CN109919002A - The recognition methods of yellow stop forbidden line, device, computer equipment and storage medium - Google Patents
The recognition methods of yellow stop forbidden line, device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN109919002A CN109919002A CN201910062723.7A CN201910062723A CN109919002A CN 109919002 A CN109919002 A CN 109919002A CN 201910062723 A CN201910062723 A CN 201910062723A CN 109919002 A CN109919002 A CN 109919002A
- Authority
- CN
- China
- Prior art keywords
- foreground area
- yellow
- line
- stop forbidden
- image
- 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
- 238000000034 method Methods 0.000 title claims abstract description 47
- 238000003860 storage Methods 0.000 title claims abstract description 14
- 238000001514 detection method Methods 0.000 claims abstract description 120
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 55
- 238000012360 testing method Methods 0.000 claims abstract description 14
- 238000012549 training Methods 0.000 claims description 48
- 238000004590 computer program Methods 0.000 claims description 13
- 238000006243 chemical reaction Methods 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 5
- 238000000151 deposition Methods 0.000 claims description 3
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 7
- 230000002411 adverse Effects 0.000 description 5
- 230000007613 environmental effect Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000001537 neural effect Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000004044 response Effects 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005260 corrosion Methods 0.000 description 2
- 230000007797 corrosion Effects 0.000 description 2
- 238000005520 cutting process Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000001149 cognitive effect Effects 0.000 description 1
- 230000010485 coping Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
Classifications
-
- 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
-
- 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
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/24—Aligning, centring, orientation detection or correction of the image
-
- 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)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of yellow stop forbidden line recognition methods, device, computer equipment and storage mediums, are related to artificial intelligence field.The yellow stop forbidden line recognition methods includes: acquisition image to be classified, and image to be classified is related to floor line；Foreground area is extracted from image to be classified using yolo object detection frame, wherein, the testing result of yolo object detection frame indicates that it is foreground area that classification, which is the candidate frame of yellow stop forbidden line, in testing result, and yellow stop forbidden line is one of floor line in the form of candidate frame；The yellow stop forbidden line in foreground area is identified based on color space, obtains the yellow stop forbidden line in foreground area；It is identified using yellow stop forbidden line of the convolutional neural networks model to foreground area non-in image to be classified, obtains the yellow stop forbidden line in non-foreground area.It can be accurately identified by realization yellow stop forbidden line near real-time under external environment influence using the yellow stop forbidden line recognition methods.
Description
[technical field]
The present invention relates to artificial intelligence field more particularly to a kind of yellow stop forbidden line recognition methods, device, computer equipment
And storage medium.
[background technique]
Yellow stop forbidden line refers to the parking restrictions indicated using yellow, belongs to one kind of floor line, including curb
Yellow stop forbidden line and the yellow stop forbidden line indicated in a grid formation on road surface.Currently, in the unfavorable ring such as illumination, ponding and corrosion
Duration needed for the identification of yellow stop forbidden line is longer under the influence of the factor of border, and the yellow for being unable to satisfy near real-time in real life is prohibited
Stop line detection demand.
[summary of the invention]
In view of this, the embodiment of the invention provides a kind of yellow stop forbidden line recognition methods, device, computer equipment and depositing
Storage media, to solve the problems, such as that being unable to satisfy yellow stop forbidden line near real-time under by external environment influence accurately identifies.
In a first aspect, the embodiment of the invention provides a kind of yellow stop forbidden line recognition methods, comprising:
Image to be classified is obtained, the image to be classified is related to floor line；
Foreground area is extracted from the image to be classified using yolo object detection frame, wherein the yolo object
The testing result of detection framework indicates that classification is that the candidate frame of yellow stop forbidden line is in testing result in the form of candidate frame
The foreground area, the yellow stop forbidden line are one of described floor line；
The yellow stop forbidden line in the foreground area is identified based on color space, is obtained in the foreground area
Yellow stop forbidden line；
It is identified using yellow stop forbidden line of the convolutional neural networks model to non-foreground area in the image to be classified,
Obtain the yellow stop forbidden line in the non-foreground area, wherein the non-foreground area refers to described in the image to be classified
Image-region other than foreground area.
The aspect and any possible implementation manners as described above, it is further provided a kind of implementation, the use
Yolo object detection frame extracts foreground area from the image to be classified, comprising:
Set the yolo object detection frame to the detection pattern of the single detection yellow stop forbidden line；
Detection zone is calculated from the image to be classified using the yolo object detection frame of the detection pattern
Objective degrees of confidence, wherein the detection zone is the image fritter divided in advance of the image to be classified, each institute
It states image fritter and represents a detection zone；
The objective degrees of confidence is compared with preset confidence threshold value, is higher than according to the objective degrees of confidence described
The detection zone of confidence threshold value obtains the foreground area.
The aspect and any possible implementation manners as described above, it is further provided a kind of implementation, it is described to be based on face
The colour space identifies the yellow stop forbidden line in the foreground area, obtains the yellow stop forbidden line in the foreground area, packet
It includes:
The conversion that the foreground area is carried out to hsv color space, determines the color space where the foreground area；
Color space where judging the foreground area whether there is color of object, and if it exists, then be based on the prospect
The color of object present in region carries out straight line fitting in the foreground area using least square method, wherein the mesh
Mark color is yellow；
The yellow stop forbidden line in the foreground area is obtained according to the result of straight line fitting.
The aspect and any possible implementation manners as described above, it is further provided a kind of implementation, in the use
Before the step of convolutional neural networks model identifies the yellow stop forbidden line of non-foreground area in the image to be classified, institute
State method further include:
Training sample is obtained, the training sample includes the training picture of yellow stop forbidden line；
Initialize convolutional neural networks；
The training sample is input in the convolutional neural networks after initialization and is trained, the convolutional Neural is obtained
Network model, the convolutional neural networks model yellow stop forbidden line for identification.
The aspect and any possible implementation manners as described above, it is further provided a kind of implementation, it is described using volume
Product neural network model identifies the yellow stop forbidden line of non-foreground area in the image to be classified, comprising:
Using the feature vector of non-foreground area described in the convolutional neural networks model extraction；
Based on described eigenvector, the classification that yellow stop forbidden line is calculated in the convolutional neural networks model is general
Rate；
The non-foreground area that the class probability of the yellow stop forbidden line is greater than default classification thresholds is determined as the yellow
Stop forbidden line.
Second aspect, the embodiment of the invention provides a kind of yellow stop forbidden line identification devices, comprising:
Image to be classified obtains module, and for obtaining image to be classified, the image to be classified is related to floor line；
Foreground area obtains module, for extracting foreground zone from the image to be classified using yolo object detection frame
Domain, wherein the testing result of the yolo object detection frame indicates that classification is yellow in testing result in the form of candidate frame
The candidate frame of stop forbidden line is the foreground area, and the yellow stop forbidden line is one of described floor line；
Foreground area yellow stop forbidden line obtains module, for being stopped based on color space to the yellow taboo in the foreground area
Line is identified, the yellow stop forbidden line in the foreground area is obtained；
Non- foreground area yellow stop forbidden line obtains module, for using convolutional neural networks model to the image to be classified
In the yellow stop forbidden line of non-foreground area identified, obtain the yellow stop forbidden line in the non-foreground area, wherein described non-
Foreground area refers to the image-region other than foreground area described in the image to be classified.
The third aspect, a kind of computer equipment, including memory, processor and storage are in the memory and can be
The computer program run on the processor, the processor realize above-mentioned yellow stop forbidden line when executing the computer program
The step of recognition methods.
Fourth aspect, the embodiment of the invention provides a kind of computer readable storage mediums, comprising: computer program, when
When the computer program is executed by the processor, to execute first aspect described in any item yellow stop forbidden lines identification side
Method.
In embodiments of the present invention, image to be classified is obtained first, from image to be classified relevant to floor line
Identify yellow stop forbidden line；Then foreground area is extracted from image to be classified using yolo object detection frame, obtains doubtful packet
The candidate frame of yellow stop forbidden line is included, can quickly and accurately determine has greater probability in image to be classified include yellow stop forbidden line
Foreground area；It is then based on color space to identify the yellow stop forbidden line in foreground area, obtain in foreground area
Yellow stop forbidden line can rapidly determine a part of yellow stop forbidden line in image to be classified；Finally, using convolutional neural networks
Model identifies that the yellow obtained in the non-foreground area is prohibited to the yellow stop forbidden line of foreground area non-in image to be classified
Stop line, can to it is first using yolo object detection frame examine the yellow stop forbidden line not measured also identify that come.The present invention is implemented
Example can be realized the near real-time identification of yellow stop forbidden line under the premise of guaranteeing recognition accuracy.
[Detailed description of the invention]
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below will be to needed in the embodiment attached
Figure is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the invention, for this field
For those of ordinary skill, without any creative labor, it can also be obtained according to these attached drawings other attached
Figure.
Fig. 1 is the flow chart based on yellow stop forbidden line recognition methods in one embodiment of the invention；
Fig. 2 is the schematic diagram based on yellow stop forbidden line identification device in one embodiment of the invention；
Fig. 3 is a schematic diagram of computer equipment in one embodiment of the invention.
[specific embodiment]
For a better understanding of the technical solution of the present invention, being retouched in detail to the embodiment of the present invention with reference to the accompanying drawing
It states.
It will be appreciated that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Base
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts it is all its
Its embodiment, shall fall within the protection scope of the present invention.
The term used in embodiments of the present invention is only to be not intended to be limiting merely for for the purpose of describing particular embodiments
The present invention.In the embodiment of the present invention and the "an" of singular used in the attached claims, " described " and "the"
It is also intended to including most forms, unless the context clearly indicates other meaning.
It should be appreciated that term "and/or" used herein is only a kind of identical field for describing affiliated partner, table
Show there may be three kinds of relationships, for example, A and/or B, can indicate: individualism A exists simultaneously A and B, individualism B this three
Kind situation.In addition, character "/" herein, typicallys represent the relationship that forward-backward correlation object is a kind of "or".
It will be appreciated that though default model may be described using term first, second, third, etc. in embodiments of the present invention
It encloses, but these preset ranges should not necessarily be limited by these terms.These terms are only used to for preset range being distinguished from each other out.For example,
In the case where not departing from range of embodiment of the invention, the first preset range can also be referred to as the second preset range, similarly,
Second preset range can also be referred to as the first preset range.
Depending on context, word as used in this " if " can be construed to " ... when " or " when ...
When " or " in response to determination " or " in response to detection ".Similarly, depend on context, phrase " if it is determined that " or " if detection
(condition or event of statement) " can be construed to " when determining " or " in response to determination " or " when the detection (condition of statement
Or event) when " or " in response to detection (condition or event of statement) ".
Fig. 1 shows a flow chart of yellow stop forbidden line recognition methods in the present embodiment.The yellow stop forbidden line recognition methods can
Apply in yellow stop forbidden line identifying system, from pavement strip identify yellow stop forbidden line can be used the yellow stop forbidden line identification
System is identified.The yellow stop forbidden line identifying system specifically can be applicable in computer equipment, wherein the computer equipment is
The equipment that can carry out human-computer interaction with user, the including but not limited to equipment such as computer, smart phone and plate.As shown in Figure 1, should
Yellow stop forbidden line recognition methods includes the following steps:
S10: obtaining image to be classified, and image to be classified is related to floor line.
It is to be appreciated that vehicle-mounted yellow stop forbidden line identifying system will be real-time by picture pick-up device in vehicle travel process
Image relevant to floor line is obtained, which is image to be classified.
In one embodiment, yellow stop forbidden line identifying system obtains image to be classified, yellow to occur in image to be classified
When color stop forbidden line, yellow stop forbidden line can be quickly identified, and make preset reaction.
S20: foreground area is extracted from image to be classified using yolo object detection frame, wherein yolo object detection
The testing result of frame indicates that it is foreground zone that classification, which is the candidate frame of yellow stop forbidden line, in testing result in the form of candidate frame
Domain, yellow stop forbidden line are one of floor line.
Wherein, yolo (You Only Look Once) object detection frame is that one kind can detect object, and
The model classified to object.Candidate frame indicates that doubtful there are the mark frames of yellow stop forbidden line.
In one embodiment, foreground area is extracted from image to be classified using yolo object detection frame.The yolo object
Body detection framework be it is pre- first pass through training obtain, wherein training the yolo object detection frame training sample include various
Different types of floor line, yellow stop forbidden line are one such training samples.Particularly, yellow stop forbidden line and other ground
The ratio of graticule training sample can all be 1:1, and yolo object detection can be effectively prevented using the training sample of the equal proportion
The over-fitting of frame improves the Detection accuracy and classification accuracy of yolo object detection frame.
It is to be appreciated that yolo object detection frame will be determined quickly when occurring yellow stop forbidden line in image to be classified
There is greater probability to belong to the region of yellow stop forbidden line in image to be classified, the region namely foreground area.
Further, in step S20, foreground area is extracted from image to be classified using yolo object detection frame,
It specifically includes:
S21: yolo object detection frame is set to the detection pattern of single detection yellow stop forbidden line.
It is to be appreciated that yolo object detection frame is classified for detection object and by object under usual scene
, it is contemplated that the embodiment of the present invention need to only detect yellow stop forbidden line, and the inspection of the floor line in addition to yellow stop forbidden line
Survey and classification be it is unnecessary, therefore, yolo object detection frame can be set to it is single detection yellow stop forbidden line detection
Mode, wherein the setting is realized by yellow stop forbidden line identifying system.
Specifically, in training yolo object detection frame, training sample is divided into yellow stop forbidden line training sample and non-
Yellow stop forbidden line training sample, wherein non-yellow stop forbidden line training sample may include the arbitrary sample except yellow stop forbidden line, no
It is limited only to floor line, in addition, equally by the ratio of yellow stop forbidden line training sample and non-yellow stop forbidden line training sample
It is set as 1:1, over-fitting occurs with the yolo object detection frame for preventing training from obtaining.Trained using the training sample
After the yolo object detection frame arrived, which will be stored in yellow stop forbidden line identification system with document form
In system, system can call file to set the single detection mould for detecting yellow stop forbidden line for yolo object detection frame at any time
Formula.
In one embodiment, yolo object detection frame is set to the detection pattern of single detection yellow stop forbidden line, energy
It is enough to be only concerned yellow stop forbidden line when detecting, other non-yellow stop forbidden line parts are classified as one kind, aid in determining whether that yellow taboo is stopped
Line further increases treatment effeciency in the position of image to be classified.
S22: it is set using the target that the yolo object detection frame of detection pattern calculates detection zone from image to be classified
Reliability, wherein detection zone is the image fritter that image to be classified is divided in advance, each image fritter represents a detection
Region.
Wherein, the yellow stop forbidden line that objective degrees of confidence shows has certain probability to fall in the degree in detection zone.
In one embodiment, image to be classified is cut before the objective degrees of confidence for calculating detection zone first
It cuts, specifically, when image to be classified is square, image to be classified can be divided into the image fritter of s*s, each image is small
Block block corresponds to a detection zone；When image to be classified is rectangle, the a*b image fritter conduct of sizes such as equally it is segmented into
Detection zone.Under normal circumstances, the cutting of image to be classified by using etc. sizes cutting by the way of obtain detection zone, Ke Yi
The connection of edge feature between different detection zones is improved when detection, improving in subsequent obtained foreground area includes that yellow taboo is stopped
The probability of line.
In one embodiment, after determining detection zone, using it is single detection yellow stop forbidden line detection pattern calculate to
The objective degrees of confidence of each detection zone in classification image, to determine foreground area according to objective degrees of confidence.
S23: objective degrees of confidence is compared with preset confidence threshold value, is higher than confidence level threshold according to objective degrees of confidence
The detection zone of value obtains foreground area.
In step S21-S23, aid in determining whether that yellow stop forbidden line exists using the detection pattern of single detection yellow stop forbidden line
The position of image to be classified calculates the mesh of detection zone using the yolo object detection frame of detection pattern from image to be classified
Confidence level is marked, and objective degrees of confidence is compared to obtain foreground area with preset confidence threshold value, can be improved foreground zone
It include yellow stop forbidden line probability in domain, so that just the major part in image to be classified when using yolo object detection frame
Yellow stop forbidden line identifies out.
S30: the yellow stop forbidden line in foreground area is identified based on color space, obtains the yellow in foreground area
Stop forbidden line.
Wherein, it is in certain standards that color space, which is also referred to as color model (also known as color space or color system) its purposes,
It is lower that colour is illustrated with generally acceptable mode.
In one embodiment, due to the objects such as yellow stop forbidden line and other floor lines in color there are difference, can be with
The yellow stop forbidden line in foreground area is identified using color space, and obtains yellow stop forbidden line from foreground area.It adopts
It can be distinguished by the feature of yellow stop forbidden line itself with other objects with color space, it can quickly and accurately in the past
Yellow stop forbidden line is obtained in scene area.
Further, in step s 30, the yellow stop forbidden line in foreground area is identified based on color space, is obtained
Yellow stop forbidden line into foreground area, specifically includes:
S31: foreground area is carried out to the conversion in hsv color space, determines the color space where foreground area.
In one embodiment, color space can specifically use hsv color space, which compares other face
Effect of the colour space on identification yellow stop forbidden line is more preferable, and accuracy rate is higher.
S32: the color space where judging foreground area whether there is color of object, and if it exists, then be based on foreground area
Present in color of object, using least square method foreground area carry out straight line fitting, wherein color of object is yellow.
It is to be appreciated that yellow stop forbidden line also needs to meet straight line other than yellow stop forbidden line is about the feature in color
Relationship excludes non-directional tool to further determine the yellow stop forbidden line in foreground area according to the requirement of the straight line
The object of standby yellow characteristic.
In one embodiment, the yellow present in the foreground area, then using least square method to existing in foreground area
Yellow carry out straight line fitting specifically can establish rectangular coordinate system, using the pixel of image as minimum unit, using seat
Target mode indicates each position for representing yellow pixel in foreground area, further according to yellow pixel coordinate is represented, using minimum
Square law carries out straight line fitting, and then identifies the yellow stop forbidden line in foreground area.The present embodiment is by considering yellow stop forbidden line
Feature in color and in shape carries out comprehensive descision, can effectively improve yellow stop forbidden line in foreground area identify it is accurate
Rate.
S33: the yellow stop forbidden line in foreground area is obtained according to the result of straight line fitting.
In one embodiment, when straight line fitting result is very, then to obtain the yellow stop forbidden line in foreground area.
In step S31-S33, the feature of yellow stop forbidden line is comprehensively considered, the method using color space and straight line fitting exists
In foreground area identification obtain yellow stop forbidden line, using this method identification foreground area in yellow stop forbidden line accuracy rate compared with
It is high.
S40: being identified using yellow stop forbidden line of the convolutional neural networks model to foreground area non-in image to be classified,
Obtain the yellow stop forbidden line in non-foreground area, wherein non-foreground area refers to the figure in image to be classified other than foreground area
As region.
It is to be appreciated that foreground area is the biggish region of yellow stop forbidden line probability of occurrence in image to be classified, not
Representing non-foreground area does not have yellow stop forbidden line.Be illuminated by the light, the adverse environmental factors such as ponding and corrosion under the influence of, non-prospect
Accounting of the yellow stop forbidden line in image to be classified will increase with the influence degree of adverse environmental factors in region, simply
It says, is exactly under the influence of adverse environmental factors, the detection effect of yolo object detection model can relative reduction.
In one embodiment, it can specifically be carried out using yellow stop forbidden line of the convolutional neural networks model to non-foreground area
Identification, obtains the yellow stop forbidden line in non-foreground area.Convolutional neural networks are a kind of deep neural networks, can extract yellow
The further feature of stop forbidden line, even if still maintaining higher recognition accuracy under the influence of adverse environmental factors.Convolutional Neural
For network compared with yolo neural network (neural network of training yolo object detection model), the speed of convolutional neural networks can phase
To slow.
Further, before step S40, further includes:
S411: training sample is obtained, training sample includes the training picture of yellow stop forbidden line.
In one embodiment, using include yellow stop forbidden line training picture as training sample, to allow convolutional neural networks
It can learn the further feature of yellow stop forbidden line, and sufficiently be distinguished with the feature of other training samples.
S412: initialization convolutional neural networks.
Wherein, convolutional neural networks include network parameter, and network parameter includes weight and biasing.In one embodiment, it enables
The weight of convolutional neural networks initialization meets formulanlIndicate the training sample inputted at l layers
Number of samples, S () indicate variance operation, WlIndicate l layers of weight,Indicate any, l is indicated in convolutional neural networks
L layers, the efficiency of convolutional neural networks model training can be improved in such a way that this is initial, help to improve convolutional Neural net
The recognition accuracy of network model.The limitation that the step of initialization convolutional neural networks and step S411 are not executed successively, can
Can also be executed before after step S411.
S413: training sample is input in the convolutional neural networks after initialization and is trained, convolutional Neural net is obtained
Network model, convolutional neural networks model yellow stop forbidden line for identification.
In one embodiment, after obtaining required training sample and initialization convolutional neural networks, training sample is defeated
Enter and be trained in the convolutional neural networks to after initialization, the network in convolutional neural networks can be joined according to training sample
Number is iterated update, and the result for exporting training sample in convolutional neural networks reaches expectation to the extent permitted by the error
As a result, obtaining the convolutional neural networks model of yellow stop forbidden line for identification.
Step S411-S413 provides a kind of training convolutional neural networks model method, can accelerate mould using this method
The process of type training, and obtain the higher convolutional neural networks model of discrimination.
Further, in step s 40, using convolutional neural networks model to the Huang of foreground area non-in image to be classified
Color stop forbidden line is identified, is specifically included:
S421: using the feature vector of the non-foreground area of convolutional neural networks model extraction.
S422: it is based on feature vector, the class probability of yellow stop forbidden line is calculated in convolutional neural networks model.
S423: the non-foreground area that the class probability of yellow stop forbidden line is greater than default classification thresholds is determined as yellow taboo and is stopped
Line.
In step S421-S423, it is possible to understand that ground, convolutional neural networks model the feature for extracting non-foreground area to
After amount, needs to be compared with the further feature of the training sample extracted when training, judge in the feature for extracting non-foreground area
Which kind of in training sample vector belong to, and when the class probability of yellow stop forbidden line is greater than default classification thresholds, then can determine
Non- foreground area corresponding to the feature vector of the extraction is yellow stop forbidden line.In the present embodiment, it provides a kind of determining yellow
The method of color stop forbidden line can effectively determine which non-foreground area is yellow stop forbidden line.
It is to be appreciated that first being prohibited in the present invention is implemented using the yellow in yolo object detection frame identification foreground area
Stop line, in conjunction with the fireballing advantage of detection of yolo object detection frame, so that most yellow taboo is stopped in images to be recognized
Line is just identified at this stage, for part because adverse environmental factors influence without what is identified, only be included in
In images to be recognized in the lesser non-foreground area of accounting, though at this time can be slow using the speed that convolution identifies, due to
Non- foreground area accounts for smaller, therefore will not consume many times in identification, and can be recognized accurately in non-foreground area
Yellow stop forbidden line.By combining two cognitive phases, reaches not only quick but also accurately identify yellow in images to be recognized and prohibit
Stop the effect of line.
In embodiments of the present invention, image to be classified is obtained first, from image to be classified relevant to floor line
Identify yellow stop forbidden line；Then foreground area is extracted from image to be classified using yolo object detection frame, obtains doubtful packet
The candidate frame of yellow stop forbidden line is included, can quickly and accurately determine has greater probability in image to be classified include yellow stop forbidden line
Foreground area；It is then based on color space to identify the yellow stop forbidden line in foreground area, obtain in foreground area
Yellow stop forbidden line can rapidly determine a part of yellow stop forbidden line in image to be classified；Finally, using convolutional neural networks
Model identifies that the yellow taboo obtained in non-foreground area is stopped to the yellow stop forbidden line of foreground area non-in image to be classified
Line, can to it is first using yolo object detection frame examine the yellow stop forbidden line not measured also identify that come.The embodiment of the present invention
Under the premise of guaranteeing recognition accuracy, the near real-time identification of yellow stop forbidden line can be realized.
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.
Based on yellow stop forbidden line recognition methods provided in embodiment, it is above-mentioned that the embodiment of the present invention further provides realization
The Installation practice of each step and method in embodiment of the method.
Fig. 2 shows the principles with the one-to-one yellow stop forbidden line identification device of yellow stop forbidden line recognition methods in embodiment
Block diagram.As shown in Fig. 2, the yellow stop forbidden line identification device includes that image to be classified obtains module 10, foreground area obtains module
20, foreground area yellow stop forbidden line obtains module 30 and non-foreground area yellow stop forbidden line obtains module 40.Wherein, figure to be sorted
As obtaining module 10, foreground area obtains module 20, foreground area yellow stop forbidden line obtains module 30 and non-foreground area yellow
The realization function step corresponding with yellow stop forbidden line recognition methods in embodiment that stop forbidden line obtains module 40 corresponds, to keep away
Exempt to repeat, the present embodiment is not described in detail one by one.
Image to be classified obtains module 10, and for obtaining image to be classified, image to be classified is related to floor line.
Foreground area obtains module 20, for extracting foreground zone from image to be classified using yolo object detection frame
Domain, wherein the testing result of yolo object detection frame indicates that classification is that yellow taboo is stopped in testing result in the form of candidate frame
The candidate frame of line is foreground area, and yellow stop forbidden line is one of floor line.
Foreground area yellow stop forbidden line obtain module 30, for based on color space to the yellow stop forbidden line in foreground area
It is identified, obtains the yellow stop forbidden line in foreground area.
Non- foreground area yellow stop forbidden line obtains module 40, for using convolutional neural networks model in image to be classified
The yellow stop forbidden line of non-foreground area is identified, the yellow stop forbidden line in non-foreground area is obtained, wherein non-foreground area is
Refer to the image-region in image to be classified other than foreground area.
Optionally, it includes detection pattern conversion unit, objective degrees of confidence computing unit and preceding that foreground area, which obtains module 20,
Scene area acquiring unit.
Detection pattern conversion unit, for setting yolo object detection frame to the detection of single detection yellow stop forbidden line
Mode.
Objective degrees of confidence computing unit, for being fallen into a trap using the yolo object detection frame of detection pattern from image to be classified
Calculate the objective degrees of confidence of detection zone, wherein detection zone is the image fritter that image to be classified is divided in advance, each
Image fritter represents a detection zone.
Foreground area acquiring unit, for objective degrees of confidence to be compared with preset confidence threshold value, according to target
The detection zone that confidence level is higher than confidence threshold value obtains foreground area.
Optionally, it includes color space determination unit, color of object judgement that foreground area yellow stop forbidden line, which obtains module 30,
Unit and foreground area yellow stop forbidden line acquiring unit.
Color space determination unit determines foreground area place for foreground area to be carried out to the conversion in hsv color space
Color space.
Color of object judging unit whether there is color of object for the color space where judging foreground area, if depositing
Then based on color of object present in foreground area, straight line fitting is being carried out in foreground area using least square method, wherein
Color of object is yellow.
Foreground area yellow stop forbidden line acquiring unit, the yellow in foreground area is obtained for the result according to straight line fitting
Stop forbidden line.
Optionally, yellow stop forbidden line identification device further includes training sample acquiring unit, initialization unit and convolutional Neural
Network model acquiring unit.
Training sample acquiring unit, for obtaining training sample, training sample includes the training picture of yellow stop forbidden line.
Initialization unit, for initializing convolutional neural networks.
Convolutional neural networks model acquiring unit, for training sample to be input in the convolutional neural networks after initialization
It is trained, obtains convolutional neural networks model, convolutional neural networks model yellow stop forbidden line for identification.
Optionally, non-foreground area yellow stop forbidden line obtain module 40 include non-foreground area characteristic vector pickup unit,
Class probability computing unit and non-foreground area yellow stop forbidden line acquiring unit.
Non- foreground area characteristic vector pickup unit, for the spy using the non-foreground area of convolutional neural networks model extraction
Levy vector.
Yellow taboo is calculated for being based on feature vector in class probability computing unit in convolutional neural networks model
Stop the class probability of line.
Non- foreground area yellow stop forbidden line acquiring unit, for the class probability of yellow stop forbidden line to be greater than default classification threshold
The non-foreground area of value is determined as yellow stop forbidden line.
In embodiments of the present invention, image to be classified is obtained first, from image to be classified relevant to floor line
Identify yellow stop forbidden line；Then foreground area is extracted from image to be classified using yolo object detection frame, obtains doubtful packet
The candidate frame of yellow stop forbidden line is included, can quickly and accurately determine has greater probability in image to be classified include yellow stop forbidden line
Foreground area；It is then based on color space to identify the yellow stop forbidden line in foreground area, obtain in foreground area
Yellow stop forbidden line can rapidly determine a part of yellow stop forbidden line in image to be classified；Finally, using convolutional neural networks
Model identifies that the yellow taboo obtained in non-foreground area is stopped to the yellow stop forbidden line of foreground area non-in image to be classified
Line, can to it is first using yolo object detection frame examine the yellow stop forbidden line not measured also identify that come.The embodiment of the present invention
Under the premise of guaranteeing recognition accuracy, the near real-time identification of yellow stop forbidden line can be realized.
The present embodiment provides a computer readable storage medium, computer journey is stored on the computer readable storage medium
Sequence realizes yellow stop forbidden line recognition methods in embodiment, to avoid repeating, herein not when the computer program is executed by processor
It repeats one by one.Alternatively, realizing in embodiment each mould in yellow stop forbidden line identification device when the computer program is executed by processor
Block/unit function does not repeat one by one herein to avoid repeating.
Fig. 3 is the schematic diagram for the computer equipment that one embodiment of the invention provides.As shown in figure 3, the calculating of the embodiment
Machine equipment 50 includes: processor 51, memory 52 and is stored in the calculating that can be run in memory 52 and on processor 51
Machine program 53 realizes the yellow stop forbidden line recognition methods in embodiment, to keep away when the computer program 53 is executed by processor 51
Exempt to repeat, not repeat one by one herein.Alternatively, realizing that yellow taboo is stopped in embodiment when the computer program 53 is executed by processor 51
The function of each model/unit does not repeat one by one herein in line identification device to avoid repeating.
Computer equipment 50 can be desktop PC, notebook, palm PC and cloud server etc. and calculate equipment.
Computer equipment 50 may include, but be not limited only to, processor 51, memory 52.It will be understood by those skilled in the art that Fig. 3 is only
It is only the example of computer equipment 50, does not constitute the restriction to computer equipment 50, may include more more or less than illustrating
Component, perhaps combine certain components or different components, for example, computer equipment can also include input-output equipment,
Network access equipment, bus etc..
Alleged processor 51 can be central processing unit (Central Processing Unit, CPU), can also be
Other general processors, digital signal processor (Digital Signal Processor, DSP), specific integrated circuit
(Application Specific Integrated Circuit, ASIC), field 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 is also possible to any conventional processor
Deng.
Memory 52 can be the internal storage unit of computer equipment 50, such as the hard disk or interior of computer equipment 50
It deposits.Memory 52 is also possible to the plug-in type being equipped on the External memory equipment of computer equipment 50, such as computer equipment 50
Hard disk, intelligent memory card (Smart Media Card, SMC), secure digital (Secure Digital, SD) card, flash card
(Flash Card) etc..Further, memory 52 can also both including computer equipment 50 internal storage unit and also including
External memory equipment.Memory 52 is for storing other programs and data needed for computer program and computer equipment.It deposits
Reservoir 52 can be also used for temporarily storing the data that has exported or will export.
It is apparent to those skilled in the art that for convenience of description and succinctly, only with above-mentioned each function
Can unit, module division progress for example, in practical application, can according to need and by above-mentioned function distribution by different
Functional unit, module are completed, i.e., the internal structure of device are divided into different functional unit or module, to complete above description
All or part of function.
The above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although with reference to the foregoing embodiments
Invention is explained in detail, those skilled in the art should understand that: it still can be to aforementioned each implementation
Technical solution documented by example is modified or equivalent replacement of some of the technical features；And these modification or
Replacement, the spirit and scope for technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution should all include
Within protection scope of the present invention.
Claims (10)
1. a kind of yellow stop forbidden line recognition methods, which is characterized in that the described method includes:
Image to be classified is obtained, the image to be classified is related to floor line；
Foreground area is extracted from the image to be classified using yolo object detection frame, wherein the yolo object detection
The testing result of frame indicates that it is described that classification, which is the candidate frame of yellow stop forbidden line, in testing result in the form of candidate frame
Foreground area, the yellow stop forbidden line are one of described floor line；
The yellow stop forbidden line in the foreground area is identified based on color space, obtains the yellow in the foreground area
Stop forbidden line；
It is identified, is obtained using yellow stop forbidden line of the convolutional neural networks model to non-foreground area in the image to be classified
Yellow stop forbidden line in the non-foreground area, wherein the non-foreground area refers to prospect described in the image to be classified
Image-region other than region.
2. the method according to claim 1, wherein it is described using yolo object detection frame from described to be sorted
Foreground area is extracted in image, comprising:
Set the yolo object detection frame to the detection pattern of the single detection yellow stop forbidden line；
The mesh of detection zone is calculated from the image to be classified using the yolo object detection frame of the detection pattern
Mark confidence level, wherein the detection zone is the image fritter that the image to be classified is divided in advance, each described figure
As fritter represents a detection zone；
The objective degrees of confidence is compared with preset confidence threshold value, the confidence is higher than according to the objective degrees of confidence
The detection zone of degree threshold value obtains the foreground area.
3. the method according to claim 1, wherein it is described based on color space to the Huang in the foreground area
Color stop forbidden line is identified, the yellow stop forbidden line in the foreground area is obtained, comprising:
The conversion that the foreground area is carried out to hsv color space, determines the color space where the foreground area；
Color space where judging the foreground area whether there is color of object, and if it exists, then be based on the foreground area
Present in the color of object, using least square method the foreground area carry out straight line fitting, wherein the target face
Color is yellow；
The yellow stop forbidden line in the foreground area is obtained according to the result of straight line fitting.
4. the method according to claim 1, wherein using convolutional neural networks model to described wait divide described
Before the step of yellow stop forbidden line of non-foreground area is identified in class image, the method also includes:
Training sample is obtained, the training sample includes the training picture of yellow stop forbidden line；
Initialize convolutional neural networks；
The training sample is input in the convolutional neural networks after initialization and is trained, the convolutional neural networks are obtained
Model, the convolutional neural networks model yellow stop forbidden line for identification.
5. the method according to claim 1, which is characterized in that described to use convolutional neural networks model
The yellow stop forbidden line of non-foreground area in the image to be classified is identified, comprising:
Using the feature vector of non-foreground area described in the convolutional neural networks model extraction；
Based on described eigenvector, the class probability of yellow stop forbidden line is calculated in the convolutional neural networks model；
The non-foreground area that the class probability of the yellow stop forbidden line is greater than default classification thresholds is determined as the yellow taboo to stop
Line.
6. a kind of yellow stop forbidden line identification device, which is characterized in that described device includes:
Image to be classified obtains module, and for obtaining image to be classified, the image to be classified is related to floor line；
Foreground area obtains module, for extracting foreground area from the image to be classified using yolo object detection frame,
Wherein, the testing result of the yolo object detection frame is indicated in the form of candidate frame, and classification is yellow taboo in testing result
The candidate frame for stopping line is the foreground area, and the yellow stop forbidden line is one of described floor line；
Foreground area yellow stop forbidden line obtain module, for based on color space to the yellow stop forbidden line in the foreground area into
Row identification, obtains the yellow stop forbidden line in the foreground area；
Non- foreground area yellow stop forbidden line obtains module, for using convolutional neural networks model to non-in the image to be classified
The yellow stop forbidden line of foreground area is identified, the yellow stop forbidden line in the non-foreground area is obtained, wherein the non-prospect
Region refers to the image-region other than foreground area described in the image to be classified.
7. device according to claim 6, which is characterized in that the foreground area obtains module, comprising:
Detection pattern conversion unit, for setting the single detection yellow stop forbidden line for the yolo object detection frame
Detection pattern；
Objective degrees of confidence computing unit, for the yolo object detection frame using the detection pattern from described to be sorted
The objective degrees of confidence of detection zone is calculated in image, wherein the detection zone is divided in advance for the image to be classified to be obtained
Image fritter, each described image fritter represents a detection zone；
Foreground area acquiring unit, for the objective degrees of confidence to be compared with preset confidence threshold value, according to described
The detection zone that objective degrees of confidence is higher than the confidence threshold value obtains the foreground area.
8. device according to claim 6, which is characterized in that the foreground area yellow stop forbidden line obtains module, comprising:
Color space determination unit determines the foreground area for the foreground area to be carried out to the conversion in hsv color space
The color space at place；
Color of object judging unit whether there is color of object for the color space where judging the foreground area, if depositing
Then based on the color of object present in the foreground area, carried out directly using least square method in the foreground area
Line fitting, wherein the color of object is yellow；
Foreground area yellow stop forbidden line acquiring unit, obtains the yellow in the foreground area for the result according to straight line fitting
Stop forbidden line.
9. a kind of computer equipment, including memory, processor and storage are in the memory and can be in the processor
The computer program of upper operation, which is characterized in that the processor realized when executing the computer program as claim 1 to
The step of any one of 5 yellow stop forbidden line recognition methods.
10. a kind of computer readable storage medium, the computer-readable recording medium storage has computer program, and feature exists
In realization yellow stop forbidden line recognition methods as described in any one of claim 1 to 5 when the computer program is executed by processor
The step of.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910062723.7A CN109919002B (en) | 2019-01-23 | 2019-01-23 | Yellow stop line identification method and device, computer equipment and storage medium |
| PCT/CN2019/115947 WO2020151299A1 (en) | 2019-01-23 | 2019-11-06 | Yellow no-parking line identification method and apparatus, computer device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910062723.7A CN109919002B (en) | 2019-01-23 | 2019-01-23 | Yellow stop line identification method and device, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109919002A true CN109919002A (en) | 2019-06-21 |
| CN109919002B CN109919002B (en) | 2024-02-27 |
Family
ID=66960527
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910062723.7A Active CN109919002B (en) | 2019-01-23 | 2019-01-23 | Yellow stop line identification method and device, computer equipment and storage medium |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN109919002B (en) |
| WO (1) | WO2020151299A1 (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111008672A (en) * | 2019-12-23 | 2020-04-14 | 腾讯科技（深圳）有限公司 | Sample extraction method, sample extraction device, computer-readable storage medium and computer equipment |
| WO2020151299A1 (en) * | 2019-01-23 | 2020-07-30 | 平安科技（深圳）有限公司 | Yellow no-parking line identification method and apparatus, computer device and storage medium |
| CN112418061A (en) * | 2020-11-19 | 2021-02-26 | 城云科技（中国）有限公司 | Method and system for determining vehicle no-parking area |
| WO2021147387A1 (en) * | 2020-01-21 | 2021-07-29 | 上海万物新生环保科技集团有限公司 | Screen scratch and crack detection method and device |
| CN113537248A (en) * | 2021-08-13 | 2021-10-22 | 珠海格力电器股份有限公司 | Image recognition method and device, electronic equipment and storage medium |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112348905B (en) * | 2020-10-30 | 2023-12-19 | 深圳市优必选科技股份有限公司 | Color recognition method and device, terminal equipment and storage medium |
| CN112613344B (en) * | 2020-12-01 | 2024-04-16 | 浙江华锐捷技术有限公司 | Vehicle track occupation detection method, device, computer equipment and readable storage medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103116985A (en) * | 2013-01-21 | 2013-05-22 | 信帧电子技术（北京）有限公司 | Detection method and device of parking against rules |
| CN107122776A (en) * | 2017-04-14 | 2017-09-01 | 重庆邮电大学 | A kind of road traffic sign detection and recognition methods based on convolutional neural networks |
| CN107134144A (en) * | 2017-04-27 | 2017-09-05 | 武汉理工大学 | A kind of vehicle checking method for traffic monitoring |
| CN107358242A (en) * | 2017-07-11 | 2017-11-17 | 浙江宇视科技有限公司 | Target area color identification method, device and monitor terminal |
| CN107480730A (en) * | 2017-09-05 | 2017-12-15 | 广州供电局有限公司 | Power equipment identification model construction method and system, the recognition methods of power equipment |
| CN107633229A (en) * | 2017-09-21 | 2018-01-26 | 北京智芯原动科技有限公司 | Method for detecting human face and device based on convolutional neural networks |
| CN107679508A (en) * | 2017-10-17 | 2018-02-09 | 广州汽车集团股份有限公司 | Road traffic sign detection recognition methods, apparatus and system |
| JP2018092501A (en) * | 2016-12-07 | 2018-06-14 | クラリオン株式会社 | On-vehicle image processing apparatus |
| US20180181819A1 (en) * | 2016-12-22 | 2018-06-28 | Denso Corporation | Demarcation line recognition device |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10242282B2 (en) * | 2017-03-20 | 2019-03-26 | Conduent Business Services, Llc | Video redaction method and system |
| CN108229421B (en) * | 2018-01-24 | 2021-07-02 | 华中科技大学 | Depth video information-based method for detecting falling-off from bed in real time |
| CN109241904B (en) * | 2018-08-31 | 2023-10-20 | 平安科技（深圳）有限公司 | Character recognition model training, character recognition method, device, equipment and medium |
| CN109919002B (en) * | 2019-01-23 | 2024-02-27 | 平安科技（深圳）有限公司 | Yellow stop line identification method and device, computer equipment and storage medium |
-
2019
- 2019-01-23 CN CN201910062723.7A patent/CN109919002B/en active Active
- 2019-11-06 WO PCT/CN2019/115947 patent/WO2020151299A1/en active Application Filing
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103116985A (en) * | 2013-01-21 | 2013-05-22 | 信帧电子技术（北京）有限公司 | Detection method and device of parking against rules |
| JP2018092501A (en) * | 2016-12-07 | 2018-06-14 | クラリオン株式会社 | On-vehicle image processing apparatus |
| US20180181819A1 (en) * | 2016-12-22 | 2018-06-28 | Denso Corporation | Demarcation line recognition device |
| CN107122776A (en) * | 2017-04-14 | 2017-09-01 | 重庆邮电大学 | A kind of road traffic sign detection and recognition methods based on convolutional neural networks |
| CN107134144A (en) * | 2017-04-27 | 2017-09-05 | 武汉理工大学 | A kind of vehicle checking method for traffic monitoring |
| CN107358242A (en) * | 2017-07-11 | 2017-11-17 | 浙江宇视科技有限公司 | Target area color identification method, device and monitor terminal |
| CN107480730A (en) * | 2017-09-05 | 2017-12-15 | 广州供电局有限公司 | Power equipment identification model construction method and system, the recognition methods of power equipment |
| CN107633229A (en) * | 2017-09-21 | 2018-01-26 | 北京智芯原动科技有限公司 | Method for detecting human face and device based on convolutional neural networks |
| CN107679508A (en) * | 2017-10-17 | 2018-02-09 | 广州汽车集团股份有限公司 | Road traffic sign detection recognition methods, apparatus and system |
Non-Patent Citations (1)
| Title |
|---|
| 姜雪宁 等: "颜色和形态学在交通标志识别中的应用研究", 测绘与空间地理信息, vol. 37, no. 12, 25 December 2014 (2014-12-25), pages 17 - 21 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020151299A1 (en) * | 2019-01-23 | 2020-07-30 | 平安科技（深圳）有限公司 | Yellow no-parking line identification method and apparatus, computer device and storage medium |
| CN111008672A (en) * | 2019-12-23 | 2020-04-14 | 腾讯科技（深圳）有限公司 | Sample extraction method, sample extraction device, computer-readable storage medium and computer equipment |
| WO2021147387A1 (en) * | 2020-01-21 | 2021-07-29 | 上海万物新生环保科技集团有限公司 | Screen scratch and crack detection method and device |
| CN112418061A (en) * | 2020-11-19 | 2021-02-26 | 城云科技（中国）有限公司 | Method and system for determining vehicle no-parking area |
| CN112418061B (en) * | 2020-11-19 | 2024-01-23 | 城云科技（中国）有限公司 | Vehicle forbidden region determining method and system |
| CN113537248A (en) * | 2021-08-13 | 2021-10-22 | 珠海格力电器股份有限公司 | Image recognition method and device, electronic equipment and storage medium |
| CN113537248B (en) * | 2021-08-13 | 2024-06-07 | 珠海格力电器股份有限公司 | Image recognition method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109919002B (en) | 2024-02-27 |
| WO2020151299A1 (en) | 2020-07-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109919002A (en) | The recognition methods of yellow stop forbidden line, device, computer equipment and storage medium | |
| CN110020592B (en) | Object detection model training method, device, computer equipment and storage medium | |
| US11455805B2 (en) | Method and apparatus for detecting parking space usage condition, electronic device, and storage medium | |
| CN107123131B (en) | Moving target detection method based on deep learning | |
| US9754160B2 (en) | Method and device for detecting gathering of objects based on stereo vision as well as non-transitory computer-readable medium | |
| CN107491762B (en) | A kind of pedestrian detection method | |
| US10062172B2 (en) | Automated tattoo recognition techniques | |
| CN108875600A (en) | A kind of information of vehicles detection and tracking method, apparatus and computer storage medium based on YOLO | |
| CN103679168B (en) | Detection method and detection device for character region | |
| CN109086724B (en) | Accelerated human face detection method and storage medium | |
| US20130243343A1 (en) | Method and device for people group detection | |
| CN103268481A (en) | Method for extracting text in complex background image | |
| CN105205486A (en) | Vehicle logo recognition method and device | |
| CN111079621B (en) | Method, device, electronic equipment and storage medium for detecting object | |
| CN109815863A (en) | Firework detecting method and system based on deep learning and image recognition | |
| CN102542244A (en) | Face detection method and system and computer program product | |
| CN106503710A (en) | A kind of automobile logo identification method and device | |
| CN115049954B (en) | Target identification method, device, electronic equipment and medium | |
| CN107194393A (en) | A kind of method and device for detecting Provisional Number Plate | |
| CN105740751A (en) | Object detection and identification method and system | |
| CN112417955A (en) | Patrol video stream processing method and device | |
| CN109255792B (en) | Video image segmentation method and device, terminal equipment and storage medium | |
| CN111695373A (en) | Zebra crossing positioning method, system, medium and device | |
| WO2023070955A1 (en) | Method and apparatus for detecting tiny target in port operation area on basis of computer vision | |
| CN109086781A (en) | A kind of cabinet lamp state identification method based on deep learning |
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 |