CN106650581A - Visitor flow rate statistics method and device - Google Patents
Visitor flow rate statistics method and device Download PDFInfo
- Publication number
- CN106650581A CN106650581A CN201610856298.5A CN201610856298A CN106650581A CN 106650581 A CN106650581 A CN 106650581A CN 201610856298 A CN201610856298 A CN 201610856298A CN 106650581 A CN106650581 A CN 106650581A
- Authority
- CN
- China
- Prior art keywords
- people
- feature
- people feature
- target area
- unit
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- 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
- G06F18/217—Validation; Performance evaluation; Active pattern learning techniques
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
- G06V20/53—Recognition of crowd images, e.g. recognition of crowd congestion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Databases & Information Systems (AREA)
- Human Computer Interaction (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Mathematical Physics (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention provides a visitor flow rate statistics method and device. The method comprises: a target detection image of a target area is obtained; head detection is carried out on the target detection image by using a deep learning method to obtain at least one head feature; a recursive analysis is carried out on the at least one head feature to obtain attribute information of the at least one head feature; and on the basis of the attribute information of the at least one head feature, statistics of the visitor flow rate in the target area is carried out. Therefore, head detection can be realized effectively and conveniently and the accuracy of the visitor flow rate statistics can be improved.
Description
Technical field
The present invention relates to Internet technical field, and in particular to technical field of image processing, more particularly to a kind of flow of the people
Statistical method and device.
Background technology
With the development of Internet technology, image processing techniques is developed rapidly.In image processing field, the stream of people
Amount statistics is an important application, is also a frontier in current intelligent video monitoring.Due to human head shapes, (class is justified
Shape) relative to human body, other position shapes are more fixed, and the possibility blocked is less, therefore number of people detection Jing is normal
In being applied to people flow rate statistical.The main task of number of people detection is, by automatically analyzing to input picture, people to be caught in time
The size and location of body contouring head.At present, based on number of people detection technique people flow rate statistical method master include it is following several,
One be based on the people flow rate statistical method of Image Edge-Detection, the method be randomly select along image border it is not conllinear several
Calculating distance between points is put so as to judge whether to there may be number of people target；The method accuracy rate is not high, and detects
Speed can not meet requirement of real-time.Its two be based on Hough (Hough) convert people flow rate statistical, the ginseng involved by the method
Number space is three dimensions, and its amount of calculation is larger and processes complicated, and accuracy rate is not high.
The content of the invention
The embodiment of the present invention provides a kind of people flow rate statistical method and device, can more effectively, more easily realize the number of people
Detection, lifts the accuracy rate of people flow rate statistical.
Embodiment of the present invention first aspect provides a kind of people flow rate statistical method, it may include：
Obtain the target detection image of target area；
Number of people detection is carried out to the target detection image using deep learning method, at least one number of people feature is obtained；
Recursive analysis is carried out at least one number of people feature, the attribute letter of at least one number of people feature is obtained
Breath；
The flow of the people of the target area is counted according to the attribute information of at least one number of people feature.
Preferably, the target detection image of the acquisition target area, including：
Determine target area to be counted, wherein, the target area is a closed area, or the target area
It is a designated area in open area；
In the target area, at least camera head all the way is set；
Image information described in synchronous acquisition at least all the way captured by camera head；
The image information at least all the way captured by camera head is carried out into panorama mosaic and forms target detection image.
Preferably, the employing deep learning method carries out number of people detection to the target detection image, obtains at least one
Individual number of people feature, including：
A sample image is randomly selected from the target detection image；
The density of stream of people reflected according to the sample image determines the level parameter of deep learning；
Level parameter builds depth convolutional neural networks model according to determined by；
Number of people detection is carried out to the target detection image using the depth convolutional neural networks model, at least one is obtained
Individual number of people feature.
Preferably, it is described that recursive analysis is carried out at least one number of people feature, obtain at least one number of people special
The attribute information levied, including：
The number of nodes of recursive analysis is determined according to the quantity of at least one number of people feature；
Number of nodes builds recurrent neural networks model according to determined by；
At least one number of people feature is analyzed using the recurrent neural networks model, obtains described at least one
The attribute information of individual number of people feature；
Wherein, the attribute information of a number of people feature includes：Confirm probability and confirm position.
Preferably, the stream of people that the target area is counted according to the attribute information of at least one number of people feature
Amount, including：
Validity sieve is carried out at least one number of people feature according to the attribute information of at least one number of people feature
Choosing is processed；
The flow of the people of target area according at least one effective number of people characteristic statisticses that screening is obtained.
Preferably, the attribute information according at least one number of people feature enters at least one number of people feature
Row validity Screening Treatment, including：
To confirm that probability is defined as the alternative number of people more than the number of people feature of preset value at least one number of people feature special
Levy；
Non- maximum suppression process is carried out to the confirmation position of each alternative number of people feature, is filtered and is confirmed that position produces what is overlapped
Alternative number of people feature, obtains remaining effective number of people feature.
Embodiment of the present invention second aspect provides a kind of people flow rate statistical device, it may include：
Acquiring unit, for obtaining the target detection image of target area；
Detector unit, for number of people detection to be carried out to the target detection image using deep learning method, obtains at least
One number of people feature；
Analytic unit, for carrying out recursive analysis at least one number of people feature, obtains at least one number of people
The attribute information of feature；
Statistic unit, for counting the stream of people of the target area according to the attribute information of at least one number of people feature
Amount.
Preferably, the acquiring unit includes：
Area determination unit, for determining target area to be counted, wherein, the target area is an enclosed area
Domain, or the target area is a designated area in open area；
Setting unit, for arranging at least camera head all the way in the target area；
Collecting unit, for the image information described in synchronous acquisition at least all the way captured by camera head；
Synthesis unit, for the image information at least all the way captured by camera head to be carried out into panorama mosaic mesh is formed
Mark detection image.
Preferably, the detector unit includes：
Unit is chosen, for randomly selecting a sample image from the target detection image；
Parameter determination unit, the density of stream of people for being reflected according to the sample image determines the level ginseng of deep learning
Number；
First model construction unit, for the level parameter according to determined by depth convolutional neural networks model is built；
Number of people detector unit, for entering pedestrian to the target detection image using the depth convolutional neural networks model
Head detection, obtains at least one number of people feature.
Preferably, the analytic unit includes：
Quantity determining unit, for determining the nodes of recursive analysis according to the quantity of at least one number of people feature
Amount；
Second model construction unit, for the number of nodes according to determined by recurrent neural networks model is built；
Recursive analysis unit, for being carried out at least one number of people feature point using the recurrent neural networks model
Analysis, obtains the attribute information of at least one number of people feature；
Wherein, the attribute information of a number of people feature includes：Confirm probability and confirm position.
Preferably, the statistic unit includes：
Screening Treatment unit, for according to the attribute information of at least one number of people feature at least one number of people
Feature carries out validity Screening Treatment；
Traffic statistics unit, for the target area according to screening at least one effective number of people characteristic statisticses for obtaining
Flow of the people.
Preferably, the Screening Treatment unit includes：
Alternative determining unit, for number of people feature of the probability more than preset value being confirmed at least one number of people feature
It is defined as alternative number of people feature；
Filter element is overlapped, is processed for carrying out non-maximum suppression to the confirmation position of each alternative number of people feature, filtered
Confirm that position produces the alternative number of people feature for overlapping, obtain remaining effective number of people feature.
In the embodiment of the present invention, deep learning is carried out by the target detection image to target area and obtains at least one people
Head feature, then the attribute information that recursive analysis obtains number of people feature is carried out to resulting number of people feature, it is last special according to the number of people
The attribute information levied carries out people flow rate statistical；Because the process of deep learning is obtained in that the abstract number of people of high-level, deep layer is special
Levy, this cause the number of people detect more effectively, it is more convenient, with reference to the process of recursive analysis, the significant increase standard of number of people detection
Exactness, improves the accuracy rate of people flow rate statistical.
Description of the drawings
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
The accompanying drawing to be used needed for having technology description is briefly described, it should be apparent that, drawings in the following description are only this
Some embodiments of invention, for those of ordinary skill in the art, on the premise of not paying creative work, can be with
Other accompanying drawings are obtained according to these accompanying drawings.
Fig. 1 is a kind of flow chart of people flow rate statistical method provided in an embodiment of the present invention；
Fig. 2 is the schematic diagram of target area provided in an embodiment of the present invention；
Fig. 3 is the flow chart of another kind of people flow rate statistical method provided in an embodiment of the present invention；
Fig. 4 is the schematic diagram of number of people testing process provided in an embodiment of the present invention；
Fig. 5 is a kind of internal structure schematic diagram of terminal provided in an embodiment of the present invention；
Fig. 6 is a kind of structural representation of people flow rate statistical device provided in an embodiment of the present invention.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation is described, it is clear that described embodiment is only a part of embodiment of the invention, rather than the embodiment of whole.It is based on
Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under the premise of creative work is not made
Embodiment, belongs to the scope of protection of the invention.
A kind of people flow rate statistical method and device detected based on the number of people is embodiments provided, first by mesh
The target detection image in mark region carries out deep learning and obtains at least one number of people feature, because the process of deep learning can be obtained
The abstract number of people feature of high-level, deep layer, this cause the number of people detect more effectively, it is more convenient.And then to number of people feature
The attribute information that recursive analysis obtains number of people feature is carried out, the process of deep learning combines the process of recursive analysis, significant increase
The degree of accuracy of number of people detection.The last attribute information according to number of people feature carries out people flow rate statistical, the people based on high accuracy
Head testing result greatly improves the accuracy rate of people flow rate statistical, while also improving the intelligent of people flow rate statistical method.
In the embodiment of the present invention, the main task of number of people detection is that detection image is automatically analyzed, and people is caught in time
Head feature (i.e. the size and location of number of people contouring)；And the task of people flow rate statistical be based on the number of people feature that captured come
Count the number total amount in certain region.The people flow rate statistical scheme detected based on the number of people that the embodiment of the present invention is provided can be answered
For several scenes, for example：Can apply to enter the indoor environment of this kind of closings such as school's self-study classroom, KTV rooms
The scene of pedestrian's traffic statistics；The outdoor to this kind of opening such as square gate area, bus platform can also be applied to
Certain specific region carries out the scene of people flow rate statistical in environment.
Based on foregoing description, a kind of people flow rate statistical method is embodiments provided, refer to Fig. 1, the method can
Comprise the following steps S101- steps S104.
S101, obtains the target detection image of target area.
Target area can be a closed area, for example：Target area can be school's self-study classroom, KTV rooms etc.
Deng；Target area can also be a designated area being located in open area, for example：Target area can be tourist attraction
Bus platform region on gate area, public way etc..Target detection image refers to the people of reflection target area
The image of traffic environment, the target detection image both can be the picture of one or more static state, or by a frame frame figure
As the dynamic video image sequence of composition.
In implementing, step S101 can be examined by setting up the target in camera head photographic subjects region in target area
Altimetric image；Camera head herein can be camera, video recording equipment etc..It is enforcement of the present invention please also refer to Fig. 2
The schematic diagram of the target area that example is provided；As shown in Fig. 2 w1 represents the length of target area, w2 represents the width of target area；
One or more camera heads can be set at H height, for the target detection image in photographic subjects region.May be appreciated
It is that the value of H can be set according to practical experience value, if the value setting of H is too small, possibly cannot completely photographs target
The contouring head of all personnel in region, therefore in order to ensure the content integrity of captured target detection image, then H
Value should be more than target area in personnel maximum height；In addition, if the value setting of H is excessive, then may make to photograph
Personnel contouring head it is less or unclear, in order to ensure the definition of the contouring head of captured personnel, the value of H should
It is too much when arranging.
S102, using deep learning method number of people detection is carried out to the target detection image, obtains at least one number of people
Feature.
Deep learning method is substantially a kind of machine learning method, its object is to：Setting up one can simulate the mankind
Brain is analyzed the neutral net of study, by imitating the mechanism of human brain come all kinds of to image, sound, text etc.
Data are explained.In implementing, deep learning method can be realized based on various models, model herein may include but
It is not limited to：Depth convolutional neural networks (Convolutional Neural Networks, CNN) model, fuzzy neural network mould
Type, BP (Back-Propagation, Back-propagation) neural network model etc..This step can be by building neutral net mould
Type, deep learning is carried out using constructed neural network model to target detection image, so as to realize that number of people detection is obtained
At least one number of people feature；Herein, the number of people is characterized in that the attributes such as the size, the position that refer to describe human head profile comprehensively
Feature.
S103, recursive analysis is carried out at least one number of people feature, obtains the category of at least one number of people feature
Property information.
The attribute information of number of people feature may include but be not limited to：Confirm probability and confirm position.Wherein, confirm that probability is referred to
Content described by number of people feature belongs to the probability of human head profile；Confirm that probability is bigger, show described by number of people feature
The possibility that content belongs to human head profile is bigger；Confirm that probability is less, show that the content described by number of people feature belongs to people
The possibility of body contouring head is less.Wherein, confirm that position refers to people's head contour described by number of people feature in target detection figure
The occupied position as in；In practical application, people can be marked by modes such as square frame, similar round frames in target detection image
The confirmation position of body contouring head, the square frame, the size of similar round frame can also reflect to a certain extent human head profile
Size.In implementing, the process of recursive analysis is referred to：It is special so as to export the number of people by carrying out depth training to number of people feature
The confirmation probability levied and confirmation position.Recursive analysis can be realized based on various models, including but not limited to：RNN
(Long-Short Term Memory, length is remembered in short-term for (Recurrent Neural Net, recurrent neural network) model, LSTM
Recall neutral net) model etc..This step can be by building recurrent neural networks model, using constructed recurrent neural net
Network model carrying out depth training and analysis to number of people feature, so as to export the attribute information of number of people feature.
S104, according to the attribute information of at least one number of people feature flow of the people of the target area is counted.
Obtained comprising at least one number of people feature for confirming probability and confirmation position by abovementioned steps, but these people
May there are some invalid number of people features in head feature, for example：The confirmation probability of some number of people features is less, shows these
Number of people feature may describe the contour feature of other objects and the not feature of human head profile, then need these
Number of people feature is filtered；For another example：The confirmation position registration of some number of people features is higher, shows that these number of people features may be retouched
What is stated is the contouring head feature of same person, then, need to carry out heavy filtration to these number of people features.Through to invalid
Number of people feature is carried out after screening and filtering, and the flow of the people of target area can be counted based on remaining effective number of people feature, that is, unite
In respect of the quantity of effect number of people feature so as to obtaining the flow of the people size of target area.
The people flow rate statistical method of the embodiment of the present invention, by the target detection image to target area deep learning is carried out
At least one number of people feature is obtained, then the attribute information that recursive analysis obtains number of people feature is carried out to resulting number of people feature,
The last attribute information according to number of people feature carries out people flow rate statistical；Because the process of deep learning is obtained in that high-level, depth
The abstract number of people feature of layer, this cause the number of people detect more effectively, it is more convenient, with reference to the process of recursive analysis, significant increase
The degree of accuracy of number of people detection, improves the accuracy rate of people flow rate statistical.
The embodiment of the present invention additionally provides another kind of people flow rate statistical method, refers to Fig. 3, and the method may include following step
Rapid S201- steps S213.
S201, determines target area to be counted, wherein, the target area is a closed area, or the mesh
Mark region is a designated area in open area.
S202, in the target area at least camera head all the way is arranged.
S203, the image information described in synchronous acquisition at least all the way captured by camera head.
S204, carries out the image information at least all the way captured by camera head panorama mosaic and forms target detection figure
Picture.
The step of the present embodiment S201-S204 S101 the step of can be embodiment illustrated in fig. 1 concrete refinement step.
In step S201-S204：Target area to be counted can be determined according to user's request is received, for example：Certain
User's request is counted to the flow of the people in certain room, then can determine that target area is when the user's request is received
The room asked；For another example：Certain user's request is counted to the flow of the people of the gateway on certain square, then receiving this
The gate area that target area is asked square is can determine that during user's request.After determining target area, can be in target
Region sets up the flow of the people environment that one or more camera head comes photographic subjects region, and synchronous acquisition one or more shooting
Image information taken by device carries out panorama composing process and forms target detection image.Because target detection image is by one
Plant or multichannel camera head carries out multi-angle, shoots in all directions, and the image that many figure panorama composing process of Jing are formed, because
This, target detection image can comprehensively, it is comprehensive without dead angle reflection target area flow of the people environment, be conducive to lifted people
The accuracy of head detection and the accuracy rate of people flow rate statistical.
S205, randomly selects a sample image from the target detection image.
S206, the density of stream of people reflected according to the sample image determines the level parameter of deep learning.
S207, level parameter builds depth convolutional neural networks model according to determined by.
S208, number of people detection is carried out using the depth convolutional neural networks model to the target detection image, is obtained
At least one number of people feature.
The step of the present embodiment S205-S208 S102 the step of can be embodiment illustrated in fig. 1 concrete refinement step.
The present embodiment by based on depth convolutional neural networks MODEL C NN model come the process for realizing deep learning as a example by carry out
Explanation, then in step S205-S208：Need to build a CNN model.The basic structure of one CNN network includes：Input
Layer, intermediate layer (being called hidden layer) and output layer；Wherein, according to actual needs, the quantity in intermediate layer can be one or more layers,
An important step for building CNN models is the hierarchical structure for determining intermediate layer.In implementing, can examine from the target
A sample image is randomly selected in altimetric image, intuitively, this process is equivalent to from this large scale of target detection image
A small images are randomly selected in image as sample image.The main task of CNN models is the sample image from this fritter
Learning is applied to any of target detection image this large-size images to some features (i.e. CNN features), then by CNN features
Place.There is corresponding relation in the density of stream of people that sample image is reflected, with level parameter if the stream of people that sample image is reflected
Density is larger, and correspondence can set the level parameter of a higher value；If the density of stream of people that sample image is reflected is less, right
The level parameter of a smaller value should be set；Level parameter herein is used to determine the hierarchical structure in intermediate layer, for example：It is false
If at most there are N=4 personnel less than or equal to the region indicated by 4, i.e. sample image in the density of stream of people of sample image, then
The level capable setting parameter of CNN models is 4, shows that CNN models, altogether comprising four layers of intermediate layer, thus can build CNN models.Enter one
Step, number of people detection is carried out using constructed CNN models to target detection image, can obtain CNN1, CNN2, CNN3 and CNN4
Totally four number of people features, four number of people features represent respectively the abstract feature for obtaining of different depth, specially：CNN1 represents
The abstract number of people feature for obtaining in one layer of intermediate layer, CNN2 represents the abstract number of people feature for obtaining in the second intermediate layer, by that analogy.Need
It is noted that the CNN models in above-mentioned four layers of intermediate layer are only for example, the hierarchical structure in the corresponding intermediate layer of CNN features is higher,
Its CNN feature is more abstract；CNN model depths constructed in practical application are deeper, carry out the CNN features that number of people detection is obtained
It is more abstract, carry out number of people detection by using the CNN features of CNN model different depths, can more comprehensively, it is more abstract, more have
Person head profile in effect, deeper description target area so that number of people testing result is more accurate.
S209, according to the quantity of at least one number of people feature number of nodes of recursive analysis is determined.
S210, number of nodes builds recurrent neural networks model according to determined by.
S211, is analyzed using the recurrent neural networks model at least one number of people feature, obtains described
The attribute information of at least one number of people feature.
The step of the present embodiment S209-S211 S103 the step of can be embodiment illustrated in fig. 1 concrete refinement step.
The present embodiment is illustrated as a example by the process of recursive analysis is realized based on recurrent neural networks model, then step
In rapid S209-S211：Need to build recurrent neural networks model, it is preferable that recurrent neural networks model can be LSTM models.
An important step for building LSTM models is the quantity for determining LSTM neural units (node).The main task of LSTM models
It is that depth training is carried out to number of people feature, so as to exporting the confirmation probability of number of people feature and confirming position.Need the number of people of analysis
There is corresponding relation in the quantity of feature, more if necessary to the quantity of the number of people feature of analysis with the quantity of LSTM neural units,
The quantity of so required LSTM neural units is more；It is fewer if necessary to the quantity of the number of people feature of analysis, then required
The quantity of LSTM neural units is fewer；Specifically, can be with number of people feature one LSTM neural unit of correspondence.Jing CNN models
The CNN features of extraction are input to respectively different LSTM neural units, after each LSTM unit is analyzed to number of people feature,
The confirmation probability of output CNN features and confirmation position.
Step S205-S211 is the flow process description of number of people detection, is provided in an embodiment of the present invention please also refer to Fig. 4
The schematic diagram of number of people testing process；In example shown in Fig. 4, CNN models include four layers of intermediate layer, and LSTM models include 4 LSTM
Neural unit.In number of people testing process：Target detection image is input to CNN models and carries out after deep learning, exports CNN1-
Tetra- number of people features of CNN4；CNN1 is input to LSTM-1 neural units and carries out after recursive analysis, exports the attribute information of CNN1；
CNN2 is input to LSTM-2 neural units and carries out after recursive analysis, exports the attribute information of CNN2；By that analogy.
S212, is carried out effectively according to the attribute information of at least one number of people feature at least one number of people feature
Property Screening Treatment.
In implementing, the method is concrete to perform following steps s11-s12 during execution step S212：
S11, will confirm that probability is defined as alternative people more than the number of people feature of preset value at least one number of people feature
Head feature.
S12, to the confirmation position of each alternative number of people feature non-maximum suppression process is carried out, and is filtered and is confirmed that position produces weight
Folded alternative number of people feature, obtains remaining effective number of people feature.
In step s11-s12：Preset value can according to actual needs or practical experience value is set, if certain number of people is special
The confirmation probability levied is more than preset value, show the content described by the number of people feature belong to the possibility of human head profile compared with
Greatly, then then select the number of people feature alternately number of people feature., whereas if the confirmation probability of certain number of people feature is less than or waits
In preset value, show the content described by the number of people feature belong to non-human contouring head possibility it is larger, this number of people feature
Interference can be produced to people flow rate statistical, then then need to give up the number of people feature.Further, if a few number of people features really
Recognize that position registration is higher, illustrate these number of people features for describing the contouring head of same personnel, these number of people features are produced
It is raw to overlap, then the statistics that should only retain one of number of people feature and filter out other number of people features to ensure flow of the people is accurate
True property；In practical application, processing method can be suppressed using non-maximum, find out the registration for confirming position more than default overlap
Two or more number of people features of threshold values (can be set according to actual needs), reservation confirms that of maximum probability
Individual number of people feature, and other number of people characteristic filters for overlapping are fallen.
S213, the flow of the people of target area according at least one effective number of people characteristic statisticses that screening is obtained.
The step of the present embodiment S212-S213 S104 the step of can be embodiment illustrated in fig. 1 concrete refinement step.
In step S212-S213, it is invalid at least one number of people feature to be fallen with screening and filtering through step s11-s12
Number of people feature, then remaining to be effective number of people feature, further can count target based on remaining effective number of people feature
The flow of the people in region, that is, count the quantity of effective number of people feature so as to obtain the flow of the people size of target area.
The people flow rate statistical method of the embodiment of the present invention, by the target detection image to target area deep learning is carried out
At least one number of people feature is obtained, then the attribute information that recursive analysis obtains number of people feature is carried out to resulting number of people feature,
The last attribute information according to number of people feature carries out people flow rate statistical；Because the process of deep learning is obtained in that high-level, depth
The abstract number of people feature of layer, this cause the number of people detect more effectively, it is more convenient, with reference to the process of recursive analysis, significant increase
The degree of accuracy of number of people detection, improves the accuracy rate of people flow rate statistical.
Based on above-mentioned people flow rate statistical method, the embodiment of the present invention additionally provides a kind of people flow rate statistical device, the dress
Putting can run in a terminal, and terminal herein may include but be not limited to：PC (Personal Computer, individual calculus
Machine), smart mobile phone, the equipment such as PAD (panel computer).Please also refer to Fig. 5, the internal structure of the terminal may include but not limit
In：Processor, user interface, network interface and memory.Wherein, in terminal processor, user interface, network interface and deposit
Reservoir can be connected by bus or other modes, in Fig. 5 shown in the embodiment of the present invention as a example by being connected by bus.
Wherein, user interface is to realize that user interacts the medium exchanged with information with terminal, and it is embodied can be with
Including the display screen (Display) for output and the keyboard (Keyboard) for input etc., it should be noted that this
The keyboard at place both can be physical keyboard, or touch screen dummy keyboard, can also be what entity was combined with touch screen virtualphase
Keyboard.Processor (or claim CPU (Central Processing Unit, central processing unit)) be the calculating core of terminal and
Control core, it can parse the Various types of data of all kinds of instructions in terminal and processing terminal, for example：CPU can be used for solution
Analysis user to the switching on and shutting down that terminal is sent are instructed, and control terminal carries out switching on and shutting down operation；For another example：CPU can be in terminal
All kinds of interaction datas, etc. are transmitted between portion's structure.Memory (Memory) is the memory device in terminal, for depositing program
And data.It is understood that memory herein can both include the internal memory of terminal, naturally it is also possible to including terminal
The extended menory supported.Memory provides memory space, and the memory space stores the operating system of terminal.It is of the invention real
In applying example, the memory space of memory also stored for people flow rate statistical device.Terminal is by the flow of the people system in run memory
Counter device is performing the corresponding steps of method flow shown in above-mentioned Fig. 1-4.Fig. 6 is referred to, the people flow rate statistical plant running is such as
Lower unit：
Acquiring unit 101, for obtaining the target detection image of target area.
Detector unit 102, for carrying out number of people detection to the target detection image using deep learning method, obtain to
A few number of people feature.
Analytic unit 103, for carrying out recursive analysis at least one number of people feature, obtains at least one people
The attribute information of head feature.
Statistic unit 104, for counting the target area according to the attribute information of at least one number of people feature
Flow of the people.
In implementing, the device during the acquiring unit 101 is run, carrying out practically such as lower unit：
Area determination unit 1001, for determining target area to be counted, wherein, the target area is a closing
Region, or the target area is a designated area in open area.
Setting unit 1002, for arranging at least camera head all the way in the target area.
Collecting unit 1003, for the image information described in synchronous acquisition at least all the way captured by camera head.
Synthesis unit 1004, for the image information at least all the way captured by camera head to be carried out into panorama mosaic shape
Into target detection image.
In implementing, the device during the detector unit 102 is run, carrying out practically such as lower unit：
Unit 2001 is chosen, for randomly selecting a sample image from the target detection image.
Parameter determination unit 2002, the density of stream of people for being reflected according to the sample image determines the layer of deep learning
Level parameter.
First model construction unit 2003, for the level parameter according to determined by depth convolutional neural networks mould is built
Type.
Number of people detector unit 2004, for being entered to the target detection image using the depth convolutional neural networks model
The detection of pedestrian's head, obtains at least one number of people feature.
In implementing, the device during the analytic unit 103 is run, carrying out practically such as lower unit：
Quantity determining unit 3001, for determining the node of recursive analysis according to the quantity of at least one number of people feature
Quantity.
Second model construction unit 3002, for the number of nodes according to determined by recurrent neural networks model is built.
Recursive analysis unit 3003, for being entered at least one number of people feature using the recurrent neural networks model
Row analysis, obtains the attribute information of at least one number of people feature；Wherein, the attribute information of a number of people feature includes：Really
Recognize probability and confirm position.
In implementing, during statistic unit 104 described in the plant running, carrying out practically such as lower unit：
Screening Treatment unit 4001, for according to the attribute information of at least one number of people feature to described at least one
Number of people feature carries out validity Screening Treatment.
Traffic statistics unit 4002, for the target area according to screening at least one effective number of people characteristic statisticses for obtaining
The flow of the people in domain.
In implementing, the device during the Screening Treatment unit 4001 is run, carrying out practically such as lower unit：
Alternative determining unit 4441, for the number of people of the probability more than preset value being confirmed at least one number of people feature
Feature is defined as alternative number of people feature；
Filter element 4442 is overlapped, is processed for carrying out non-maximum suppression to the confirmation position of each alternative number of people feature,
Filter and confirm that position produces the alternative number of people feature for overlapping, obtain remaining effective number of people feature.
Because terminal performs the method shown in Fig. 1-Fig. 4 by running the people flow rate statistical device shown in Fig. 6, therefore,
The function of each unit of Fig. 6 shown devices can be found in the associated description of each step of method shown in Fig. 1-Fig. 4, will not be described here.
With method in the same manner, the people flow rate statistical device of the embodiment of the present invention, by the target detection image to target area
At least one number of people feature is obtained and carries out deep learning, then recursive analysis is carried out to resulting number of people feature obtaining number of people feature
Attribute information, the last attribute information according to number of people feature carries out people flow rate statistical；Because the process of deep learning can be obtained
The abstract number of people feature of high-level, deep layer, this cause the number of people detect more effectively, it is more convenient, with reference to the mistake of recursive analysis
Journey, the degree of accuracy of significant increase number of people detection, improves the accuracy rate of people flow rate statistical.
One of ordinary skill in the art will appreciate that realizing all or part of flow process in above-described embodiment method, can be
Related hardware is instructed to complete by computer program, described program can be stored in a computer read/write memory medium
In, the program is upon execution, it may include such as the flow process of the embodiment of above-mentioned each method.Wherein, described storage medium can be magnetic
Dish, CD, read-only memory (Read-Only Memory, ROM) or random access memory (Random Access
Memory, RAM) etc..
Above disclosed is only present pre-ferred embodiments, can not limit the right model of the present invention with this certainly
Enclose, therefore the equivalent variations made according to the claims in the present invention, still belong to the scope that the present invention is covered.
Claims (12)
1. a kind of people flow rate statistical method, it is characterised in that include：
Obtain the target detection image of target area；
Number of people detection is carried out to the target detection image using deep learning method, at least one number of people feature is obtained；
Recursive analysis is carried out at least one number of people feature, the attribute information of at least one number of people feature is obtained；
The flow of the people of the target area is counted according to the attribute information of at least one number of people feature.
2. the method for claim 1, it is characterised in that the target detection image of the acquisition target area, including：
Determine target area to be counted, wherein, the target area is a closed area, or the target area is position
A designated area in open area；
In the target area, at least camera head all the way is set；
Image information described in synchronous acquisition at least all the way captured by camera head；
The image information at least all the way captured by camera head is carried out into panorama mosaic and forms target detection image.
3. method as claimed in claim 1 or 2, it is characterised in that the employing deep learning method is to the target detection
Image carries out number of people detection, obtains at least one number of people feature, including：
A sample image is randomly selected from the target detection image；
The density of stream of people reflected according to the sample image determines the level parameter of deep learning；
Level parameter builds depth convolutional neural networks model according to determined by；
Number of people detection is carried out to the target detection image using the depth convolutional neural networks model, at least one people is obtained
Head feature.
4. method as claimed in claim 3, it is characterised in that described that recurrence point is carried out at least one number of people feature
Analysis, obtains the attribute information of at least one number of people feature, including：
The number of nodes of recursive analysis is determined according to the quantity of at least one number of people feature；
Number of nodes builds recurrent neural networks model according to determined by；
At least one number of people feature is analyzed using the recurrent neural networks model, obtains at least one people
The attribute information of head feature；
Wherein, the attribute information of a number of people feature includes：Confirm probability and confirm position.
5. method as claimed in claim 4, it is characterised in that the attribute information according at least one number of people feature
The flow of the people of the target area is counted, including：
At least one number of people feature is carried out at validity screening according to the attribute information of at least one number of people feature
Reason；
The flow of the people of target area according at least one effective number of people characteristic statisticses that screening is obtained.
6. method as claimed in claim 5, it is characterised in that the attribute information according at least one number of people feature
Validity Screening Treatment is carried out at least one number of people feature, including：
To confirm that probability is defined as alternative number of people feature more than the number of people feature of preset value at least one number of people feature；
Non- maximum suppression process is carried out to the confirmation position of each alternative number of people feature, is filtered and is confirmed that position produces the alternative of overlap
Number of people feature, obtains remaining effective number of people feature.
7. a kind of people flow rate statistical device, it is characterised in that include：
Acquiring unit, for obtaining the target detection image of target area；
Detector unit, for number of people detection to be carried out to the target detection image using deep learning method, obtains at least one
Number of people feature；
Analytic unit, for carrying out recursive analysis at least one number of people feature, obtains at least one number of people feature
Attribute information；
Statistic unit, for counting the flow of the people of the target area according to the attribute information of at least one number of people feature.
8. device as claimed in claim 7, it is characterised in that the acquiring unit includes：
Area determination unit, for determining target area to be counted, wherein, the target area is a closed area, or
Target area described in person is a designated area in open area；
Setting unit, for arranging at least camera head all the way in the target area；
Collecting unit, for the image information described in synchronous acquisition at least all the way captured by camera head；
Synthesis unit, for the image information at least all the way captured by camera head to be carried out into panorama mosaic target inspection is formed
Altimetric image.
9. device as claimed in claim 7 or 8, it is characterised in that the detector unit includes：
Unit is chosen, for randomly selecting a sample image from the target detection image；
Parameter determination unit, the density of stream of people for being reflected according to the sample image determines the level parameter of deep learning；
First model construction unit, for the level parameter according to determined by depth convolutional neural networks model is built；
Number of people detector unit, for carrying out number of people inspection to the target detection image using the depth convolutional neural networks model
Survey, obtain at least one number of people feature.
10. device as claimed in claim 9, it is characterised in that the analytic unit includes：
Quantity determining unit, for determining the number of nodes of recursive analysis according to the quantity of at least one number of people feature；
Second model construction unit, for the number of nodes according to determined by recurrent neural networks model is built；
Recursive analysis unit, for being analyzed at least one number of people feature using the recurrent neural networks model,
Obtain the attribute information of at least one number of people feature；
Wherein, the attribute information of a number of people feature includes：Confirm probability and confirm position.
11. devices as claimed in claim 10, it is characterised in that the statistic unit includes：
Screening Treatment unit, for according to the attribute information of at least one number of people feature at least one number of people feature
Carry out validity Screening Treatment；
Traffic statistics unit, for the stream of people of the target area according to screening at least one effective number of people characteristic statisticses for obtaining
Amount.
12. devices as claimed in claim 11, it is characterised in that the Screening Treatment unit includes：
Alternative determining unit, for will confirm that probability determines more than the number of people feature of preset value at least one number of people feature
For alternative number of people feature；
Filter element is overlapped, is processed for carrying out non-maximum suppression to the confirmation position of each alternative number of people feature, filtered and confirm
Position produces the alternative number of people feature for overlapping, and obtains remaining effective number of people feature.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610856298.5A CN106650581B (en) | 2016-09-27 | A kind of people flow rate statistical method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610856298.5A CN106650581B (en) | 2016-09-27 | A kind of people flow rate statistical method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106650581A true CN106650581A (en) | 2017-05-10 |
| CN106650581B CN106650581B (en) | 2019-07-16 |
Family
ID=
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107392097A (en) * | 2017-06-15 | 2017-11-24 | 中山大学 | A kind of 3 D human body intra-articular irrigation method of monocular color video |
| CN107844832A (en) * | 2017-10-30 | 2018-03-27 | 上海寒武纪信息科技有限公司 | A kind of information processing method and Related product |
| CN108171162A (en) * | 2017-12-27 | 2018-06-15 | 重庆交通开投科技发展有限公司 | Crowded degree detection method, apparatus and system |
| CN108345854A (en) * | 2018-02-08 | 2018-07-31 | 腾讯科技（深圳）有限公司 | Information processing method, device, system based on image analysis and storage medium |
| CN108615020A (en) * | 2018-04-28 | 2018-10-02 | 东莞市华睿电子科技有限公司 | A kind of floating population number statistical method in video monitoring regional |
| CN108647615A (en) * | 2018-04-28 | 2018-10-12 | 东莞市华睿电子科技有限公司 | A kind of cinema steals a ride the implementation method of early warning |
| CN108830231A (en) * | 2018-06-21 | 2018-11-16 | 芜湖宝文信息科技有限公司 | The statistics equipment of people information statistics for height classification |
| CN108830145A (en) * | 2018-05-04 | 2018-11-16 | 深圳技术大学（筹） | A kind of demographic method and storage medium based on deep neural network |
| CN109117741A (en) * | 2018-07-20 | 2019-01-01 | 苏州中德宏泰电子科技股份有限公司 | Offline object identifying method and device to be detected |
| CN109740444A (en) * | 2018-12-13 | 2019-05-10 | 深圳云天励飞技术有限公司 | Flow of the people information displaying method and Related product |
| CN110084173A (en) * | 2019-04-23 | 2019-08-02 | 精伦电子股份有限公司 | Number of people detection method and device |
| CN110096959A (en) * | 2019-03-28 | 2019-08-06 | 上海拍拍贷金融信息服务有限公司 | Flow of the people calculation method, device and computer storage medium |
| US10373332B2 (en) | 2017-12-08 | 2019-08-06 | Nvidia Corporation | Systems and methods for dynamic facial analysis using a recurrent neural network |
| CN110262496A (en) * | 2019-06-27 | 2019-09-20 | 柴冠华 | A kind of intelligent mobile dustbin and its control method |
| CN112348765A (en) * | 2020-10-23 | 2021-02-09 | 深圳市优必选科技股份有限公司 | Data enhancement method and device, computer readable storage medium and terminal equipment |
| US11990137B2 (en) | 2018-09-13 | 2024-05-21 | Shanghai Cambricon Information Technology Co., Ltd. | Image retouching method and terminal device |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101877058A (en) * | 2010-02-10 | 2010-11-03 | 杭州海康威视软件有限公司 | People flow rate statistical method and system |
| CN104992223A (en) * | 2015-06-12 | 2015-10-21 | 安徽大学 | Dense population estimation method based on deep learning |
| CN105303193A (en) * | 2015-09-21 | 2016-02-03 | 重庆邮电大学 | People counting system for processing single-frame image |
| CN105321187A (en) * | 2014-07-22 | 2016-02-10 | 无锡慧眼电子科技有限公司 | Pedestrian counting method based on head detection |
| CN105631427A (en) * | 2015-12-29 | 2016-06-01 | 北京旷视科技有限公司 | Suspicious personnel detection method and system |
| CN105654104A (en) * | 2014-11-28 | 2016-06-08 | 无锡慧眼电子科技有限公司 | Pedestrian detection method based on multi-granularity feature |
| US20160224829A1 (en) * | 2015-02-04 | 2016-08-04 | UDP Technology Ltd. | People counter using tof camera and counting method thereof |
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101877058A (en) * | 2010-02-10 | 2010-11-03 | 杭州海康威视软件有限公司 | People flow rate statistical method and system |
| CN105321187A (en) * | 2014-07-22 | 2016-02-10 | 无锡慧眼电子科技有限公司 | Pedestrian counting method based on head detection |
| CN105654104A (en) * | 2014-11-28 | 2016-06-08 | 无锡慧眼电子科技有限公司 | Pedestrian detection method based on multi-granularity feature |
| US20160224829A1 (en) * | 2015-02-04 | 2016-08-04 | UDP Technology Ltd. | People counter using tof camera and counting method thereof |
| CN104992223A (en) * | 2015-06-12 | 2015-10-21 | 安徽大学 | Dense population estimation method based on deep learning |
| CN105303193A (en) * | 2015-09-21 | 2016-02-03 | 重庆邮电大学 | People counting system for processing single-frame image |
| CN105631427A (en) * | 2015-12-29 | 2016-06-01 | 北京旷视科技有限公司 | Suspicious personnel detection method and system |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107392097A (en) * | 2017-06-15 | 2017-11-24 | 中山大学 | A kind of 3 D human body intra-articular irrigation method of monocular color video |
| CN107392097B (en) * | 2017-06-15 | 2020-07-07 | 中山大学 | Three-dimensional human body joint point positioning method of monocular color video |
| CN109117947A (en) * | 2017-10-30 | 2019-01-01 | 上海寒武纪信息科技有限公司 | Profile testing method and Related product |
| US11762631B2 (en) | 2017-10-30 | 2023-09-19 | Shanghai Cambricon Information Technology Co., Ltd. | Information processing method and terminal device |
| US11922132B2 (en) | 2017-10-30 | 2024-03-05 | Shanghai Cambricon Information Technology Co., Ltd. | Information processing method and terminal device |
| CN107844832A (en) * | 2017-10-30 | 2018-03-27 | 上海寒武纪信息科技有限公司 | A kind of information processing method and Related product |
| US10373332B2 (en) | 2017-12-08 | 2019-08-06 | Nvidia Corporation | Systems and methods for dynamic facial analysis using a recurrent neural network |
| CN108171162B (en) * | 2017-12-27 | 2021-05-11 | 重庆交通开投科技发展有限公司 | Crowd crowding degree detection method, device and system |
| CN108171162A (en) * | 2017-12-27 | 2018-06-15 | 重庆交通开投科技发展有限公司 | Crowded degree detection method, apparatus and system |
| CN108345854A (en) * | 2018-02-08 | 2018-07-31 | 腾讯科技（深圳）有限公司 | Information processing method, device, system based on image analysis and storage medium |
| CN108615020A (en) * | 2018-04-28 | 2018-10-02 | 东莞市华睿电子科技有限公司 | A kind of floating population number statistical method in video monitoring regional |
| CN108647615A (en) * | 2018-04-28 | 2018-10-12 | 东莞市华睿电子科技有限公司 | A kind of cinema steals a ride the implementation method of early warning |
| CN108830145A (en) * | 2018-05-04 | 2018-11-16 | 深圳技术大学（筹） | A kind of demographic method and storage medium based on deep neural network |
| CN108830231A (en) * | 2018-06-21 | 2018-11-16 | 芜湖宝文信息科技有限公司 | The statistics equipment of people information statistics for height classification |
| CN109117741A (en) * | 2018-07-20 | 2019-01-01 | 苏州中德宏泰电子科技股份有限公司 | Offline object identifying method and device to be detected |
| US11990137B2 (en) | 2018-09-13 | 2024-05-21 | Shanghai Cambricon Information Technology Co., Ltd. | Image retouching method and terminal device |
| US11996105B2 (en) | 2018-09-13 | 2024-05-28 | Shanghai Cambricon Information Technology Co., Ltd. | Information processing method and terminal device |
| CN109740444A (en) * | 2018-12-13 | 2019-05-10 | 深圳云天励飞技术有限公司 | Flow of the people information displaying method and Related product |
| CN110096959A (en) * | 2019-03-28 | 2019-08-06 | 上海拍拍贷金融信息服务有限公司 | Flow of the people calculation method, device and computer storage medium |
| CN110096959B (en) * | 2019-03-28 | 2021-05-28 | 上海拍拍贷金融信息服务有限公司 | People flow calculation method, device and computer storage medium |
| CN110084173B (en) * | 2019-04-23 | 2021-06-15 | 精伦电子股份有限公司 | Human head detection method and device |
| CN110084173A (en) * | 2019-04-23 | 2019-08-02 | 精伦电子股份有限公司 | Number of people detection method and device |
| CN110262496B (en) * | 2019-06-27 | 2021-12-07 | 柴冠华 | Intelligent mobile garbage can and control method thereof |
| CN110262496A (en) * | 2019-06-27 | 2019-09-20 | 柴冠华 | A kind of intelligent mobile dustbin and its control method |
| CN112348765A (en) * | 2020-10-23 | 2021-02-09 | 深圳市优必选科技股份有限公司 | Data enhancement method and device, computer readable storage medium and terminal equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105550678B (en) | Human action feature extracting method based on global prominent edge region | |
| CN111126399B (en) | Image detection method, device and equipment and readable storage medium | |
| CN112464971A (en) | Method for constructing pest detection model | |
| CN109697416A (en) | A kind of video data handling procedure and relevant apparatus | |
| CN111832489A (en) | Subway crowd density estimation method and system based on target detection | |
| CN108288027A (en) | A kind of detection method of picture quality, device and equipment | |
| CN110491480A (en) | A kind of medical image processing method, device, electromedical equipment and storage medium | |
| CN108334847A (en) | A kind of face identification method based on deep learning under real scene | |
| CN108416314B (en) | Picture important face detection method | |
| CN109584204A (en) | A kind of image noise intensity estimation method, storage medium, processing and identification device | |
| CN109816745A (en) | Human body thermodynamic chart methods of exhibiting and Related product | |
| CN106570478A (en) | Object loss determine method and device in visual tracking | |
| CN110390673A (en) | Cigarette automatic testing method based on deep learning under a kind of monitoring scene | |
| CN112668375B (en) | Tourist distribution analysis system and method in scenic spot | |
| CN112417990A (en) | Examination student violation behavior identification method and system | |
| CN114093501A (en) | Intelligent auxiliary analysis method for children's motor epilepsy based on synchronous video and electroencephalogram | |
| CN110363087A (en) | A kind of Long baselines binocular human face in-vivo detection method and system | |
| CN113850183A (en) | Method for judging behaviors in video based on artificial intelligence technology | |
| CN110505440A (en) | A kind of area monitoring method and device | |
| Luong et al. | Towards window state detection using image processing in residential and office building facades | |
| CN107133964A (en) | A kind of stingy image space method based on Kinect | |
| CN107958231A (en) | Light field image filter method, human face analysis method and electronic equipment | |
| CN106650581A (en) | Visitor flow rate statistics method and device | |
| CN106650581B (en) | A kind of people flow rate statistical method and device | |
| CN108040244A (en) | Grasp shoot method and device, storage medium based on light field video flowing |
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 |