The traffic information is inputted into deep learning model, obtains the preliminary recognition result of the traffic information;
Control signal is sent to the first harvester according to preset rules and the preliminary recognition result, the control signal is used for
Indicate the reference information of the first harvester acquisition road conditions;First harvester is that the preliminary recognition result corresponds to
Harvester;
Final anomalous identification result is determined according to the reference information and the preliminary recognition result.
2. a kind of abnormal traffic information recognition methods as described in claim 1, which is characterized in that the harvester includes regarding
Frequency harvester, the traffic information include video information, and the deep learning model includes first nerves network model and
Two neural network models, the preliminary recognition result include target object recognition result;
Described that the traffic information is input to deep learning model, the preliminary recognition result for obtaining the traffic information includes:
The video information is inputted into the first nerves network model, detects the target object in the video information, and from
The structured features of the target object are extracted in the video information;
The structured features of the target object are inputted into the nervus opticus network model, obtain the target object identification knot
Fruit.
3. a kind of abnormal traffic information recognition methods as described in claim 1, which is characterized in that the harvester includes regarding
Frequency harvester, the traffic information include video information, and the deep learning model includes third nerve network model and
Four neural network models, the preliminary recognition result include video anomalous identification result;
Described that the traffic information is inputted deep learning model, the preliminary recognition result for obtaining the traffic information includes:
The video information is inputted into the third nerve network model, detection obtains the anomalous event in the video information
Structured features and location information;
The structured features of the anomalous event and location information are inputted into the fourth nerve network model, obtain video exception
Recognition result.
4. a kind of abnormal traffic information recognition methods as described in claim 1, which is characterized in that the harvester includes sound
Frequency harvester, the traffic information include audio-frequency information, and the preliminary recognition result includes audio anomalous identification result;
Described that the traffic information is inputted deep learning model, the preliminary recognition result for obtaining the traffic information includes:
The audio frequency characteristics and sound source information of the audio-frequency information are extracted from the audio-frequency information;
The audio frequency characteristics and sound source information are inputted into the deep learning model, obtain the audio anomalous identification result.
5. a kind of abnormal traffic information recognition methods as described in claim 1, which is characterized in that the harvester includes ring
Border data acquisition device, the traffic information include environmental information, and the preliminary recognition result includes environmental information recognition result;
It is described that the traffic information is inputted into deep learning model, the preliminary recognition result of the traffic information is obtained, including:
The environmental information is inputted into the deep learning model, obtains environmental information recognition result.
6. a kind of abnormal traffic information recognition methods as described in claim 1, which is characterized in that the harvester includes regarding
Frequency harvester, audio collecting device and environment data acquisition device, the traffic information include video information, audio-frequency information and
Environmental information;The preliminary recognition result includes video anomalous identification result, audio anomalous identification result and environmental information identification
As a result;
The preset rules include:
It is different to the audio after detecting audio anomalous identification result in the audio-frequency information that the audio collecting device acquires
The normal recognition result corresponding video acquisition device transmission control signal;
Or it after detecting video anomalous identification result in the video information that the video acquisition device acquires, is regarded to described
The corresponding audio collecting device of frequency anomalous identification result sends the control signal;
Or it after detecting video anomalous identification result in the video information that the video acquisition device acquires, is regarded to described
The corresponding environment data acquisition device of frequency anomalous identification result sends the control signal.
7. a kind of abnormal traffic information recognition methods as claimed in any one of claims 1 to 6, which is characterized in that in the basis
After the reference information and the preliminary recognition result determine final anomalous identification result, further include:
According to the final anomalous identification as a result, sending warning message to corresponding terminal, the warning message is used to indicate institute
Terminal is stated to be alarmed according to the warning message.
8. a kind of exception traffic information identification device, which is characterized in that including:
Traffic information acquisition module, the traffic information sent for obtaining at least one harvester;
Preliminary recognition result acquisition module obtains the traffic information for the traffic information to be inputted deep learning model
Preliminary recognition result;
Signal transmitting module is controlled, is controlled for being sent to the first harvester according to preset rules and the preliminary recognition result
Signal, the control signal are used to indicate the reference information of the first harvester acquisition road conditions;First harvester
For the corresponding harvester of the preliminary recognition result;
Final anomalous identification result acquisition module, it is final different for being determined according to the reference information and the preliminary recognition result
Normal recognition result.
9. a kind of terminal device, including memory, processor and it is stored in the memory and can be on the processor
The computer program of operation, which is characterized in that the processor realizes such as claim 1 to 7 when executing the computer program
The step of any one the method.
10. a kind of computer readable storage medium, the computer-readable recording medium storage has computer program, feature to exist
In when the computer program is executed by processor the step of any one of such as claim 1 to 7 of realization the method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810381142.5A CN108764042B (en) | 2018-04-25 | 2018-04-25 | Abnormal road condition information identification method and device and terminal equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810381142.5A CN108764042B (en) | 2018-04-25 | 2018-04-25 | Abnormal road condition information identification method and device and terminal equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108764042A true CN108764042A (en) | 2018-11-06 |
CN108764042B CN108764042B (en) | 2021-05-28 |
Family
ID=64011889
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810381142.5A Active CN108764042B (en) | 2018-04-25 | 2018-04-25 | Abnormal road condition information identification method and device and terminal equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108764042B (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109410582A (en) * | 2018-11-27 | 2019-03-01 | 易念科技(深圳)有限公司 | Traffic condition analysis method and terminal device |
CN109785633A (en) * | 2019-03-14 | 2019-05-21 | 百度在线网络技术(北京)有限公司 | Dangerous road conditions based reminding method, device, car-mounted terminal, server and medium |
CN110363070A (en) * | 2019-05-31 | 2019-10-22 | 合刃科技(武汉)有限公司 | The method, apparatus and computer program product of intelligent recognition road condition |
CN110781834A (en) * | 2019-10-28 | 2020-02-11 | 上海眼控科技股份有限公司 | Traffic abnormality image detection method, device, computer device and storage medium |
CN111310607A (en) * | 2020-01-22 | 2020-06-19 | 交通运输部公路科学研究所 | Highway safety risk identification method and system based on computer vision and artificial intelligence |
CN111357011A (en) * | 2019-01-31 | 2020-06-30 | 深圳市大疆创新科技有限公司 | Environment sensing method and device, control method and device and vehicle |
CN111931753A (en) * | 2020-10-13 | 2020-11-13 | 蘑菇车联信息科技有限公司 | Internet of vehicles AI network interface implementation method and open platform system |
CN112149763A (en) * | 2020-11-25 | 2020-12-29 | 江苏量动信息科技有限公司 | Method and device for improving road surface abnormity detection by using crowdsourcing concept |
CN112585656A (en) * | 2020-02-25 | 2021-03-30 | 华为技术有限公司 | Method and device for identifying special road conditions, electronic equipment and storage medium |
CN111357011B (en) * | 2019-01-31 | 2024-04-30 | 深圳市大疆创新科技有限公司 | Environment sensing method and device, control method and device and vehicle |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP4242422B2 (en) * | 2004-03-31 | 2009-03-25 | 日立造船株式会社 | Sudden event recording and analysis system |
CN101751782A (en) * | 2009-12-30 | 2010-06-23 | 北京大学深圳研究生院 | Crossroad traffic event automatic detection system based on multi-source information fusion |
CN103824455A (en) * | 2014-02-24 | 2014-05-28 | 宁波众信智能科技有限公司 | Road patrol system and control method thereof |
CN204168409U (en) * | 2014-07-21 | 2015-02-18 | 深圳市力艺市政工程有限公司 | Town road video and the audio frequency contrast supervisory control system of being with self-checking function |
CN106375713A (en) * | 2016-08-30 | 2017-02-01 | 黄爱国 | Intelligent video monitoring skynet system and working method thereof |
WO2017095559A1 (en) * | 2015-12-01 | 2017-06-08 | Qualcomm Incorporated | Determining audio event based on location information |
CN106907697A (en) * | 2017-04-19 | 2017-06-30 | 杭州派尼澳电子科技有限公司 | A kind of intelligent road-lamp of built-in collision sound detection function |
CN107180535A (en) * | 2017-07-12 | 2017-09-19 | 安徽金赛弗信息技术有限公司 | The joyride behavior automatic identification equipment and method of a kind of automatic sound detection based on deep learning |
-
2018
- 2018-04-25 CN CN201810381142.5A patent/CN108764042B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP4242422B2 (en) * | 2004-03-31 | 2009-03-25 | 日立造船株式会社 | Sudden event recording and analysis system |
CN101751782A (en) * | 2009-12-30 | 2010-06-23 | 北京大学深圳研究生院 | Crossroad traffic event automatic detection system based on multi-source information fusion |
CN103824455A (en) * | 2014-02-24 | 2014-05-28 | 宁波众信智能科技有限公司 | Road patrol system and control method thereof |
CN204168409U (en) * | 2014-07-21 | 2015-02-18 | 深圳市力艺市政工程有限公司 | Town road video and the audio frequency contrast supervisory control system of being with self-checking function |
WO2017095559A1 (en) * | 2015-12-01 | 2017-06-08 | Qualcomm Incorporated | Determining audio event based on location information |
CN106375713A (en) * | 2016-08-30 | 2017-02-01 | 黄爱国 | Intelligent video monitoring skynet system and working method thereof |
CN106907697A (en) * | 2017-04-19 | 2017-06-30 | 杭州派尼澳电子科技有限公司 | A kind of intelligent road-lamp of built-in collision sound detection function |
CN107180535A (en) * | 2017-07-12 | 2017-09-19 | 安徽金赛弗信息技术有限公司 | The joyride behavior automatic identification equipment and method of a kind of automatic sound detection based on deep learning |
Non-Patent Citations (2)
Title |
---|
NAOYA TAKAHASHI 等: "AENet: Learning Deep Audio Features for Video Analysis", 《IEEE TRANSACTIONS ON MULTIMEDIA》 * |
姚明海 等: "基于特征融合的视频异常事件检测方法", 《吉林大学学报( 信息科学版)》 * |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109410582A (en) * | 2018-11-27 | 2019-03-01 | 易念科技(深圳)有限公司 | Traffic condition analysis method and terminal device |
CN111357011A (en) * | 2019-01-31 | 2020-06-30 | 深圳市大疆创新科技有限公司 | Environment sensing method and device, control method and device and vehicle |
WO2020155020A1 (en) * | 2019-01-31 | 2020-08-06 | 深圳市大疆创新科技有限公司 | Environment perception method and device, control method and device, and vehicle |
CN111357011B (en) * | 2019-01-31 | 2024-04-30 | 深圳市大疆创新科技有限公司 | Environment sensing method and device, control method and device and vehicle |
CN109785633A (en) * | 2019-03-14 | 2019-05-21 | 百度在线网络技术(北京)有限公司 | Dangerous road conditions based reminding method, device, car-mounted terminal, server and medium |
CN110363070A (en) * | 2019-05-31 | 2019-10-22 | 合刃科技(武汉)有限公司 | The method, apparatus and computer program product of intelligent recognition road condition |
CN110781834A (en) * | 2019-10-28 | 2020-02-11 | 上海眼控科技股份有限公司 | Traffic abnormality image detection method, device, computer device and storage medium |
CN111310607A (en) * | 2020-01-22 | 2020-06-19 | 交通运输部公路科学研究所 | Highway safety risk identification method and system based on computer vision and artificial intelligence |
CN112585656A (en) * | 2020-02-25 | 2021-03-30 | 华为技术有限公司 | Method and device for identifying special road conditions, electronic equipment and storage medium |
CN112585656B (en) * | 2020-02-25 | 2022-06-17 | 华为技术有限公司 | Method and device for identifying special road conditions, electronic equipment and storage medium |
CN111931753A (en) * | 2020-10-13 | 2020-11-13 | 蘑菇车联信息科技有限公司 | Internet of vehicles AI network interface implementation method and open platform system |
CN112149763A (en) * | 2020-11-25 | 2020-12-29 | 江苏量动信息科技有限公司 | Method and device for improving road surface abnormity detection by using crowdsourcing concept |
Also Published As
Publication number | Publication date |
---|---|
CN108764042B (en) | 2021-05-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108764042A (en) | A kind of exception traffic information recognition methods, device and terminal device | |
CN110060484B (en) | Road passenger traffic violation real-time early warning system and method based on block chain | |
CN111462484A (en) | Congestion state determination method, device, equipment and computer readable storage medium | |
CN107798876B (en) | Road traffic abnormal jam judging method based on event | |
CN105844904B (en) | A kind of detection of abnormal vehicle behavior and method for tracing based on DSRC | |
CN103258430B (en) | Road traveling time calculating and traffic road condition judging method and road traveling time calculating and traffic road condition judging device | |
Zhang et al. | Data-driven intelligent transportation systems: A survey | |
WO2018122803A1 (en) | Smart road traffic anomaly detection method | |
KR101803891B1 (en) | The apparatus and method for smart reporting of traffic information with traffic factor and object tracking | |
CN109993969A (en) | A kind of road conditions determine information acquisition method, device and equipment | |
CN109655107A (en) | Air quality monitoring method, device, vehicle and computer readable storage medium | |
CN107705551A (en) | The appraisal procedure and device of traffic | |
CN108346288A (en) | Section operating status method for early warning, device and electronic equipment | |
CN107316457B (en) | Method for judging whether road traffic condition accords with automatic driving of automobile | |
CN108122408B (en) | Road condition monitoring method and device and system for monitoring road conditions | |
CN106228830A (en) | A kind of bus arrival time real-time estimate system and method | |
CN108701396A (en) | A kind of detection alarm method, storage medium and the server of Chinese herbaceous peony accumulated snow and icing | |
CN107689161A (en) | The intelligent automobile Real-road Driving Cycle constructing system of people's car traffic interconnection | |
CN105279963B (en) | A kind of method and device for obtaining traffic information | |
CN114140300A (en) | Method, device, storage medium and terminal for identifying vehicle stop points based on GPS data | |
CN113223296A (en) | Warning system used during highway maintenance construction | |
CN110136442A (en) | Pedestrian based on shared bicycle GPS track data occupies road Activity recognition method | |
CN110930715B (en) | Method and system for identifying red light running of non-motor vehicle and violation processing platform | |
CN115797403A (en) | Traffic accident prediction method and device, storage medium and electronic device | |
WO2015188450A1 (en) | Pre-alarm processing 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 |