CN109190507A - A kind of passenger flow crowding calculation method and device based on rail transit train - Google Patents
A kind of passenger flow crowding calculation method and device based on rail transit train Download PDFInfo
- Publication number
- CN109190507A CN109190507A CN201810917472.1A CN201810917472A CN109190507A CN 109190507 A CN109190507 A CN 109190507A CN 201810917472 A CN201810917472 A CN 201810917472A CN 109190507 A CN109190507 A CN 109190507A
- Authority
- CN
- China
- Prior art keywords
- compartment
- grades
- neural networks
- convolutional neural
- training
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to a kind of passenger flow crowding calculation method and device based on rail transit train, the described method includes: S1: acquisition video simultaneously carries out sub-frame processing to collected video, compartment image is obtained, training set, verifying collection and test set are separately constituted by compartment image described in multiple;S2: standee in all compartment images is labeled;S3: the accuracy that the training set is trained and imports verifying collection verifying convolutional neural networks model is imported to depth convolutional neural networks model, to obtain the convolutional neural networks model after final training;S4: corresponding region of interest ROI is arranged according to the environment in compartment, and according to the convolutional neural networks model after the ROI and the training set, to collected video described in the convolutional neural networks mode input after the training or the image in individual compartment in the test set;S5: counting the passenger marked in the ROI, and the grade of intensity of passenger flow is divided according to count results.
Description
Technical field
The present invention relates to density data detection technique field more particularly to a kind of passenger flow based on rail transit train are crowded
Spend calculation method and device.
Background technique
2017,34, China urban track traffic distance travelled 5021.7km, 18,400,000,000 person-times of passenger capacity of year.It counts herein
Under, it is extremely urgent how to solve contradiction brought by operational safety and efficiency.Effective passenger flow managing and control system can be protected
It demonstrate,proves rail traffic safety and improves conevying efficiency under the premise of evading passenger flow risk, and the crowded environment of rail train carriage is visitor
Flow the center of gravity in managing and control system.Because under the premise that security is guaranteed, every train transport passenger as much as possible, will be reduced
Passenger is stayed in standing, and is improved passenger flow and is managed efficiency.But such a case is often had, (escalator is downstairs located in certain compartments
Compartment) saturation is had reached, it is crowded to capacity, and some compartments are quite loose.Therefore, it is united in every section of compartment by monitoring camera
The crowding of each compartment is counted, selects loose compartment to be of practical significance very much when waiting for passenger.
At this stage, camera detection is based on to most of compartment congestion levels, by camera come the people in detection image
Honorable portion/head number, as people on vehicle, this algorithm can accurately calculate the number in compartment, but when compartment is crowded
When, block excessively serious, and human body face/head size is inconsistent, and coverage extent is inconsistent, cause statistics human body face/
Head number is excessive with actual number error and then influences the precision of crowding.
Subway crowding predicts existing problems and disadvantages at this stage: occlusion issue is serious in compartment, and the passenger that stand hides
Passenger on block seat position, mutually block between the passenger that stand and compartment in the shelters such as vehicle handrail easily passenger is caused to block,
Passenger when very crowded in compartment on seat is almost kept off, and when the compartment degree of crowding is generally crowded, it stands
People can detect while the passenger on seat can also detect, the people that detected when be easy to causeing very crowded and general crowded
Number gap is little, and then influences the precision of detection crowding;The angle of grade classification, the camera in each compartment is different,
Monitor that range is also different, everyone may be different to the rank of same image definition, is lack of objective appraisal index.
Summary of the invention
The present invention is directed to current technology problem, provides a kind of passenger flow crowding calculating side based on rail transit train
Method and device.Lower, low efficiency that this method solve the arithmetic accuracies using the prior art, does not have objective appraisal index, no
The technical issues of being easy to practical application.
In a first aspect, the present invention provides a kind of passenger flow crowding calculation method based on rail transit train, comprising:
S1: acquisition video simultaneously carries out sub-frame processing to collected video, compartment image is obtained, by compartment figure described in multiple
As separately constituting training set, verifying collection and test set;
S2: standee in all compartment images is labeled;
S3: being trained to the depth convolutional neural networks model importing training set and imports verifying collection verifying convolution mind
Accuracy through network model, the convolutional neural networks model after obtaining final training;
S4: corresponding region of interest ROI is arranged according to the environment in compartment, and according to the ROI and the training set
Convolutional neural networks model afterwards, to collected video described in the convolutional neural networks mode input after the training or institute
State the image in individual compartment in test set;
S5: counting the passenger marked in the ROI, and the grade of intensity of passenger flow is divided according to count results.
Further, collected video described in the S1 is different time sections, different compartment videos collected.
Further, the training set in the S1 is made of 1800 compartment images, and the verifying collection is by 200
The compartment image composition, the test set are made of 800 compartment images.
Further, the grade of the intensity of passenger flow is divided into 1 grade, 2 grades and 3 grades total three-level or 1 grade, 2 grades, 3 grades and 4 grades totally four
Grade, wherein 1 grade of expression compartment is loose, 2 grades expression is generally crowded, 3 grades of very crowded and 4 grades of expression crowdings of expression are moderate.
Further, red dot mark is used to the process that standee in all compartment images is labeled in the S2
The mode of note is labeled human body face/head.
Second aspect, the invention discloses a kind of passenger flow crowding computing device based on rail transit train, comprising:
Framing module is acquired, for acquiring video and to the progress sub-frame processing of collected video, obtains compartment image, by
Multiple described compartment images separately constitute training set, verifying collection and test set;
Labeling module, for being labeled to standee in all compartment images;
Training module collects for importing the training set to depth convolutional neural networks model and being trained and import verifying
Verify the accuracy of convolutional neural networks model, the convolutional neural networks model after obtaining final training;
Input module, for corresponding region of interest ROI to be arranged according to the environment in compartment, and according to the ROI set
And the convolutional neural networks model after the training, to collected described in the convolutional neural networks mode input after the training
The image in individual compartment in video or the test set;
Diversity module, for being counted to the passenger that is marked in the ROI, and according to count results divide intensity of passenger flow etc.
Grade.
Further, the collected video of the acquisition framing module is different time sections, different compartment videos collected.
Further, the training set in the acquisition framing module is made of 1800 compartment images, described to test
Card collection is made of 200 compartment images, and the test set is made of 800 compartment images.
Further, the grade of the intensity of passenger flow in the diversity module be divided into 1 grade, 2 grades and 3 grades total three-level or 1 grade, 2 grades,
3 grades and 4 grades total level Four, wherein 1 grade of expression compartment is loose, general crowded, the 3 grades of very crowded and 4 grades of expressions of expression of 2 grades of expressions are gathered around
It is moderate to squeeze degree.
Further, the labeling module is labeled human body face/head by the way of red dot mark.
The beneficial effects of the present invention are:
Based on the present invention is counted by human body face/head, first to the human body face/head for the passenger that stand in compartment
It is labeled, abstract-learning is then carried out to human body face/head feature by deep learning again and obtains high-order feature, it is final to obtain
Human body face/head density figure in compartment, the gray value of statistical regression density map obtain the number of whole image.Due to vehicle
It is blocked inside compartment excessively seriously, only judges that the crowding in compartment can increase crowding etc. by human body face/head counting statistics
The error that grade differentiates, therefore ROI is carried out using the middle section to compartment image, the range in unified fixed compartment counts same
The different standees in different compartments in range, part replaces whole, and the congestion levels in compartment are predicted by single camera.
Measured precision is more than 90%.
The present invention uses box to mark human body face/head method compared with prior art, marks human body face by red dot
Portion/head convenience and high-efficiency, and in Dense crowd, red dot can accurately mark out human body face/head of partial occlusion
Portion, this is that effect be not achieved is marked using box.
Detailed description of the invention
Fig. 1 is a kind of process signal of passenger flow crowding calculation method based on rail transit train provided by the invention
Figure;
Fig. 2 is a kind of structural representation of the passenger flow crowding computing device based on rail transit train provided by the invention
Figure.
Specific embodiment
In being described below, for illustration and not for limitation, propose such as specific system structure, interface, technology it
The detail of class, to understand thoroughly the present invention.However, it will be clear to one skilled in the art that there is no these specific
The present invention also may be implemented in the other embodiments of details.In other situations, omit to well-known device, circuit and
The detailed description of method, in case unnecessary details interferes description of the invention.
Fig. 1 is a kind of process signal of passenger flow crowding calculation method based on rail transit train provided by the invention
Figure.
As shown in Figure 1, the present invention provides a kind of passenger flow crowding calculation method based on rail transit train, the side
Method includes:
S1: acquisition video simultaneously carries out sub-frame processing to collected video, compartment image is obtained, by compartment figure described in multiple
As separately constituting training set, verifying collection and test set;
S2: standee in all compartment images is labeled;
S3: being trained to the depth convolutional neural networks model importing training set and imports verifying collection verifying convolution mind
Accuracy through network model, the convolutional neural networks model after obtaining final training;
S4: corresponding region of interest ROI is arranged according to the environment in compartment, and according to the ROI and the training set
Convolutional neural networks model afterwards, to collected video described in the convolutional neural networks mode input after the training or institute
State the image in individual compartment in test set;
S5: counting the passenger marked in the ROI, and the grade of intensity of passenger flow is divided according to count results.
In some illustrative embodiments, collected video described in the S1 is different time sections, different compartment institutes
The video of acquisition.
In some illustrative embodiments, the training set in the S1 is made of 1800 compartment images, institute
It states verifying collection to be made of 200 compartment images, the test set is made of 800 compartment images.
The present invention successively different time sections difference compartment acquire totally 100 be all 5 minutes video sets, then pass through
Video takes out frame and has screened 2800 compartment images from 100 videos, wherein 1800 refreshing for subsequent training as training set
Through network model, 200 as verifying collects accuracy for verifying neural network, and 800 are used to test mind as test sets
Whether the function through network meets desired effect.
In some illustrative embodiments, standee in all compartment images is labeled in the S2
Process is labeled human body face/head by the way of red dot mark.
In S2, the present invention uses box to mark human body face/head method compared with prior art, is marked by red dot
Human body face/head convenience and high-efficiency, and in Dense crowd, red dot can accurately mark out the human body face of partial occlusion
Portion/head, this is that effect be not achieved is marked using box.
In some illustrative embodiments, the grade of the intensity of passenger flow is divided into 1 grade, 2 grades and 3 grades total three-level or 1 grade, 2
The total level Four of grade, 3 grades and 4 grades, wherein 1 grade of expression compartment is loose, general crowded, the 3 grades of very crowded and 4 grades of tables of expression of 2 grades of expressions
Show that crowding is moderate.
More than one single camera in each section compartment is illustrated with regard to Changsha subway, and there are four cameras for each compartment, therefore
Only differentiate that the degree of crowding in compartment is excessively unilateral by single camera, the present invention differentiates compartment using comprehensive four cameras
The degree of crowding.Concrete model is following (four, each compartment camera, be divided into four congestion levels):
The number that four cameras are detected by human body face/head density detection algorithm, according to the number detected
And grade threshold determines congestion levels, obtains four respective congestion levels of camera, according to 1,2,3,4 grades it is corresponding obtain 1,
2,3,4 points.The degree of crowding of final compartment is determined by score.
In view of (the multiple cameras in some region detect simultaneously for repetition detection of four cameras to compartment in single compartment
To human body face/head) and missing inspection region, it is difficult to accurately detect the ridership in entire compartment by four camera detections
Amount, therefore the present invention is by taking four camera detections, as one of the benchmark for differentiating congestion levels, to obtain vehicle to number mean value
After the mean value number of compartment as the grade threshold of step 1, corresponding mean value grade and score are obtained.
Seek vehicle-mounted congestion levels score S=0.2 × (S1+S2+S3+S4+S?), wherein S1、S2、S3、S4、S?Respectively four
Camera detection human body face/head institute's score and human body face/head mean value score.Gathering around for entire compartment is determined according to score
It squeezes grade: as S < 1.5, taking 1 grade;When 1.5≤S < 2.5,2 grades are taken;3 grades are taken when 2.5≤S < 3.5;4 are taken when S >=3.5
Grade.
About convolutional neural networks structure of the invention, the present invention includes six using six layers of convolutional network structure altogether
Layer convolutional layer, the structure of entire convolutional neural networks are input layer, the first convolutional layer, the first pond layer, the second convolutional layer, second
Pond layer, third convolutional layer, Volume Four lamination, the 5th convolutional layer, the 6th convolutional layer, output layer.What each convolutional layer successively used
It is the convolution kernel of 11 × 11,11 × 11,9 × 9,1 × 1,1 × 1,1 × 1 size, too small convolution kernel cannot extract people well
Honorable portion/header data carries out abstract-learning, and excessive convolution kernel increases network parameter, influences network detection efficiency and human body
Face/head detection precision is not improved.Two pond layers are used in network and carry out down-sampling, and pond window 2 × 2 will
Original image is long and width is reduced into original a quarter.And network structure end 1 × 1 full articulamentum of convolution nuclear subsitution, pole
The big parameter for reducing whole network structure, and there are also slightly promoted precision when with 1 × 1 convolution kernel than full articulamentum.
The algorithm of convolutional neural networks is applied in the classification of subway carriage density rating by the present invention, compared to traditional image procossing skill
Art, precision are enhanced.And the image for testing a frame 1280*720 pixel only needs the time of 30ms or so, actual measurement
Precision can achieve practical application more than 90%, be saved AT STATION by display directing passengers where in platform or compartment
It waits in compartment.
Convolutional neural networks update to obtain neurode in neural network by continuous iteration propagated forward and backpropagation
The optimized parameter of (convolution kernel), each parameter are also referred to as weight, and institute in entire convolutional neural networks model is obtained after network training
There is the set of weight to be referred to as weight model.
Different for the detection range of different compartment cameras, present invention employs the settings of region of interest ROI, often
The ROI range of a camera interception fixed size, finally obtains final congestion levels according to region area.It is led in image procossing
Domain, region of interest ROI are the image-regions selected from image, this region is image analysis emphasis of interest.
The region is drawn a circle to approve to be further processed.The target for needing to read using ROI delineation, it is possible to reduce the processing time increases
Precision.
For intensity of passenger flow detection is according to Changsha subway carriage:
1. acquiring a large amount of Changsha subway carriage images first, by video framing, 2,800 images are obtained, wherein
Two thousand sheets of training set, test set 800.
2. the human body head in pair all images is labeled, and generates the position of red dot and the mat of red dot number
(matlab) file, wherein red dot position is to do gaussian filtering in next step to prepare, and red dot number is statistics compartment number, each
Red dot represents a people, and an image is corresponding to generate a .mat file.
3. doing data by gaussian filtering existence .csv file with original image and mat file for convolutional neural networks training
Pretreatment.After csv file is handled through Gauss, each red dot generates a matrix with numerical value, the sum of data in a matrix
It is 1, remaining place data is then 0,
4. designing convolutional network model.
Other opposite convolutional networks, multiple row network or more depth network structure, network proposed by the present invention are simple and efficient,
In the case where precision obtains preferable situation, while there are preferable timeliness and lower memory again, on Nvidia 1080Ti video card
Detect the time that individual 1080p image only needs 80ms or so.
5 training networks
Data prediction has prepared the required data of training, while entire convolutional network design is completed, and next needs to instruct
Practice entire convolutional network, that is, obtain optimal parameter for whole network, enables to predict the number come and true people
Number error is minimum, meanwhile, the set of whole network parameter is called weight model.
Training process: first expanding data, and by each 1080*720 image, three 640*352 are cut into life at random
Small figure (had between small figure part repetition), it is 6000 images that 2000 thousand sheets training images, which just increase, takes 570,000 works
For training image collection, 30,000 are used to select as proof diagram image set the best initial weights model for training and, and such benefit is to increase
Training sample has been added to improve the robustness of entire model, while small image can improve training speed again, obtained weight model
.h5 file.
6. test
800 test image testing algorithm precision are taken, 91.6% can be reached, accuracy computation formula:Wherein M is the number of test image, pi、giFor the detection people of i-th image
Several and effective strength.
Test process: trained best weight value model .h5 file is imported into convolutional network model and activates network, so
It is tested test image importing network to obtain detection number p afterwards.
In addition, as shown in Fig. 2, the invention discloses a kind of passenger flow crowding computing device based on rail transit train,
Described device includes:
Framing module 100 is acquired, for acquiring video and carrying out sub-frame processing to collected video, obtains compartment figure
Picture separately constitutes training set, verifying collection and test set by compartment image described in multiple;
Labeling module 200, for being labeled to standee in all compartment images;
Training module 300 is tested for being trained and importing to the depth convolutional neural networks model importing training set
The accuracy of card collection verifying convolutional neural networks model, the convolutional neural networks model after obtaining final training;
Input module 400, for being arranged corresponding region of interest ROI according to the environment in compartment, and according to setting
Convolutional neural networks model after ROI and the training is acquired to described in the convolutional neural networks mode input after the training
To video or the test set in individual compartment image;
Diversity module 500 for counting to the passenger marked in the ROI, and divides intensity of passenger flow according to count results
Grade.
In some illustrative embodiments, the collected video of the acquisition framing module 100 is different time sections, different
Compartment video collected.
In some illustrative embodiments, the training set in the acquisition framing module 100 is by 1800 vehicles
Compartment image composition, the verifying collection are made of 200 compartment images, and the test set is by 800 compartment image groups
At.
Preferably, the grade of the intensity of passenger flow in the diversity module 500 is divided into 1 grade, 2 grades and 3 grades total three-level or 1 grade, 2
The total level Four of grade, 3 grades and 4 grades, wherein 1 grade of expression compartment is loose, general crowded, the 3 grades of very crowded and 4 grades of tables of expression of 2 grades of expressions
Show that crowding is moderate.
In some illustrative embodiments, the labeling module 200 is by the way of red dot mark to human body face/head
It is labeled.
Reader should be understood that in the description of this specification reference term " one embodiment ", " is shown " some embodiments "
The description of example ", " specific example " or " some examples " etc. mean specific features described in conjunction with this embodiment or example, structure,
Material or feature are included at least one embodiment or example of the invention.In the present specification, above-mentioned term is shown
The statement of meaning property need not be directed to identical embodiment or example.Moreover, particular features, structures, materials, or characteristics described
It may be combined in any suitable manner in any one or more of the embodiments or examples.In addition, without conflicting with each other, this
The technical staff in field can be by the spy of different embodiments or examples described in this specification and different embodiments or examples
Sign is combined.
Although the embodiments of the present invention has been shown and described above, it is to be understood that above-described embodiment is example
Property, it is not considered as limiting the invention, those skilled in the art within the scope of the invention can be to above-mentioned
Embodiment is changed, modifies, replacement and variant.
Claims (10)
1. a kind of passenger flow density detection method based on rail transit train, which is characterized in that the described method includes:
S1: acquisition video simultaneously carries out sub-frame processing to collected video, compartment image is obtained, by compartment image described in multiple point
It Zu Cheng not training set, verifying collection and test set;
S2: standee in all compartment images is labeled;
S3: being trained to the depth convolutional neural networks model importing training set and imports verifying collection verifies the convolution mind
Accuracy through network model, the convolutional neural networks model after obtaining final training;
S4: being arranged corresponding region of interest ROI according to the environment in compartment, and according to the ROI and the training set after
Convolutional neural networks model, to collected video described in the convolutional neural networks mode input after the training or the survey
Try the image in individual compartment concentrated;
S5: counting the passenger marked in the ROI, and the grade of intensity of passenger flow is divided according to count results.
2. the method according to claim 1, wherein collected video described in the S1 is different time
Section, different compartment videos collected.
3. the method according to claim 1, wherein the training set in the S1 is by 1800 compartments
Image composition, the verifying collection are made of 200 compartment images, and the test set is made of 800 compartment images.
4. the method according to claim 1, wherein the grade of the intensity of passenger flow is divided into 1 grade, 2 grades and 3 grades altogether
The total level Four of three-level or 1 grade, 2 grades, 3 grades and 4 grades, wherein 1 grade of expressions compartment is loose, 2 grades indicate general crowded, 3 grades of expressions very
Crowded and 4 grades of expression crowdings are moderate.
5. the method according to claim 1, wherein multiplying in the S2 to what is stood in all compartment images
The process that visitor is labeled is labeled human body face/head by the way of red dot mark.
6. a kind of intensity of passenger flow detection device based on rail transit train characterized by comprising
Framing module is acquired, for acquiring video and carrying out sub-frame processing to collected video, compartment image is obtained, by multiple
The compartment image separately constitutes training set, verifying collection and test set;
Labeling module, for being labeled to standee in all compartment images;
Training module collects verifying for importing the training set to depth convolutional neural networks model and being trained and import verifying
The accuracy of convolutional neural networks model, the convolutional neural networks model after obtaining final training;
Input module, for corresponding region of interest ROI to be arranged according to the environment in compartment, and according to the ROI and institute set
Convolutional neural networks model after stating training, to collected video described in the convolutional neural networks mode input after the training
Or the image in individual compartment in the test set;
Diversity module for counting to the passenger marked in the ROI, and divides the grade of intensity of passenger flow according to count results.
7. device according to claim 6, which is characterized in that when the collected video of the acquisition framing module is different
Between section, different compartment videos collected.
8. device according to claim 6, which is characterized in that the training set in the acquisition framing module is by 1800
The compartment Zhang Suoshu image composition, the verifying collection are made of 200 compartment images, and the test set is by 800 vehicles
Compartment image composition.
9. device according to claim 6, which is characterized in that the grade of the intensity of passenger flow in the diversity module is divided into 1
The total three-level of grade, 2 grades and 3 grades or 1 grade, 2 grades, 3 grades and 4 grades total level Four, wherein 1 grade of expression compartment is loose, 2 grades of expressions are generally gathered around
Squeeze, 3 grades of very crowded and 4 grades of expression crowdings of expression it is moderate.
10. device according to claim 6, which is characterized in that the labeling module is by the way of red dot mark to people
Honorable portion/head is labeled.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810917472.1A CN109190507A (en) | 2018-08-13 | 2018-08-13 | A kind of passenger flow crowding calculation method and device based on rail transit train |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810917472.1A CN109190507A (en) | 2018-08-13 | 2018-08-13 | A kind of passenger flow crowding calculation method and device based on rail transit train |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109190507A true CN109190507A (en) | 2019-01-11 |
Family
ID=64921655
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810917472.1A Pending CN109190507A (en) | 2018-08-13 | 2018-08-13 | A kind of passenger flow crowding calculation method and device based on rail transit train |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109190507A (en) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109815882A (en) * | 2019-01-21 | 2019-05-28 | 南京行者易智能交通科技有限公司 | A kind of subway carriage intensity of passenger flow monitoring system and method based on image recognition |
CN109919066A (en) * | 2019-02-27 | 2019-06-21 | 湖南信达通信息技术有限公司 | The method and apparatus of passenger's density anomaly in a kind of detection rail transit cars |
US20200186743A1 (en) * | 2016-11-24 | 2020-06-11 | Hanwha Techwin Co., Ltd. | Apparatus and method for displaying images and passenger density |
CN111582251A (en) * | 2020-06-15 | 2020-08-25 | 江苏航天大为科技股份有限公司 | Method for detecting passenger crowding degree of urban rail transit based on convolutional neural network |
CN111723739A (en) * | 2020-06-19 | 2020-09-29 | 艾索信息股份有限公司 | Congestion state monitoring method in bus based on convolutional neural network |
CN112241688A (en) * | 2020-09-24 | 2021-01-19 | 厦门卫星定位应用股份有限公司 | Carriage congestion degree detection method and system |
CN112699741A (en) * | 2020-12-10 | 2021-04-23 | 广州广电运通金融电子股份有限公司 | Method, system and equipment for calculating internal congestion degree of bus |
US11030465B1 (en) | 2019-12-01 | 2021-06-08 | Automotive Research & Testing Center | Method for analyzing number of people and system thereof |
CN113159004A (en) * | 2021-05-27 | 2021-07-23 | 苏州大学 | Passenger flow estimation method for rail transit carriage |
CN113408369A (en) * | 2021-05-31 | 2021-09-17 | 广州忘平信息科技有限公司 | Passenger flow detection method, system, device and medium based on convolutional neural network |
CN113553921A (en) * | 2021-07-02 | 2021-10-26 | 兰州交通大学 | Convolutional neural network-based subway carriage congestion degree identification method |
CN114217761A (en) * | 2021-12-16 | 2022-03-22 | 深圳市点创科技有限公司 | Screen energy-saving adjusting method based on passenger flow number and related device |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101944289A (en) * | 2010-08-31 | 2011-01-12 | 中山大学 | Subway passenger conducting management control system and method thereof |
-
2018
- 2018-08-13 CN CN201810917472.1A patent/CN109190507A/en active Pending
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101944289A (en) * | 2010-08-31 | 2011-01-12 | 中山大学 | Subway passenger conducting management control system and method thereof |
Non-Patent Citations (2)
Title |
---|
袁烨: "《面向智能监控的人流量计数及人群密度检测研究与实现》", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
马海军 等: "《基于卷积神经网络的监控视频人数统计算法》", 《安徽大学学报(自然科学版)》 * |
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20200186743A1 (en) * | 2016-11-24 | 2020-06-11 | Hanwha Techwin Co., Ltd. | Apparatus and method for displaying images and passenger density |
US10841654B2 (en) * | 2016-11-24 | 2020-11-17 | Hanwha Techwin Co., Ltd. | Apparatus and method for displaying images and passenger density |
CN109815882B (en) * | 2019-01-21 | 2020-04-10 | 南京行者易智能交通科技有限公司 | Subway carriage passenger flow density monitoring system and method based on image recognition |
CN109815882A (en) * | 2019-01-21 | 2019-05-28 | 南京行者易智能交通科技有限公司 | A kind of subway carriage intensity of passenger flow monitoring system and method based on image recognition |
CN109919066A (en) * | 2019-02-27 | 2019-06-21 | 湖南信达通信息技术有限公司 | The method and apparatus of passenger's density anomaly in a kind of detection rail transit cars |
CN109919066B (en) * | 2019-02-27 | 2021-05-25 | 湖南信达通信息技术有限公司 | Method and device for detecting density abnormality of passengers in rail transit carriage |
US11030465B1 (en) | 2019-12-01 | 2021-06-08 | Automotive Research & Testing Center | Method for analyzing number of people and system thereof |
CN111582251A (en) * | 2020-06-15 | 2020-08-25 | 江苏航天大为科技股份有限公司 | Method for detecting passenger crowding degree of urban rail transit based on convolutional neural network |
CN111723739A (en) * | 2020-06-19 | 2020-09-29 | 艾索信息股份有限公司 | Congestion state monitoring method in bus based on convolutional neural network |
CN112241688A (en) * | 2020-09-24 | 2021-01-19 | 厦门卫星定位应用股份有限公司 | Carriage congestion degree detection method and system |
CN112699741A (en) * | 2020-12-10 | 2021-04-23 | 广州广电运通金融电子股份有限公司 | Method, system and equipment for calculating internal congestion degree of bus |
CN113159004A (en) * | 2021-05-27 | 2021-07-23 | 苏州大学 | Passenger flow estimation method for rail transit carriage |
CN113408369A (en) * | 2021-05-31 | 2021-09-17 | 广州忘平信息科技有限公司 | Passenger flow detection method, system, device and medium based on convolutional neural network |
CN113553921A (en) * | 2021-07-02 | 2021-10-26 | 兰州交通大学 | Convolutional neural network-based subway carriage congestion degree identification method |
CN113553921B (en) * | 2021-07-02 | 2022-06-10 | 兰州交通大学 | Convolutional neural network-based subway carriage congestion degree identification method |
CN114217761A (en) * | 2021-12-16 | 2022-03-22 | 深圳市点创科技有限公司 | Screen energy-saving adjusting method based on passenger flow number and related device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109190507A (en) | A kind of passenger flow crowding calculation method and device based on rail transit train | |
CN106650913B (en) | A kind of vehicle density method of estimation based on depth convolutional neural networks | |
CN107103754A (en) | A kind of road traffic condition Forecasting Methodology and system | |
CN109977793A (en) | Trackside image pedestrian's dividing method based on mutative scale multiple features fusion convolutional network | |
CN109767427A (en) | The detection method of train rail fastener defect | |
CN102324016B (en) | Statistical method for high-density crowd flow | |
CN104881662B (en) | A kind of single image pedestrian detection method | |
CN110245644A (en) | A kind of unmanned plane image transmission tower lodging knowledge method for distinguishing based on deep learning | |
CN104268528B (en) | A kind of crowd massing method for detecting area and device | |
CN109446920B (en) | Method for detecting passenger crowding degree of urban rail transit based on convolutional neural network | |
CN109479206B (en) | Method for detecting pedestrian flow on road by using WI-FI probe | |
CN108334847A (en) | A kind of face identification method based on deep learning under real scene | |
CN109241858A (en) | A kind of passenger flow density detection method and device based on rail transit train | |
CN105160678A (en) | Convolutional-neural-network-based reference-free three-dimensional image quality evaluation method | |
CN106652483A (en) | Method for arranging traffic information detection points in local highway network by utilizing detection device | |
CN107945153A (en) | A kind of road surface crack detection method based on deep learning | |
CN107705560A (en) | A kind of congestion in road detection method for merging visual signature and convolutional neural networks | |
CN105513354A (en) | Video-based urban road traffic jam detecting system | |
CN106096561A (en) | Infrared pedestrian detection method based on image block degree of depth learning characteristic | |
CN108074244A (en) | A kind of safe city wagon flow statistical method for merging deep learning and Background difference | |
CN106772695B (en) | A kind of Along Railway wind speed forecasting method for merging many air measuring station measured datas | |
CN106991666A (en) | A kind of disease geo-radar image recognition methods suitable for many size pictorial informations | |
CN107038416A (en) | A kind of pedestrian detection method based on bianry image modified HOG features | |
CN112633149B (en) | Domain-adaptive foggy-day image target detection method and device | |
CN105513362B (en) | A kind of bus platform adjacent area bus running state evaluation verification method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190111 |