CN108288063A - The meteorology on road surface determines method, apparatus and system - Google Patents
The meteorology on road surface determines method, apparatus and system Download PDFInfo
- Publication number
- CN108288063A CN108288063A CN201810017719.4A CN201810017719A CN108288063A CN 108288063 A CN108288063 A CN 108288063A CN 201810017719 A CN201810017719 A CN 201810017719A CN 108288063 A CN108288063 A CN 108288063A
- Authority
- CN
- China
- Prior art keywords
- meteorology
- road surface
- image
- image block
- feature vector
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
This application discloses a kind of meteorologies on road surface to determine method, apparatus and system, belongs to technical field of intelligent traffic.The method includes：Obtain the image on road surface；The multiple images block in the image on road surface is extracted according to default extracting mode；Obtain the target feature vector of each image block in multiple images block；The meteorology of each image block is determined according to the target feature vector of each image block；The meteorology distribution situation on road surface is determined according to the meteorology of multiple images block.Present application addresses only the relatively low problem of the accuracy of the meteorology on road surface is determined by the corresponding meteorology in the central area of image, the accuracy of the meteorology on determining road surface is improved.Determination of the application for the meteorology on road surface.
Description
Technical field
This application involves technical field of intelligent traffic, more particularly to a kind of meteorology on road surface determine method, apparatus and
System.
Background technology
The meteorology on road surface includes：Drying, humidity, ponding, accumulated snow, icing etc., with the development of intelligent transport technology,
Determine that the meteorology on road surface is more and more important for communications and transportation.
In the related technology, the meteorology on road surface is determined by analyzing the image on road surface.It is exemplary, it can obtain first
The image of traffic monitoring apparatus shooting, due to the image that the central area of the image of traffic monitoring apparatus shooting is road surface, so
Need to extract the part of the central area of the image.Then the image section of the extraction is analyzed to determine the image section
Corresponding meteorology, and the meteorology is determined as to the meteorology on road surface.
Due in the related technology, the meteorological shape on road surface only being determined by the corresponding meteorology in the central area of image
State, it is thus determined that pavement state accuracy it is relatively low.
Invention content
This application provides a kind of meteorologies on road surface to determine method, apparatus and system, can solve only to pass through image
The corresponding meteorology in central area determine the relatively low problem of the accuracy of the meteorology on road surface.The technical solution is such as
Under：
On the one hand, it provides a kind of meteorology on road surface and determines method, the method includes：
Obtain the image on road surface；
The multiple images block in the image on the road surface is extracted according to default extracting mode；
Obtain the target feature vector of each image block in described multiple images block；
The meteorology of each image block is determined according to the target feature vector of each image block；
The meteorology distribution situation on the road surface is determined according to the meteorology of described multiple images block.
Optionally, the target feature vector according to each image block determines the meteorological shape of each image block
State, including：
It obtains and presets disaggregated model, the default disaggregated model is used to determine image according to the target feature vector of image block
The meteorology of block, the default disaggregated model include at least one function, and the parameter at least one function is basis
What multiple feature vector sample trainings obtained；
The target feature vector of each image block is inputted into the default disaggregated model, to obtain each image
The meteorology of block.
Optionally, the default disaggregated model includes support vector machines.
Optionally, the meteorology of each image block is determined in the target feature vector according to each image block
After state, the method further includes：
Determine the corresponding number accounting of each meteorology, wherein the corresponding number accounting of described each meteorology is：
Image block with each meteorology is in described multiple images number accounting in the block.
Optionally, the target feature vector for obtaining each image block in described multiple images block, including：
Obtain the color feature vector of each image block；
Obtain the texture feature vector of each image block；
The color feature vector and texture feature vector of each image block are normalized, obtained described every
The target feature vector of a image block.
Optionally, the color feature vector is nine dimension color moment vectors, and the texture feature vector is four dimensional vectors, institute
It is ten three-dimensional vectors to state target feature vector.
Optionally, the amount of pixels positive correlation that the image on the amount of pixels that each described image block includes and the road surface includes.
On the other hand, a kind of meteorology determining device on road surface, the meteorology determining device on the road surface are provided
Including：
First acquisition module, the image for obtaining road surface；
Extraction module, the multiple images block in image for extracting the road surface according to default extracting mode；
Second acquisition module, the target feature vector for obtaining each image block in described multiple images block；
First determining module, for determining each image block according to the target feature vector of each image block
Meteorology；
Second determining module, the meteorology point for determining the road surface according to the meteorology of described multiple images block
Cloth situation.
Optionally, first determining module is additionally operable to：
It obtains and presets disaggregated model, the default disaggregated model is used to determine image according to the target feature vector of image block
The meteorology of block, the default disaggregated model include at least one function, and the parameter at least one function is basis
What multiple feature vector sample trainings obtained；
The target feature vector of each image block is inputted into the default disaggregated model, to obtain each image
The meteorology of block.
Optionally, the default disaggregated model includes support vector machines.
Optionally, the meteorology determining device on the road surface further includes：
Third determining module, for determining the corresponding number accounting of each meteorology, wherein each described meteorology
Corresponding number accounting is：Image block with each meteorology is in described multiple images number accounting in the block.
Optionally, second acquisition module is additionally operable to：
Obtain the color feature vector of each image block；
Obtain the texture feature vector of each image block；
The color feature vector and texture feature vector of each image block are normalized, obtained described every
The target feature vector of a image block.
Optionally, the color feature vector is nine dimension color moment vectors, and the texture feature vector is four dimensional vectors, institute
It is ten three-dimensional vectors to state target feature vector.
The amount of pixels positive correlation that the image of amount of pixels and the road surface that each described image block includes includes.
In another aspect, providing a kind of meteorology on road surface determines that system, the meteorology on the road surface determine system
Meteorology determining device including above-mentioned road surface.
The advantageous effect that technical solution provided by the present application is brought is：
This application provides a kind of meteorologies on road surface to determine method, apparatus and system, and the meteorology on road surface is true
Determine in method, the multiple images block in the image on road surface can be extracted, and determine according to the target feature vector of each image block
The meteorology of each image block, and then determine the meteorology distribution situation on road surface.Since it is determined that road surface meteorology
State distribution situation can characterize the meteorology of all areas on road surface, therefore improve the standard of the meteorology on determining road surface
True property.
Description of the drawings
To describe the technical solutions in the embodiments of the present invention more clearly, make required in being described below to embodiment
Attached drawing is briefly described, it should be apparent that, the accompanying drawings in the following description is only some embodiments of the present application, for
For those of ordinary skill in the art, without creative efforts, other are can also be obtained according to these attached drawings
Attached drawing.
Fig. 1 is that a kind of meteorology on road surface provided in an embodiment of the present invention determines the schematic diagram of system；
Fig. 2 is that a kind of meteorology on road surface provided in an embodiment of the present invention determines the method flow diagram of method；
Fig. 3 is that the meteorology on another road surface provided in an embodiment of the present invention determines the method flow diagram of method；
Fig. 4 is a kind of schematic diagram of the image on road surface provided in an embodiment of the present invention；
Fig. 5 is a kind of schematic diagram of multiple images block provided in an embodiment of the present invention；
Fig. 6 is a kind of schematic diagram of the image block on the road surface of known meteorology provided in an embodiment of the present invention；
Fig. 7 is a kind of meteorology distribution schematic diagram on road surface provided in an embodiment of the present invention；
Fig. 8 is that the meteorology distribution situation on another road surface provided in an embodiment of the present invention shows schematic diagram；
Fig. 9 is a kind of structural schematic diagram of the meteorology determining device on road surface provided in an embodiment of the present invention；
Figure 10 is the structural schematic diagram of the meteorology determining device on another road surface provided in an embodiment of the present invention.
Specific implementation mode
To keep the purpose, technical scheme and advantage of the application clearer, below in conjunction with attached drawing to the application embodiment party
Formula is described in further detail.
With the development of intelligent transport technology, determine that the meteorology on road surface is most important for communications and transportation.If can be with
The meteorology for accurately knowing each local road surface, then can preferably carry out traffic administration, or avoid by meteorology
Traffic safety is improved on the road surface of poor (such as accumulated snow, the meteorology of icing).
Fig. 1 is that a kind of meteorology on road surface provided in an embodiment of the present invention determines the schematic diagram of system.As shown in Figure 1,
The meteorology on the road surface determines that system 10 may include：The meteorology determining device 101 on road surface.Optionally, the road surface
Meteorology determines that system 10 can also include：Image collecting device 102 and client 103.
Exemplary, the meteorology determining device 101 on road surface can be computer, server or server cluster；The figure
As harvester 102 can be traffic monitoring apparatus (camera installed on such as road), it should be noted that image collector
Set or can acquire road image other equipment, it is not limited in the embodiment of the present invention；Client 103 can be with
It is deployed in terminal, terminal can be that the electronics such as mobile phone, laptop, desktop computer, tablet computer, intelligent TV set are set
It is standby.
The image collecting device 102 and the meteorology determining device 101 and image collecting device 102 on road surface and this
Client 103 can pass through cable network or wireless network connection, wherein cable network can include but is not limited to：It is logical
With universal serial bus (English：Universal Serial Bus；Referred to as：USB), wireless network can include but is not limited to：Wirelessly
Fidelity (English：Wireless Fidelity；Referred to as：WIFI), bluetooth, infrared, purple honeybee (English：Zigbee), data etc..
The meteorology determining device 101 on road surface can obtain the image on road surface by image collecting device 102, and carry out
Then the meteorology distribution situation on the road surface can be sent to and road by analysis with determining the meteorology distribution situation on road surface
The client 103 that the meteorology determining device 101 in face is connected.
Optionally, the meteorology on the road surface determines that system 10 can also include light compensating apparatus (Fig. 1 is not shown), light filling dress
Setting can be arranged on image collecting device 102.When ambient brightness residing for the road surface is relatively low, which can be with road pavement
Light filling is carried out, so that image collecting device can acquire the image on road surface.
Optionally, which can be with image collecting device, the meteorology determining device on road surface and light compensating apparatus
It is all connected with.User can obtain the image on the road surface of image acquisition device by client, or be controlled by client
Light compensating lamp road pavement carries out light filling and adjusts light filling intensity, and the meteorology determining device on road surface can also be controlled by client
Determine the frequency meteorology of a determining road surface (such as every 5 hours) of the meteorology on road surface.
Fig. 2 is that a kind of meteorology on road surface provided in an embodiment of the present invention determines the method flow diagram of method.This method
The meteorology determining device that can be used for the road surface in Fig. 1, as shown in Fig. 2, this method may include：
Step 201, the image for obtaining road surface.
Step 202, according to default extracting mode extract road surface image in multiple images block.
Step 203, the target feature vector for obtaining each image block in multiple images block.
Step 204, the meteorology that each image block is determined according to the target feature vector of each image block.
Step 205, the meteorology distribution situation that road surface is determined according to the meteorology of multiple images block.
In conclusion the meteorology on road surface provided in an embodiment of the present invention determines in method, the figure on road surface can be extracted
Multiple images block as in, and determine according to the target feature vector of each image block the meteorology of each image block, into
And determine the meteorology distribution situation on road surface.Since it is determined that the meteorology distribution situation on road surface can characterize road surface
The meteorology of all areas, therefore improve the accuracy of the meteorology on determining road surface.
Fig. 3 is that the meteorology on another road surface provided in an embodiment of the present invention determines the method flow diagram of method.The party
Method can be used for the meteorology determining device on the road surface in Fig. 1, as shown in figure 3, this method may include：
Step 301, the image for obtaining road surface.
Image collecting device can acquire image in real time, and image acquisition device to image include road surface figure
Picture, the meteorology determining device on road surface can obtain the image on road surface by image collecting device connected to it.
Exemplary, the image collected can be sent to the meteorology determining device on road surface by image collecting device, with
Road surface can be identified and extracted from image collecting device acquired image convenient for the meteorology determining device on road surface
Image.The image resolution ratio of image acquisition device can be 1080P (English：Progressive；Chinese：Progressive scan),
Pickup area may range from 10 square metres, and the image of acquisition can be stored as the picture format of " .jpg ", if the Image Acquisition
Device acquires image at night, then the corresponding light compensating lamp of the image collection assembly can be to image acquisition region light filling, the light filling
Intensity can be 500~1000 luxs.
Step 302, according to default extracting mode extract road surface image in multiple images block.
When extracting image block, the meteorology determining device on road surface can extract the figure on road surface according to default extracting mode
Multiple images block as in.Such as, the meteorology determining device on road surface can obtain amount of pixels range and unit pixel amount first
Default correspondence, and the amount of pixels for including according to the image on road surface, where determining the amount of pixels that the image on road surface includes
Amount of pixels range and the corresponding target unit amount of pixels of the amount of pixels range；Then, the meteorology determining device on road surface can
According to the target unit amount of pixels, to extract the multiple images block in the image on road surface, and can be according to the target unit pixel
Amount determine each image block the length of side and comprising amount of pixels.
It should be noted that the image for the amount of pixels and road surface for including according to each image block that default extracting mode extracts
Including amount of pixels can be with positive correlation namely the amount of pixels range and amount of pixels range in the default correspondence of unit pixel amount
Can be with positive correlation with unit pixel amount, the amount of pixels that the image on road surface includes determines the amount of pixels that each image block includes.If
The amount of pixels that the image on road surface includes is larger, then the amount of pixels that each image block extracted includes is also larger, avoids and carries
The number of the image block obtained is excessive, ensure that road surface meteorology determining device can faster to these image blocks into
Row processing.
Exemplary, the default correspondence of amount of pixels range and unit pixel amount can be as shown in table 1.If image collector
Set the image of acquisition as shown in figure 4, and the image includes the image 40 on road surface, then can extract the image in step 301
The image section on middle road surface.Assuming that the amount of pixels that the image on the road surface is included is 2500000, then the pixel where the amount of pixels
Amount ranging from [2000000,3000000), target unit amount of pixels be 200 × 200.Then the meteorology determining device on road surface
It can be extracted according to the target unit amount of pixels by the image processing function of the softwares such as MATLAB (also referred to as matrix labotstory)
Multiple images block in the image on road surface so that each image block of extraction is the square that the length of side is 200 pixels.Assuming that road surface
Image border region can not according to target unit amount of pixels formed the length of side be 200 pixels square, then can be by the nothing
The region that method forms complete square is ignored into image block, so that in extracting the image on road surface shown in Fig. 5
Multiple images block.
It should be noted that the embodiment of the present invention is only for extracting obtained image block and be square, practical application
In, the image block extracted may be rectangle, and the embodiment of the present invention is not construed as limiting this.
Table 1
| Amount of pixels range | Unit pixel amount |
| [100000,1000000) | 80×80 |
| [1000000,2000000) | 100×100 |
| [2000000,3000000) | 200×200 |
| [3000000,5000000) | 300×300 |
| [5000000,8500000) | 500×500 |
Step 303, the target feature vector for obtaining each image block in multiple images block.
Each the target feature vector of image block may include：Color feature vector (namely vector of characterization color characteristic)
With texture feature vector (namely vector of characterization textural characteristics).Wherein, color characteristic may include：Tone, saturation degree and bright
Degree, textural characteristics may include：Energy, entropy, contrast and correlation.The meteorology determining device on road surface may include image
Processor, image processor can obtain the color feature vector and texture feature vector of each image block, then by the color
Feature vector and texture feature vector are normalized, to obtain the target feature vector of each image block.
Exemplary, first, the color characteristic that three rank color moments methods extract each image block may be used in image processor, figure
The expression formula of the color characteristic of picture is：
Wherein, i is Color Channel, and j is grey scale pixel value, μiThe single order color moment for being image under i Color Channels, σiFor
Second order color moment of the image under i Color Channels, siThe three rank color moments for being image under i Color Channels, pi,jFor i in image
The frequency that the pixel that gray value is j in Color Channel occurs, N are the number of pixels in image.
The color feature vector of each image block can be used based on HSV (English：Hue, Saturation, Value；Chinese：
Tone, saturation degree, brightness) color model nine dimension color moment vectors indicate, this nine dimension color moment vector expression formula be：Fcolor
=[μH,σH,sH,μS,σS,sS,μV,σV,sV]；
Wherein, FcolorIndicate the nine dimensions color moment vector, μHThe single order color moment for being the image block under chrominance component, σH
The second order color moment for being the image block under chrominance component, sHThe three rank color moments for being the image block under chrominance component, μSFor this
Single order color moment of the image block under saturation degree component, σSThe second order color moment for being the image block under saturation degree component, sSFor this
Three rank color moments of the image block under saturation degree component, μVThe single order color moment for being the image block under luminance component, σVFor the figure
As second order color moment of the block under luminance component, sVThe three rank color moments for being the image block under luminance component.
Then, four textural characteristics that gray level co-occurrence matrixes extract each image block may be used in image processor：Energy
Feature ASM=∑ ∑s q (x, y | d, θ)2, entropy feature ENT=- ∑ ∑s q (x, y | d, θ) logq (x, y | d, θ), contrast metric
CON=∑s ∑ (x-y)2q(x,y|d,θ) and correlative character
Wherein, X, y indicate that the gray value of pixel, d indicate the spatial relation between two pixels, θ
Indicate the generation direction of gray level co-occurrence matrixes；Q indicates that position relationship is d on image, and gray value is respectively x, and two pixels of y go out
Existing number.
Finally, obtained color feature vector and textural characteristics can be normalized in image processor, obtain
Ten objective feature vectors of each image block：
Feature=[μH,σH,sH,μS,σS,sS,μV,σV,sV,ASM,ENT,CON,COR]。
Optionally, the tonal gradation of image can be compressed to 16 grades namely above-mentioned gray value be tonal gradation is 16 grades
Image in pixel gray value, with improve obtain the speed of target feature vector.
Step 304, the meteorology that each image block is determined according to the target feature vector of each image block.
When the meteorology determining device on road surface gets the target of each image block in multiple images block in step 303
After feature vector, the default classification of the meteorology for determining image block according to the target feature vector of image block can be obtained
Model.Then, the target feature vector of each image block can be inputted the default classification by the meteorology determining device on road surface
Model, to obtain the meteorology of each image block of the default disaggregated model output.
Exemplary, default disaggregated model may include at least one function, and the parameter at least one function can root
It is trained to obtain according to multiple feature vector samples, and default disaggregated model may include support vector machines (English：support
vector machine；Referred to as：SVM), and using SVM determine that the rate of the meteorology of each image block is very fast.
It should be noted that the default disaggregated model can be that staff is trained to obtain to disaggregated model in advance
's.Before train classification models, staff can be by the image on the road surface of multiple known meteorologies in SQL Server
Block (the parts of images block that Fig. 6 is shown in which) sequentially inputs image processor so that image processor can extract multiple
The target feature vector of each image block in image block, to obtain sample characteristics database.The sample characteristics database can be with
Including：Each meteorology of image block and the correspondence of target feature vector.In order to ensure the default classification of training gained
Model can effectively determine that the meteorology of each image block, the number of the pavement image block of multiple known meteorology need
More than or equal to 2500, and the number of the pavement image of each meteorology need to be more than or equal to 500.
It should be noted that the full name in English of SQL Server is Structured Query Language Server,
Also it is called structured query language management system.SQL Server can update and upgrade, and that is to say more in SQL Server
The image block on the road surface of a known meteorology can update.After SQL Server update or upgrade every time, image processor
It is required to extract the target feature vector of updated image block again, and obtains new sample characteristics database.
Exemplary, the disaggregated model being trained to may includeA SVM, wherein k is of preset meteorology
Number, each SVM can obtain meteorology according to the target feature vector of each image block.Disaggregated model can should
The meteorology that a SVM is obtained is counted, and the highest meteorology of accounting is determined as the image block that disaggregated model obtains
Meteorology.Later, disaggregated model can compare the meteorology that this is obtained with the actual meteorology of the image block
Compared with, and the parameter in SVM is adjusted according to comparison result, and the process of the adjusting parameter is repeated several times, so that finally obtained
Disaggregated model can determine the meteorology of accurate image block according to the target feature vector of image block.
Exemplary, each SVM has gaussian kernel function factor C and penalty factor g, is trained to the disaggregated model
When, multiple target feature vectors in sample characteristics database can be sequentially input SVM by disaggregated model so that SVM is according to this
Each target feature vector exports meteorology.Then, disaggregated model can pass through grid search (English：Grid Search)
Method or least square method adjust the gaussian kernel function factor C and penalty factor g of SVM until determining preferably Gaussian kernel in SVM
The value of function factor C and penalty factor g.Later, staff can will have preferably gaussian kernel function factor C and punish
The SVM of penalty factor g is determined as default disaggregated model.If SVM has the preferably gaussian kernel function factor C and penalty factor g,
The SVM can export the higher meteorology of accuracy according to the target feature vector of input.
Optionally, staff can be by the target feature vector of the parts of images block in sample characteristics database for instructing
Practice disaggregated model to obtain default disaggregated model, by the target feature vector of another part image block for verifying the default classification
The accuracy of model.This for train classification models partial target feature vector can account for image target signature in the block to
The 90% of amount, the 20% of image target feature vector in the block can be accounted for by verifying the partial target feature vector of accuracy.
It is exemplary, it is assumed that the number of the target feature vector in sample characteristics database is 2500, each meteorology
The number of image block is 500, and ranging from [the 2 of gaussian kernel function factor C-8, 28], ranging from [the 2 of penalty factor g-8, 28],
Can then the target feature vector of 2000 image blocks be used to be trained disaggregated model, training result can be as shown in table 2；
The target feature vector for 500 image blocks for being not used for training can be used to verify the accuracy for presetting disaggregated model,
Verification result can be as shown in table 3.By table 2 with table 3 it is found that the time consumption for training for disaggregated model is shorter, and what is obtained presets
The accuracy of identification of disaggregated model is more than 90%, and the accuracy of identification is higher.
Table 2
| Training image block number | Time consumption for training (second) | The relatively figure of merit of C | The relatively figure of merit of g | Accuracy of identification |
| 2000 | 2.6482 | 16 | 0.5 | 83.57% |
Table 3
| Authentication image block number | Accuracy of identification |
| 500 | 81.02% |
Step 305, the meteorology distribution situation that road surface is determined according to the meteorology of multiple images block.
In the image for determining road surface after the meteorology of each image block, it may be determined that the meteorology on road surface is distributed feelings
Condition, and can be that corresponding mark is arranged in each meteorology, intuitively to indicate the meteorology distribution situation on road surface.If road
The different location in face is in different meteorologies, then the distribution situation of the meteorology on the road surface that the embodiment of the present invention determines can
To show the mixing meteorology on the road surface.
It is exemplary, different meteorologies can be identified with different numbers, such as the corresponding mark of drying regime
It is 1, dampness is corresponding to be identified as 2, and ponding state is corresponding to be identified as 3, and accumulated snow state is corresponding to be identified as 4, icing shape
State is corresponding to be identified as 5.After the meteorology determining device on road surface determines the meteorology of each image block, can directly it exist
The position of each image block, shows the corresponding number of the meteorology of each image block in the image on road surface.It is assumed that in step
The meteorology of each image block shown in fig. 6 determined in 304 is drying, then the position of all image blocks can be shown
1, meteorology distribution situation can be as shown in Figure 7；Assuming that there is the meteorology of parts of images block to be in image block shown in fig. 6
The meteorology of humidity, remaining image block is drying, then the position of the image block of the dampness can show 2, remaining image
The position of block can show 1, and meteorology distribution situation can be as shown in Figure 8.
Step 306 determines the corresponding number accounting of each meteorology.
The corresponding number accounting of meteorology is：Image block with this kind of meteorology is in multiple images number in the block
Accounting.It is exemplary, if the image on road surface is extracted 48 image blocks altogether, the meteorology for there are 9 image blocks is determined in step 304
State is humidity, and the meteorology of remaining 39 image blocks is drying, then the meteorology determining device on road surface can determine tide
The corresponding number accounting of wet meteorology isThe corresponding number accounting of dry meteorology is
Optionally, the meteorology determining device on road surface can be by the meteorology point on road surface determining in step 305
Cloth situation, and the corresponding number accounting of each meteorology determined within step 306 are sent to its client connected.
It, can be by the highest gas of number accounting after the corresponding number accounting of user's each meteorology in the image for receiving the road surface
As state is determined as the meteorology on the road surface.If the number accounting difference of several meteorologies is smaller, user can also basis
Specific requirements determine the meteorology on road surface.Exemplary, if the number accounting of drying regime is 90%, the number of dampness accounts for
Than being 10%, then drying regime can be determined as the main meteorological state on road surface by user；If the number accounting of drying regime is
10%, the number accounting of ponding state is 44%, and the number accounting of accumulated snow state is 46%, then user can be true by ponding state
It is set to the main meteorological state on road surface, accumulated snow state can also be determined as to the main meteorological state on road surface.
The meteorology on road surface provided in an embodiment of the present invention determines in method, can obtain the image on road surface first, and
By the multiple images block in the image for extracting the road surface according to default extracting mode；It is then possible to extract each image block respectively
Target feature vector, and then carry out the identification of meteorology respectively to each image block, finally obtain the meteorology on road surface
Distribution situation and the corresponding number accounting of each meteorology that has of road surface.
In conclusion the meteorology on road surface provided in an embodiment of the present invention determines in method, the figure on road surface can be extracted
Multiple images block as in, and determine according to the target feature vector of each image block the meteorology of each image block, into
And determine the meteorology distribution situation on road surface.Since it is determined that the meteorology distribution situation on road surface can characterize road surface
The meteorology of all areas, therefore improve the accuracy of the meteorology on determining road surface.
Fig. 9 is a kind of structural schematic diagram of the meteorology determining device on road surface provided in an embodiment of the present invention.The road surface
Meteorology determining device can be Fig. 1 in road surface meteorology determining device, as shown in figure 9, the meteorology on the road surface
State determination device 90 may include：
First acquisition module 901, the image for obtaining road surface.
Extraction module 902, the multiple images block in image for extracting road surface according to default extracting mode.
Second acquisition module 903, the target feature vector for obtaining each image block in multiple images block.
First determining module 904, the meteorology for determining each image block according to the target feature vector of each image block
State.
Second determining module 905, for determining that the meteorology on road surface is distributed feelings according to the meteorology of multiple images block
Condition.
In conclusion in the meteorology determining device on road surface provided in an embodiment of the present invention, extraction module can extract
Multiple images block in the image on road surface, the first determining module can determine that this is every according to the target feature vector of each image block
The meteorology of a image block, and then the second determining module can determine the meteorology distribution situation on road surface.Really due to second
The meteorology distribution situation on the road surface that cover half block is determined can characterize the meteorology of all areas on road surface, therefore improve
The accuracy of the meteorology on determining road surface.
Optionally, the first determining module can be also used for：
It obtains and presets disaggregated model, default disaggregated model is used to determine image block according to the target feature vector of image block
Meteorology, default disaggregated model include at least one function, and the parameter at least one function is according to multiple feature vectors
What sample training obtained；
Disaggregated model is preset into the target feature vector input of each image block, to obtain the meteorological shape of each image block
State.
Optionally, default disaggregated model may include support vector machines.
Optionally, Figure 10 is a kind of structural representation of the meteorology determining device on road surface provided in an embodiment of the present invention
Figure.On the basis of Fig. 9, the meteorology determining device 90 on the road surface can also include：
Third determining module 906, for determining the corresponding number accounting of each meteorology, wherein each meteorology
Corresponding number accounting is：Image block with each meteorology is in multiple images number accounting in the block.
Optionally, the second acquisition module can be also used for：
Obtain the color feature vector of each image block；
Obtain the texture feature vector of each image block；
The color feature vector and texture feature vector of each image block are normalized, each image block is obtained
Target feature vector.
Optionally, color feature vector be nine dimension color moments vector, texture feature vector be four dimensional vectors, target signature to
Amount is ten three-dimensional vectors.
Optionally, the amount of pixels positive correlation that the image for the amount of pixels and road surface that each image block includes includes.
In conclusion in the meteorology determining device on road surface provided in an embodiment of the present invention, extraction module can extract
Multiple images block in the image on road surface, the first determining module can determine that this is every according to the target feature vector of each image block
The meteorology of a image block, and then the second determining module can determine the meteorology distribution situation on road surface.Really due to second
The meteorology distribution situation on the road surface that cover half block is determined can characterize the meteorology of all areas on road surface, therefore improve
The accuracy of the meteorology on determining road surface.
The embodiment of the present invention additionally provides a kind of meteorology on road surface and determines system, and the meteorology on the road surface determines system
System can be with as shown in Figure 1, the meteorology on the road surface determines that system may include the meteorology on the road surfaces Fig. 8 or shown in Fig. 9
Determining device.
It should be noted that：The meteorology determining device on the road surface that above-described embodiment provides is in the meteorological shape for determining road surface
It, only the example of the division of the above functional modules, can be as needed and by above-mentioned function in practical application when state
Distribution is completed by different function modules, i.e., the internal structure that the meteorology on road surface determines is divided into different function moulds
Block, to complete all or part of the functions described above.
It should be noted that embodiment of the method provided in an embodiment of the present invention can mutually join with corresponding device embodiment
It examines, it is not limited in the embodiment of the present invention.The sequencing of embodiment of the method step provided in an embodiment of the present invention can be into
The appropriate adjustment of row, step also according to circumstances can accordingly be increased and decreased, and any one skilled in the art is in this Shen
In the technical scope that please be disclosed, the method that can readily occur in variation should all cover within the protection domain of the application, therefore not
It repeats again.
One of ordinary skill in the art will appreciate that realizing that all or part of step of above-described embodiment can pass through hardware
It completes, relevant hardware can also be instructed to complete by program, the program can be stored in a kind of computer-readable
In storage medium, storage medium mentioned above can be read-only memory, disk or CD etc..
The foregoing is merely the alternative embodiments of the application, not to limit the application, it is all in spirit herein and
Within principle, any modification, equivalent replacement, improvement and so on should be included within the protection domain of the application.
Claims (10)
1. a kind of meteorology on road surface determines method, which is characterized in that the method includes：
Obtain the image on road surface；
The multiple images block in the image on the road surface is extracted according to default extracting mode；
Obtain the target feature vector of each image block in described multiple images block；
The meteorology of each image block is determined according to the target feature vector of each image block；
The meteorology distribution situation on the road surface is determined according to the meteorology of described multiple images block.
2. according to the method described in claim 1, it is characterized in that, the target feature vector according to each image block
Determine the meteorology of each image block, including：
It obtains and presets disaggregated model, the default disaggregated model is used to determine image block according to the target feature vector of image block
Meteorology, the default disaggregated model include at least one function, and the parameter at least one function is according to multiple
Feature vector sample training obtains；
The target feature vector of each image block is inputted into the default disaggregated model, to obtain each image block
Meteorology.
3. according to the method described in claim 2, it is characterized in that, the default disaggregated model includes support vector machines.
4. according to the method described in claim 1, it is characterized in that, the target signature according to each image block to
After amount determines the meteorology of each image block, the method further includes：
Determine the corresponding number accounting of each meteorology, wherein the corresponding number accounting of described each meteorology is：Have
The image block of each meteorology is in described multiple images number accounting in the block.
5. according to the method described in claim 1, it is characterized in that, described obtain each image block in described multiple images block
Target feature vector, including：
Obtain the color feature vector of each image block；
Obtain the texture feature vector of each image block；
The color feature vector and texture feature vector of each image block are normalized, each figure is obtained
As the target feature vector of block.
6. according to the method described in claim 5, it is characterized in that, the color feature vector ties up color moments vector, institute for nine
It is four dimensional vectors to state texture feature vector, and the target feature vector is ten three-dimensional vectors.
7. according to the method described in claim 1, it is characterized in that,
The amount of pixels positive correlation that the image of amount of pixels and the road surface that each described image block includes includes.
8. a kind of meteorology determining device on road surface, which is characterized in that the meteorology determining device on the road surface includes：
First acquisition module, the image for obtaining road surface；
Extraction module, the multiple images block in image for extracting the road surface according to default extracting mode；
Second acquisition module, the target feature vector for obtaining each image block in described multiple images block；
First determining module, the meteorology for determining each image block according to the target feature vector of each image block
State；
Second determining module, for determining that the meteorology on the road surface is distributed feelings according to the meteorology of described multiple images block
Condition.
9. the meteorology determining device on road surface according to claim 8, which is characterized in that first determining module is also
For：
It obtains and presets disaggregated model, the default disaggregated model is used to determine image block according to the target feature vector of image block
Meteorology, the default disaggregated model include at least one function, and the parameter at least one function is according to multiple
Feature vector sample training obtains；
The target feature vector of each image block is inputted into the default disaggregated model, to obtain each image block
Meteorology.
10. a kind of meteorology on road surface determines system, which is characterized in that the meteorology on the road surface determines that system includes power
Profit requires the meteorology determining device on the road surface described in 8 or 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810017719.4A CN108288063A (en) | 2018-01-09 | 2018-01-09 | The meteorology on road surface determines method, apparatus and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810017719.4A CN108288063A (en) | 2018-01-09 | 2018-01-09 | The meteorology on road surface determines method, apparatus and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108288063A true CN108288063A (en) | 2018-07-17 |
Family
ID=62835193
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810017719.4A Pending CN108288063A (en) | 2018-01-09 | 2018-01-09 | The meteorology on road surface determines method, apparatus and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108288063A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114241777A (en) * | 2022-01-06 | 2022-03-25 | 交通运输部路网监测与应急处置中心 | Multi-source heterogeneous networking road condition monitoring early warning system and method |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101610357A (en) * | 2008-06-18 | 2009-12-23 | 株式会社理光 | Camera head and method of detecting road status |
| CN102695040A (en) * | 2012-05-03 | 2012-09-26 | 中兴智能交通（无锡）有限公司 | Parallel high definition video vehicle detection method based on GPU |
| US8436902B2 (en) * | 2007-08-30 | 2013-05-07 | Valeo Schalter And Sensoren Gmbh | Method and system for weather condition detection with image-based road characterization |
| CN104134080A (en) * | 2014-08-01 | 2014-11-05 | 重庆大学 | Method and system for automatically detecting roadbed collapse and side slope collapse of road |
| US20150363654A1 (en) * | 2014-06-12 | 2015-12-17 | GM Global Technology Operations LLC | Vision-based wet road surface detection using mirrored and real images |
| CN105469099A (en) * | 2015-11-20 | 2016-04-06 | 南京理工大学 | Sparse-representation-classification-based pavement crack detection and identification method |
| CN102901489B (en) * | 2011-07-25 | 2016-09-07 | 烟台惠通网络技术有限公司 | Surface gathered water accumulated ice detection method and device |
| CN106156793A (en) * | 2016-06-27 | 2016-11-23 | 西北工业大学 | Extract in conjunction with further feature and the classification method of medical image of shallow-layer feature extraction |
| CN106210471A (en) * | 2016-07-19 | 2016-12-07 | 成都百威讯科技有限责任公司 | A kind of outdoor face recognition method and system |
| CN106803252A (en) * | 2017-01-16 | 2017-06-06 | 广东容祺智能科技有限公司 | A kind of dirty positioning of power transmission line column number plate and automatic testing method |
-
2018
- 2018-01-09 CN CN201810017719.4A patent/CN108288063A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8436902B2 (en) * | 2007-08-30 | 2013-05-07 | Valeo Schalter And Sensoren Gmbh | Method and system for weather condition detection with image-based road characterization |
| CN101610357A (en) * | 2008-06-18 | 2009-12-23 | 株式会社理光 | Camera head and method of detecting road status |
| CN102901489B (en) * | 2011-07-25 | 2016-09-07 | 烟台惠通网络技术有限公司 | Surface gathered water accumulated ice detection method and device |
| CN102695040A (en) * | 2012-05-03 | 2012-09-26 | 中兴智能交通（无锡）有限公司 | Parallel high definition video vehicle detection method based on GPU |
| US20150363654A1 (en) * | 2014-06-12 | 2015-12-17 | GM Global Technology Operations LLC | Vision-based wet road surface detection using mirrored and real images |
| CN104134080A (en) * | 2014-08-01 | 2014-11-05 | 重庆大学 | Method and system for automatically detecting roadbed collapse and side slope collapse of road |
| CN105469099A (en) * | 2015-11-20 | 2016-04-06 | 南京理工大学 | Sparse-representation-classification-based pavement crack detection and identification method |
| CN106156793A (en) * | 2016-06-27 | 2016-11-23 | 西北工业大学 | Extract in conjunction with further feature and the classification method of medical image of shallow-layer feature extraction |
| CN106210471A (en) * | 2016-07-19 | 2016-12-07 | 成都百威讯科技有限责任公司 | A kind of outdoor face recognition method and system |
| CN106803252A (en) * | 2017-01-16 | 2017-06-06 | 广东容祺智能科技有限公司 | A kind of dirty positioning of power transmission line column number plate and automatic testing method |
Non-Patent Citations (3)
| Title |
|---|
| 武洪强: ""基于支持向量机的路面状态视频图像识别技术研究"", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 * |
| 虞益诚主编: "《电子商务概论》", 30 June 2006 * |
| 韩敏: "综合contourlet变换和颜色直方图的图像检索技术", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114241777A (en) * | 2022-01-06 | 2022-03-25 | 交通运输部路网监测与应急处置中心 | Multi-source heterogeneous networking road condition monitoring early warning system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ying et al. | A new image contrast enhancement algorithm using exposure fusion framework | |
| Xu et al. | Wheat ear counting using K-means clustering segmentation and convolutional neural network | |
| Dev et al. | Color-based segmentation of sky/cloud images from ground-based cameras | |
| CN104596929B (en) | Determine the method and apparatus of air quality | |
| CN110263717B (en) | Method for determining land utilization category of street view image | |
| CN104881865A (en) | Forest disease and pest monitoring and early warning method and system based on unmanned plane image analysis | |
| CN106777177A (en) | Search method and device | |
| US9305208B2 (en) | System and method for recognizing offensive images | |
| CN108234882B (en) | Image blurring method and mobile terminal | |
| CN110084782B (en) | Full-reference image quality evaluation method based on image significance detection | |
| CN104268590B (en) | The blind image quality evaluating method returned based on complementary combination feature and multiphase | |
| JP2021531571A (en) | Certificate image extraction method and terminal equipment | |
| CN103366178A (en) | Method and device for carrying out color classification on target image | |
| CN109376256A (en) | Image search method and device | |
| CN108198130A (en) | Image processing method, device, storage medium and electronic equipment | |
| CN111986785A (en) | Medical image labeling method and device, equipment and storage medium | |
| CN107943811A (en) | The dissemination method and device of content | |
| CN108846330A (en) | A kind of paper calligraphy and painting micro-image intelligent identifying system and recognition methods | |
| CN114863299A (en) | Fine identification system for aerial image target | |
| CN114374760A (en) | Image testing method and device, computer equipment and computer readable storage medium | |
| CN111256315A (en) | Neural network-based air conditioner control model creation method, control method and air conditioner | |
| CN108288063A (en) | The meteorology on road surface determines method, apparatus and system | |
| CN109784357A (en) | A kind of image based on statistical model retakes detection method | |
| CN110135274B (en) | Face recognition-based people flow statistics method | |
| CN110210401B (en) | Intelligent target detection method under weak light |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180717 |
|
| RJ01 | Rejection of invention patent application after publication |