CN108629323A - A kind of integrated providing method of scenic spot tourist chain type trip - Google Patents
A kind of integrated providing method of scenic spot tourist chain type trip Download PDFInfo
- Publication number
- CN108629323A CN108629323A CN201810450485.2A CN201810450485A CN108629323A CN 108629323 A CN108629323 A CN 108629323A CN 201810450485 A CN201810450485 A CN 201810450485A CN 108629323 A CN108629323 A CN 108629323A
- Authority
- CN
- China
- Prior art keywords
- tourist
- time
- image
- sight spot
- spot
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 58
- 230000008569 process Effects 0.000 claims abstract description 24
- 238000004458 analytical method Methods 0.000 claims abstract description 20
- 238000005457 optimization Methods 0.000 claims abstract description 18
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 14
- 238000007418 data mining Methods 0.000 claims abstract description 11
- 238000011156 evaluation Methods 0.000 claims abstract description 5
- 238000012549 training Methods 0.000 claims description 30
- 230000006870 function Effects 0.000 claims description 18
- 238000013528 artificial neural network Methods 0.000 claims description 10
- 230000000694 effects Effects 0.000 claims description 10
- 238000013135 deep learning Methods 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 8
- 230000008859 change Effects 0.000 claims description 7
- 238000005516 engineering process Methods 0.000 claims description 5
- 230000002068 genetic effect Effects 0.000 claims description 5
- 230000000644 propagated effect Effects 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 239000012141 concentrate Substances 0.000 claims description 3
- 230000003247 decreasing effect Effects 0.000 claims description 3
- 230000005284 excitation Effects 0.000 claims description 3
- 230000004907 flux Effects 0.000 claims description 3
- 230000013016 learning Effects 0.000 claims description 3
- 238000007477 logistic regression Methods 0.000 claims description 3
- 238000013178 mathematical model Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000011176 pooling Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 230000003252 repetitive effect Effects 0.000 claims description 3
- 230000007704 transition Effects 0.000 claims description 3
- 230000005540 biological transmission Effects 0.000 claims description 2
- 230000035945 sensitivity Effects 0.000 claims description 2
- 230000000452 restraining effect Effects 0.000 claims 1
- 238000012800 visualization Methods 0.000 claims 1
- 238000011161 development Methods 0.000 description 6
- 238000007726 management method Methods 0.000 description 6
- 230000010354 integration Effects 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 238000013079 data visualisation Methods 0.000 description 2
- 241000287127 Passeridae Species 0.000 description 1
- 241000031708 Saprospiraceae Species 0.000 description 1
- 241000256856 Vespidae Species 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
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
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
- G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/12—Hotels or restaurants
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Economics (AREA)
- Data Mining & Analysis (AREA)
- Strategic Management (AREA)
- Biophysics (AREA)
- Tourism & Hospitality (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- General Business, Economics & Management (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Biomedical Technology (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Primary Health Care (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physiology (AREA)
- Genetics & Genomics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Image Analysis (AREA)
Abstract
A kind of integrated providing method of scenic spot tourist chain type trip, is related to tourist service technical field of software development.To solve the problems, such as that tourist lacks effective real time data guiding in the process in journey trip and generates not good enough Tourist Experience.This method includes that sight spot queuing crowding predicts that circuit of most preferably playing precisely pushes in real time, tourist's feedback data mining analysis.Specifically scenic spot information and the identification of each sight spot tourist quantity, and calculated, differentiate sight spot crowding according to evaluation index, and crowding is presented in by method for visualizing in map, gives tourist's intuitive judgment.Relatively reasonable route is precisely recommended using path optimization's algorithm according to the user's geographical coordinate and user demand data of system platform simultaneously, reduces user's queue waiting time to greatest extent, improves traffic experience.To form benign feedback, tourist is analyzed using data, therefrom obtains and is easy to happen the more serious sight spot for being lined up congestion and the preference etc. to different recommended routes.
Description
Technical field
Present invention relates particularly to a kind of scenic spot tourist chain type trip Integrated service methods, are related to tourist service software development
Technical field.
Background technology
With the continuous development of Chinese national economy and the continuous propulsion of urbanization process, China's Tourism Economy is also quick
Development.Vast territory and abundant resources in China simultaneously, and region span is big, possesses a large amount of, good, various, comprehensive tourist resources, and
Unique cultural features.However, due to each regional economic development imbalance so that the development of tourism resources is irregular with management level
It is uneven;At present due to lacking effective tourism guiding, path planning so that the tourist industry in China is widely present showing for " flocking together "
As.Some scenic spots are full, and you can catch sparrows on the doorstep at some scenic spots, is equally present in the same scenic spot, has a sight spot overstaffed, has
Sight spot but without tourist, form the domestic visitors unbalanced phenomena at scenic spot.
Certainly, it with the development of Tourism Economy, is more concerned with and plays itself, but have ignored and experienced on way, i.e. trip
Trip in journey.When in a city, due to lacking integrated tourist service chain so that tourist does not enjoy maximumlly
By the route in the city.When in scenic spot, during tourist's trip, due to lacking rational benign interaction with scenic spot information,
So that majority of populations select hot spot with usually blindly following the wind, and less consider the tourism wish of oneself and oneself preference
Characteristic tourism and traffic path.Simultaneously because whether the queuing at each sight spot is crowded indefinite, queue waiting time is also unknown
Really, in addition the geographical coordinate to sight spot is indefinite so that visit circuit and do not optimize, waste a large amount of time, also reduce out
Trip experience.The target that the resource that scenic spot is pursued always makes full use of lacks the support of real time data and intelligent algorithm, crowded sight spot
It can not obtain income from the crowd of queuing, and the idle sight spot wasting of resources.
At present under policy support, domestic online tourism flourishes, and online tourism marketing scale reaches within 2015
4737.7 hundred million yuan, increase by 49.6% on a year-on-year basis.It is that leading tourist market is quickly grown with internet, accounting reaches 70.2%.Trip
The consumption habit of trip user is changing, and mobile, personalization, chain type integration become new trend, also derive new
Product form:Such as location-based navigation Service, information publication and the travel information recommendation service based on tourist demand.But
Existing product currently on the market, such as boat trip in length and breadth, 12306 have the function of stroke reservation, way ox takes journey, winged pig etc.
Have the function of that route reservation, masses' comment, U.S. group etc. have journey consumption function, go where, hornet's nest etc. there is guide-book
Function.
But real-time journey information exchange is provided without any a software, at sight spot, queueing message is dynamically shown in real time
Show, optimal path intelligently pushing aspect is the market vacancy.Lack the integration to tourist communications resource in route, lacks to stroke
The concern of traffic experience lacks integrated chain type trip service and multidate information obtains in real time.
Invention content
The purpose of the present invention is providing a kind of integrated providing method of scenic spot tourist chain type trip, due to existing service product
Journey information exchange cannot be carried out, sight spot information Real time dynamic display cannot be carried out, and do not have optimal path intelligently pushing
Function, cause tourist during journey trip in the presence of lacking the guiding of effective real time data and to generate Tourist Experience not good enough
Problem.
The present invention adopts the technical scheme that solve above-mentioned technical problem:
A kind of integrated providing method of scenic spot tourist chain type trip, is predicted in real time the method includes sight spot queuing time
The step of the step of step, best circuit of playing precisely push and tourist's feedback data mining analysis;
One, sight spot is lined up the real-time prediction of crowding, queuing time:The tourist for being obtained transmission in real time using camera is lined up
Image, and image is lined up to tourist using video identification and is handled, recycle TensorFlow deep learning methods to identify
Queue number outside sight spot;
According to the actual conditions at sight spot establish congestion indication system combine above-mentioned queue number obtain sight spot queuing it is crowded
Degree;
It is lined up image by above-mentioned tourist and obtains tourist's queuing flow velocity, sight spot queuing time is obtained in conjunction with queue number;
The TensorFlow deep learnings method is:Image is compared using images match, description and identification and
Registration obtains the description of image symbolization to determine its classification by extracting the feature and correlation of image;Described image
With attempting to establish the geometrical correspondence between two pictures, its similar or different degree is measured;
In the identification of crowding, the base image that present situation tourist is lined up to image and standard is needed to compare, basis
Image includes the image acquired under different weather different time points, then calculates pixel size, i.e. gray scale, identifies waiting area
Tourist's quantity in domain, and carry out crowding evaluation according to certain index;Meanwhile it can also be arranged by video identification technology
The flow velocity that team tourist moves forward, and in this, as calculation basis, calculate estimated queuing time;
Two, best route of playing precisely pushes:It is described it is best play route precisely push be it is a kind of based on sequential access and
The path optimization that genetic algorithm is realized, process are:
First, using the sight spot queuing time, each sight spot tourist queuing time variation feelings within the scope of certain time are estimated
Condition;Secondly, the duration data of playing per capita for calling the scenic spot sight spots Nei Ge, if scenic spot defines the visit per capita at specific sight spot
Between, then the data are previously stored and are loaded by the step and called, if the visit duration at part sight spot is not limited, certainly by tourist
The estimated visit duration of row input;Finally, according to the geographical distribution at the position data of tourist and the scenic spot sight spots Nei Ge, tourist is obtained
The walking time for reaching a certain sight spot is calculated in conjunction with the data such as above-mentioned queuing time and the visiting time at each sight spot using heredity
Method optimizes the tour of tourist, so that its total queuing time is minimized, and then improve the tourism level of comfort of tourist;
Three, tourist's feedback data mining analysis, tourist's feedback data mining analysis are that traffic for tourism services big data
Application process, specially:User's service condition is analyzed using big data means, therefrom can get tourist like situation,
It is easy to happen the preference of the more serious sight spot for being lined up congestion and user to different recommended routes;
Meanwhile data are used according to usage history, passenger's activity time signature analysis, activity space feature point can be carried out
Analysis, geographical flux and flow direction signature analysis;It accordingly, can be to the quality of scenic spot service facility, the degree of route optimization and equilibrium
The problem of mode of passenger flow is analyzed, finds out its is precisely improved according to tourist demand, to make promotion scenic spot
Service quality, suction effect and the experience of playing of tourist.
Further, it during best route of playing precisely pushes, according to the optimization process, provides consideration and minimizes
The mathematical model of queuing time is as follows:
In formula:H --- it is time interval;
I --- it is range;
I --- for the point bit number in range, i.e. sight spot is numbered;
--- it is the estimated queuing time of each point (sight spot);
--- decision variable, tourist reach i points in the time range of h, and arrival takes 1, otherwise takes 0;
Constraints:
In formula:--- decision variable, tourist reach i points (representing sight spot) in the time range of h, and arrival takes 1, otherwise
Take 0;V indicates walking speed;Indicate that tourist reaches i point time range end values in the time range of h respectively;
Ti--- at the time of reaching i points;
di-1,i--- the distance at the sight spots i is walked to from the sight spots i-1;
tw,i-1--- in playing the time for the sight spots i-1;
The meaning of constraints is respectively:Each sight spot is only by way of primary;The total time for reaching each sight spot calculates;It reaches
Period judgement belonging to the time at each sight spot.
Further, during best route of playing precisely pushes, the value of I is 500 meters of radius.
Further, sight spot be lined up crowding, queuing time real-time prediction during, TensorFlow deep learnings
The model training of method is as follows:
Step 1, data loading phase
1) each pictures, reading process reserved folder information are read from image training dataset file, and are incited somebody to action
Picture is Tensor forms according to grayvalue transition;Each file represents an independent classification, all figures under this document presss from both sides
Piece belongs to same classification;
2) standard base image is stored classifiedly and is numbered, classification logotype is also converted into Tensor forms;
3) corresponding with classification logotype according to the affiliated folder information of every pictures;
4) queue minimum samples are determined, subregion are carried out to queue, and training sample picture is input to training nerve
Network;
Step 2, image pre-processing phase
1) picture can't directly be handled it after entering neural network, and neural network starts before extraction feature
Picture need to be pre-processed;Because training data concentrates the Pixel Dimensions of picture not of uniform size, to ensure input layer input
There is picture the size of unified size, picture the gray value square of 128X128 pixel sizes is cut in middle section or randomly
Battle array;
2) approximate white balance processing is carried out to image, it is sensitive to the dynamic range variation of training sample weakens pre-training model
Degree;
Step 3, feature extraction and forward-propagating stage
1) after image enters neural network, each layer convolutional layer uses different convolution kernels to image zooming-out different characteristic, defeated
Go out different characteristic images to lower layer;Convolutional layer characteristic extraction step is as follows:
Wherein l layers are convolutional layer, and l+1 layers are next layer,For l j-th of characteristic image of layer, * essence is volume
Product core k does convolution algorithm on the l-1 layers of related characteristic image of institute, then sums, and adds an offset parameter, takes ReLU
Obtain the process of final excitation value;Convolution results are inputted as lower layer, are propagated to deep layer along network;
2) pond layer is input with the characteristic image that convolutional layer exports, and defines pond range on the image in each channel
Maximum pond (max-pooling) is carried out with the region of 3*3;Maximum pond process is as follows:
Wherein l-1 layers are last layer, and l layers are pond layer,Be to l-1 layers carry out it is down-sampled
(maximum pond), and keep its Output Size identical as l layers of holding using same-padding;Characteristic image is grasped through maximum pondization
After work, size does not change, and is propagated to deep layer along network as next layer of input;
Step 4, back-propagation phase
1) loss function of training neural network uses multinomial logistic regression, during regularization, to all study
Variable application weight attenuation losses, the object function of model be ask intersect entropy loss and all weight attenuation terms and;
2) after the calculated prediction result of forward-propagating does cross entropy calculating with image category information, using under batch gradient
It drops algorithm and carries out backpropagation, update each layer parameter, and loss functions are reduced with this iteration;
Step 5, repetitive exercise stage
Procedure script can export the value of loss and the processing speed of last batch of data after every 10 step training process;Instruction
Linearity configuration, the variation tendency of Grad, the output valve of activation primitive and the Study rate parameter that loss changes during white silk are borrowed
Data visualization tool is helped to check;
Step 6, convergence maturity model
Model parameter in training process obtained by each step is periodically stored in check point file, when loss functions
It is no longer decreased obviously within a certain period of time, can determine whether to restrain;Judging that loss restrains and substantially after range of normal value, is terminating
The training of network model.
Advantages of the present invention and advantageous effect:On the one hand, it arranges in real time at the scenic spot that the method for the present invention Dock With Precision Position tourist most needs
Team's information requirement, crowding real-time query demand are that it precisely pushes integration that is personalized, focusing on experience according to user demand
Tourism trip service.On the other hand, user is accurate:Tourist can save queuing time, enjoy unimpeded route, administrative staff can
To optimize service level, for garden, using this technology, can be traveled order with orderly management garden, uniform Trip distribution, from length
It obtains tourism garden passenger flow steady in a long-term from the point of view of remote to increase, scenic spot facility utilization rate improves.On the whole, it is tourist and scenic spot
Management company provides more accurately scenic spot real time information, is conducive to the organization and management of the stream of people in scenic spot.
The invention enables tourists to realize interactive cooperation by real-time data and sight spot during journey is gone on a journey.This hair
It is bright that scenic spot information and each sight spot tourist quantity are identified, and calculated, differentiate sight spot crowding according to evaluation index, and will
Crowding is presented in by method for visualizing in map, and tourist's intuitive judgment is given.It is sat simultaneously according to user's geography of system platform
It is marked with and user demand data precisely recommends relatively reasonable route using path optimization's algorithm, reduce user row to greatest extent
Team's stand-by period improves traffic experience.To form benign feedback, tourist is analyzed using data, therefrom obtains and is easy hair
The raw more serious sight spot for being lined up congestion and the preference etc. to different recommended routes.The present invention effectively can avoid tourist from pricking
Heap simultaneously can reduce queue waiting time so that experience higher, scenic spot can also obtain benign development.
Description of the drawings
Fig. 1 is the schematic block diagram of the method for the present invention entirety, and Fig. 2 is the particular flow sheet of the method for the present invention.
Specific implementation mode
The present invention includes three modules, and sight spot queuing time predicts that circuit of most preferably playing precisely pushes in real time, tourist's feedback
Data mining analysis.Its method is as follows:
One, the real-time prediction module of sight spot queuing time
It is a kind of passenger flow identification based on image processing techniques that the sight spot, which is lined up crowding prediction in real time,.Utilize camera shooting
Head obtains image, and is handled image using video identification, and TensorFlow deep learning methods is recycled to identify scape
Queue number outside point, and then obtain crowded state.
Images match, description and identification are compared and are registrated to image, by point feature of system extraction image and mutually
Relationship obtains the description of image symbolization, then it is compared with model, to determine its classification.Images match attempts to establish two
Geometrical correspondence between picture measures its similar or different degree.
In the identification of crowding, the base image that present situation tourist is lined up to image and acquisition is needed to compare, basis
Image includes the image acquired under different weather different time points, then calculates pixel size, i.e. gray scale, identifies waiting area
Tourist's quantity in domain, and carry out crowded evaluation according to certain index.Meanwhile it can also be lined up by video identification technology
The flow velocity that tourist moves forward, and in this, as calculation basis, calculate estimated queuing time.
Main training pattern is as follows:
(1) data loading phase
1) each pictures, reading process reserved folder information are read from image training dataset file, and are incited somebody to action
Picture is Tensor forms according to grayvalue transition.Each file represents an independent classification, all figures under this document presss from both sides
Piece belongs to same classification;
2) original image is stored classifiedly and is numbered, classification logotype is also converted into Tensor forms;
3) corresponding with classification logotype according to the affiliated folder information of every pictures;
4) queue minimum samples are determined, subregion are carried out to queue, and training sample picture is input to trained network.
(2) image pre-processing phase
1) picture can't directly be handled it after entering network, and network starts to need to picture before extracting feature
It is pre-processed.Because training data concentrates the Pixel Dimensions of picture not of uniform size, to ensure that the picture of input layer input has
The size of unified size, picture are cut to the gray scale value matrix of 128X128 pixel sizes in middle section or randomly;
2) approximate white balance processing is carried out to image, it is sensitive to the dynamic range variation of training sample weakens pre-training model
Degree.
(3) feature extraction and forward-propagating stage
1) after image enters neural network, each layer convolutional layer uses different convolution kernels to image zooming-out different characteristic, defeated
Go out different characteristic images to lower layer.Convolutional layer characteristic extraction step is as follows:
Wherein l layers are convolutional layer, and l+1 layers are next layer,For l j-th of characteristic image of layer, * essence is volume
Product core k does convolution algorithm on the l-1 layers of related characteristic image of institute, then sums, and adds an offset parameter, takes ReLU
Obtain the process of final excitation value.Convolution results are inputted as lower layer, are propagated to deep layer along network;
2) pond layer is input with the characteristic image that convolutional layer exports, and defines pond range on the image in each channel
Maximum pond (max-pooling) is carried out with the region of 3X3.Maximum pond process is as follows:
Wherein l-1 layers are last layer, and l layers are pond layer,Be to l-1 layers carry out it is down-sampled
(maximum pond), and keep its Output Size identical as l layers of holding using same-padding.Characteristic image is grasped through maximum pondization
After work, size does not change, and propagates to deep layer along network as next layer of input
(4) back-propagation phase
1) loss function of this paper image recognitions submodule network uses multinomial logistic regression, during regularization,
To all Variable Learning application weight attenuation losses, the object function of model, which is asked, intersects entropy loss and all weight attenuation terms
With;
2) after the calculated prediction result of forward-propagating does cross entropy calculating with image category information, using under batch gradient
It drops algorithm and carries out backpropagation, update each layer parameter, and loss functions are reduced with this iteration.For the updated of each layer parameter
Journey is as follows:
Wherein, node sensitivity δ is change rate of the error to output:
ul=Wlxl-1+bl
(5) the repetitive exercise stage
Procedure script can export the value of loss and the processing speed of last batch of data after every 10 step training process.Instruction
The parameters such as linearity configuration, the variation tendency of Grad, the output valve of activation primitive and the learning rate that loss changes during white silk
It can be checked by data visualization tool.
(6) maturity model is restrained
Model parameter in training process obtained by each step is periodically stored in check point file, when loss functions
It is no longer decreased obviously within a certain period of time, can determine whether to restrain.Judging that loss restrains and substantially after range of normal value, is terminating
The training of network model.
Two, the best accurate pushing module of route of playing (uses closed-loop fashion, due to affirming no tourist's history number at the beginning
According to so first carrying out route push of most preferably playing.If the user for having received these routes will leave data from the background, then into
Row data mining so recycles, and obtained historical data is analyzed, re-optimization second step, i.e., route of most preferably playing is accurate
Push)
Precisely push is a kind of path optimization realized based on sequential access and genetic algorithm to the best route of playing.
Realize that the main target of the precise positioning optimization of user is to realize that user is optimal and system optimal first.User is optimal to be embodied in,
For user's individual, it can precisely recommend relatively reasonable route according to user demand for user, reduce to the maximum extent
The time that user waits in line reduces the probability that tourist retraces one's steps, and improves the experience that user plays.System optimal major embodiment
For scenic spot management person, passenger flow, the distribution of balanced passenger flow can reasonably organized to ensure one by management means
Under the premise of determining service quality, the tourist capacity that scenic spot can receive is improved.
Route optimization method is as follows:First, using the processing result image of the first module of this software, current each sight spot is calculated
The queue length of tourist and queuing time, and using the data processed result of this software third module, estimate certain time range
Interior each sight spot tourist queuing time situation of change;Secondly, the duration data of playing per capita at the scenic spot sight spots Nei Ge are called, if scenic spot is advised
Determine the visiting time per capita at specific sight spot, then the data has been previously stored and is loaded by the step and called, if part sight spot
Visit duration is not limited, then estimated visit duration is voluntarily inputted by tourist;Finally, according in the position data of tourist and scenic spot
The geographical distribution at each sight spot is lined up the data such as duration and the visit duration at each sight spot in conjunction with above-mentioned, utilizes genetic algorithm optimization
The tour of tourist makes its total queuing time minimize, and then improves the tourism level of comfort of tourist.
According to the above optimization method, it is as follows to provide the mathematical model for considering to minimize queuing time:
In formula:H --- it is time interval;
I --- it is range, such as 500 meters of radius;
I --- for the point bit number in range, i.e. sight spot is numbered;
--- it is the estimated queuing time of each point (sight spot);
--- decision variable, tourist reach i points in the time range of h, and arrival takes 1, otherwise takes 0.
Constraints:
In formula:--- decision variable, tourist reach i points in the time range of h, and arrival takes 1, otherwise takes 0;
TI--- at the time of reaching i points;
di-1,i--- the distance at the sight spots i is walked to from the sight spots i-1;
tw,i-1--- in playing the time for the sight spots i-1.
The meaning of constraints is respectively:Each sight spot is only by way of primary;The total time for reaching each sight spot calculates;It reaches
Period judgement belonging to the time at each sight spot.
Three, tourist's feedback data mining analysis module
Tourist's feedback data mining analysis is a kind of traffic for tourism service big data application process.Utilize big data
Means analyze user's service condition, can therefrom obtain tourist and like situation, are easy to happen more serious queuing congestion
The preference etc. of sight spot and user to different recommended routes.
Meanwhile data are used according to usage history, passenger's activity time signature analysis, activity space feature point can be carried out
Analysis, geographical flux and flow direction signature analysis.It accordingly, can be to the quality of scenic spot service facility, the degree of route optimization and equilibrium
The problem of mode of passenger flow is analyzed, finds out its is precisely improved according to tourist demand, to make promotion scenic spot
Service quality, suction effect and the experience of playing of tourist.
Embodiment
1, sight spot queuing time is predicted
(1) gray-scale map is processed the image into MATLAB, converts the image into matrix deep learning convolutional neural networks
Classification;
(2) picture pixels value size is calculated, queue number is obtained;
(3) estimated queuing time is calculated again according to service rate.
2, it most preferably plays path optimization
If co-existing in 10 sight spots in tourist neighboring area, the coordinate at tourist and each sight spot is generated using random generating mode
And the data such as the queuing duration at each sight spot, their spacing, concrete numerical value such as 1 institute of table are indicated using the Euclidean distance between each point
Show.Tourist's serial number 0, sight spot serial number 1-10, and set walking speed about 1m/s per capita.
1 point information table of table
Based on above-mentioned data the shortest road of tourist's queuing time is obtained using the itinerary of genetic algorithm optimization tourist
Line is 0-4-7-6-1-3-8-2-9-5-10, it is contemplated that total queuing time is 20 minutes.It is random as a comparison to generate a travelling road
Line:Estimated total queuing time of 0-9-5-6-2-7-4-3-10-1-8, this itinerary are 99 minutes.Through above-mentioned to score
Analysis, it is seen that the present invention effectively optimizes the touring line of passenger, reduces the queuing time during tourist plays, and then improve
Function passenger travel is experienced.
Claims (4)
1. a kind of integrated providing method of scenic spot tourist chain type trip, which is characterized in that the method includes sight spot queuing times
The step of the step of the step of prediction in real time, best circuit of playing precisely push and tourist's feedback data mining analysis;
One, sight spot is lined up the real-time prediction of crowding, queuing time:Obtain tourist's queuing figure of transmission in real time using camera
Picture, and image is lined up to tourist using video identification and is handled, recycle TensorFlow deep learning methods to identify scape
Queue number outside point;
The congestion indication system established according to the actual conditions at sight spot obtains sight spot in conjunction with above-mentioned queue number and is lined up crowding;
It is lined up image by above-mentioned tourist and obtains tourist's queuing flow velocity, sight spot queuing time is obtained in conjunction with queue number;
The TensorFlow deep learnings method is:Image is compared and is registrated using images match, description and identification,
By extracting the feature and correlation of image, the description of image symbolization is obtained to determine its classification;Described image matching examination
Figure establishes the geometrical correspondence between two pictures, measures its similar or different degree;
In the identification of crowding, the base image that present situation tourist is lined up to image and standard is needed to compare, base image
Including the image acquired under different weather different time points, pixel size, i.e. gray scale are then calculated, identifies queue area
Tourist's quantity, and carry out crowding evaluation according to certain index;Meanwhile it can also obtain being lined up trip by video identification technology
The flow velocity that visitor moves forward, and in this, as calculation basis, calculate estimated queuing time;
Two, best route of playing precisely pushes:Precisely push is a kind of based on sequential access and heredity to the best route of playing
The path optimization that algorithm is realized, process are:
First, using the sight spot queuing time, each sight spot tourist queuing time situation of change within the scope of certain time is estimated;Its
It is secondary, the duration data of playing per capita at the scenic spot sight spots Nei Ge are called, it, will if scenic spot defines the visiting time per capita at specific sight spot
The data are previously stored and are loaded by the step and called, if the visit duration at part sight spot is not limited, are voluntarily inputted by tourist
It is expected that visit duration;Finally, it according to the geographical distribution at the position data of tourist and the scenic spot sight spots Nei Ge, obtains tourist and reaches certain
The walking time at one sight spot utilizes genetic algorithm optimization in conjunction with the data such as above-mentioned queuing time and the visiting time at each sight spot
The tour of tourist makes its total queuing time minimize, and then improves the tourism level of comfort of tourist;
Three, tourist's feedback data mining analysis, tourist's feedback data mining analysis are that traffic for tourism services big data application
Process, specially:User's service condition is analyzed using big data means, tourist is therefrom can get and likes situation, is easy
The preference of the more serious sight spot for being lined up congestion and user to different recommended routes occurs;
Meanwhile data are used according to usage history, and it can carry out passenger's activity time signature analysis, activity space signature analysis,
Geographical flux and flow direction signature analysis;It accordingly, can be to the quality of scenic spot service facility, the degree of route optimization and balanced passenger flow
Mode the problem of being analyzed, finding out its, precisely improved according to tourist demand, to make the clothes at promotion scenic spot
Business quality, suction effect and the experience of playing of tourist.
2. a kind of integrated providing method of scenic spot tourist chain type trip according to claim 1, which is characterized in that best
During route of playing precisely pushes,
According to the optimization process, it is as follows to provide the mathematical model for considering to minimize queuing time:
In formula:H --- it is time interval;
I --- it is range;
I --- for the point bit number in range, i.e. sight spot is numbered;
--- it is the estimated queuing time of each point (sight spot);
--- decision variable, tourist reach i points in the time range of h, and arrival takes 1, otherwise takes 0;
Constraints:
In formula:--- decision variable, tourist reach i points in the time range of h, and arrival takes 1, otherwise takes 0;V indicates walking
Speed;Indicate that tourist reaches i point time range end values in the time range of h respectively;
Ti--- at the time of reaching i points;
di-1,i--- the distance at the sight spots i is walked to from the sight spots i-1;
tw,i-1--- in playing the time for the sight spots i-1;
The meaning of constraints is respectively:Each sight spot is only by way of primary;The total time for reaching each sight spot calculates;It reaches each
Period judgement belonging to the time at sight spot.
3. a kind of integrated providing method of scenic spot tourist chain type trip according to claim 1 or 2, which is characterized in that
During best route of playing precisely pushes, the value of I is 500 meters of radius.
4. a kind of integrated providing method of scenic spot tourist chain type trip according to claim 3, which is characterized in that at sight spot
Be lined up crowding, queuing time it is real-time predict during, the model training of TensorFlow deep learning methods is as follows:
Step 1, data loading phase
1) each pictures are read from image training dataset file, reading process reserved folder information, and by picture
It is Tensor forms according to grayvalue transition;Each file represents an independent classification, all picture categories under this document presss from both sides
In same classification;
2) standard base image is stored classifiedly and is numbered, classification logotype is also converted into Tensor forms;
3) corresponding with classification logotype according to the affiliated folder information of every pictures;
4) queue minimum samples are determined, subregion are carried out to queue, and training sample picture is input to trained neural network;
Step 2, image pre-processing phase
1) picture enters after neural network and can't directly handle it, and neural network starts to need before extracting feature pair
Picture is pre-processed;Because training data concentrates the Pixel Dimensions of picture not of uniform size, to ensure the picture of input layer input
Size with unified size, picture are cut to the gray scale value matrix of 128X128 pixel sizes in middle section or randomly;
2) approximate white balance processing is carried out to image, weakens dynamic range change sensitivity of the pre-training model to training sample;
Step 3, feature extraction and forward-propagating stage
1) after image enters neural network, each layer convolutional layer is using different convolution kernels to image zooming-out different characteristic, and output is not
With characteristic image to lower layer;Convolutional layer characteristic extraction step is as follows:
Wherein l layers are convolutional layer, and l+1 layers are next layer,For l j-th of characteristic image of layer, * essence is convolution kernel k
Convolution algorithm is done on the l-1 layers of related characteristic image of institute, then is summed, an offset parameter is added, ReLU is taken to obtain
The process of final excitation value;Convolution results are inputted as lower layer, are propagated to deep layer along network;
2) pond layer is input with the characteristic image that convolutional layer exports, and defines pond range on the image in each channel with 3*
3 region carries out maximum pond (max-pooling);Maximum pond process is as follows:
Wherein l-1 layers are last layer, and l layers are pond layer,It is l-1 layers to be carried out down-sampled, and makes
Keep its Output Size identical as l layers of holding with same-padding;After the operation of maximum pondization, size does not occur characteristic image
Change, and is propagated to deep layer along network as next layer of input;
Step 4, back-propagation phase
1) loss function of training neural network uses multinomial logistic regression, during regularization, to all Variable Learnings
Using weight attenuation losses, the object function of model be ask intersect entropy loss and all weight attenuation terms and;
2) after the calculated prediction result of forward-propagating does cross entropy calculating with image category information, declined using batch gradient and calculated
Method carries out backpropagation, update each layer parameter, and reduces loss functions with this iteration;
Step 5, repetitive exercise stage
Procedure script can export the value of loss and the processing speed of last batch of data after every 10 step training process;It trained
Linearity configuration, the variation tendency of Grad, the output valve of activation primitive and the Study rate parameter that loss changes in journey are by number
It is checked according to visualization tool;
Step 6, convergence maturity model
Model parameter in training process obtained by each step is periodically stored in check point file, when loss functions are one
It is no longer decreased obviously in fixing time, is judged as restraining;Judging that loss restrains and substantially after range of normal value, is terminating network
The training of model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810450485.2A CN108629323B (en) | 2018-05-11 | 2018-05-11 | Chain type travel integrated providing method for scenic spot tourists |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810450485.2A CN108629323B (en) | 2018-05-11 | 2018-05-11 | Chain type travel integrated providing method for scenic spot tourists |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108629323A true CN108629323A (en) | 2018-10-09 |
CN108629323B CN108629323B (en) | 2021-10-01 |
Family
ID=63692718
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810450485.2A Active CN108629323B (en) | 2018-05-11 | 2018-05-11 | Chain type travel integrated providing method for scenic spot tourists |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108629323B (en) |
Cited By (23)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109341693A (en) * | 2018-11-06 | 2019-02-15 | 大国创新智能科技(东莞)有限公司 | Public place of entertainment air navigation aid and system based on big data and deep learning |
CN109872145A (en) * | 2019-01-29 | 2019-06-11 | 中智沃达(北京)电力科技有限公司 | Charging field intelligent management system based on CNN convolutional neural networks |
CN110148071A (en) * | 2019-04-12 | 2019-08-20 | 深圳市轱辘汽车维修技术有限公司 | A kind of scenic spot reminding method, device and terminal device |
CN110427571A (en) * | 2019-07-13 | 2019-11-08 | 恒大智慧科技有限公司 | A kind of recommending scenery spot method, computer equipment and readable storage medium storing program for executing |
CN110941776A (en) * | 2020-01-03 | 2020-03-31 | 成都中科大旗软件股份有限公司 | Scenic spot guide system based on mobile intelligent terminal |
CN111063091A (en) * | 2019-12-27 | 2020-04-24 | 惠州视维新技术有限公司 | Play yard management method, system, terminal equipment and storage medium |
CN111126715A (en) * | 2020-01-03 | 2020-05-08 | 成都中科大旗软件股份有限公司 | Scenic spot passenger flow volume management and control system |
CN111144652A (en) * | 2019-12-26 | 2020-05-12 | 浙江力石科技股份有限公司 | Tour comfort degree algorithm and trend prediction method, system and device |
CN111191150A (en) * | 2019-12-03 | 2020-05-22 | 恒大智慧科技有限公司 | Tour route recommendation method and device and storage medium |
CN111191149A (en) * | 2019-11-26 | 2020-05-22 | 恒大智慧科技有限公司 | Shortest time-based travel route recommendation method and device and storage medium |
CN111198903A (en) * | 2019-11-26 | 2020-05-26 | 恒大智慧科技有限公司 | Poor trip route recommendation method and device and storage medium |
CN111210047A (en) * | 2019-11-12 | 2020-05-29 | 恒大智慧科技有限公司 | Scenic spot service time estimation method, device and storage medium |
CN111428858A (en) * | 2020-03-04 | 2020-07-17 | 东软睿驰汽车技术(沈阳)有限公司 | Method and device for determining number of samples, electronic equipment and storage medium |
CN112200956A (en) * | 2020-09-27 | 2021-01-08 | 北京百度网讯科技有限公司 | Access control method, system, electronic device and storage medium |
CN112668375A (en) * | 2019-10-16 | 2021-04-16 | 广东轻工职业技术学院 | System and method for analyzing tourist distribution in scenic spot |
CN112824833A (en) * | 2019-11-20 | 2021-05-21 | 深圳云天励飞技术有限公司 | Route planning method and related product |
CN113077102A (en) * | 2021-04-16 | 2021-07-06 | 合肥工业大学 | Landscape route optimization method |
CN113298452A (en) * | 2020-06-01 | 2021-08-24 | 阿里巴巴集团控股有限公司 | Commodity inventory determining method and device, and commodity processing method and device |
CN113472902A (en) * | 2021-09-02 | 2021-10-01 | 环球数科集团有限公司 | Scenic spot visitor flow monitoring and auxiliary system based on cloud technology |
CN113536090A (en) * | 2021-07-07 | 2021-10-22 | 中国电子科技集团公司第五十四研究所 | Tourist group-following line people number management and control method |
CN113780613A (en) * | 2020-12-15 | 2021-12-10 | 京东城市(北京)数字科技有限公司 | Path recommendation method and device |
CN114330796A (en) * | 2022-03-04 | 2022-04-12 | 环球数科集团有限公司 | Scenic spot queuing time prediction method and device and computer equipment |
CN115470418A (en) * | 2022-09-15 | 2022-12-13 | 安徽大学 | Queuing point recommendation method and system based on unmanned aerial vehicle aerial photography |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130260791A1 (en) * | 2012-04-02 | 2013-10-03 | University of Washington Through Its Center For Commercialization | Travel pattern discovery using mobile device sensors |
CN103413509A (en) * | 2013-08-20 | 2013-11-27 | 苏州跨界软件科技有限公司 | Guide method based on geographical positions |
CN106485217A (en) * | 2016-09-30 | 2017-03-08 | 四川世纪云道科技有限公司 | A kind of method and system of identification visit scenic spot stream of people's saturation degree |
CN106934736A (en) * | 2017-03-27 | 2017-07-07 | 中国农业大学 | A kind of scenic spot guidance method and device based on positioning service |
CN107832894A (en) * | 2017-11-27 | 2018-03-23 | 浙江卓锐科技股份有限公司 | A kind of intelligent tour schedule planing method based on ant group algorithm |
-
2018
- 2018-05-11 CN CN201810450485.2A patent/CN108629323B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130260791A1 (en) * | 2012-04-02 | 2013-10-03 | University of Washington Through Its Center For Commercialization | Travel pattern discovery using mobile device sensors |
CN103413509A (en) * | 2013-08-20 | 2013-11-27 | 苏州跨界软件科技有限公司 | Guide method based on geographical positions |
CN106485217A (en) * | 2016-09-30 | 2017-03-08 | 四川世纪云道科技有限公司 | A kind of method and system of identification visit scenic spot stream of people's saturation degree |
CN106934736A (en) * | 2017-03-27 | 2017-07-07 | 中国农业大学 | A kind of scenic spot guidance method and device based on positioning service |
CN107832894A (en) * | 2017-11-27 | 2018-03-23 | 浙江卓锐科技股份有限公司 | A kind of intelligent tour schedule planing method based on ant group algorithm |
Non-Patent Citations (1)
Title |
---|
张笑白: "高峰期内旅游路线实时优化调度仿真研究", 《计算机仿真》 * |
Cited By (33)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109341693B (en) * | 2018-11-06 | 2021-02-05 | 大国创新智能科技(东莞)有限公司 | Entertainment place navigation method and system based on big data and deep learning |
CN109341693A (en) * | 2018-11-06 | 2019-02-15 | 大国创新智能科技(东莞)有限公司 | Public place of entertainment air navigation aid and system based on big data and deep learning |
CN109872145A (en) * | 2019-01-29 | 2019-06-11 | 中智沃达(北京)电力科技有限公司 | Charging field intelligent management system based on CNN convolutional neural networks |
CN109872145B (en) * | 2019-01-29 | 2020-12-08 | 中智沃达(北京)电力科技有限公司 | Charging field intelligent management system based on CNN convolutional neural network |
CN110148071A (en) * | 2019-04-12 | 2019-08-20 | 深圳市轱辘汽车维修技术有限公司 | A kind of scenic spot reminding method, device and terminal device |
CN110427571A (en) * | 2019-07-13 | 2019-11-08 | 恒大智慧科技有限公司 | A kind of recommending scenery spot method, computer equipment and readable storage medium storing program for executing |
CN112668375B (en) * | 2019-10-16 | 2023-07-07 | 广东轻工职业技术学院 | Tourist distribution analysis system and method in scenic spot |
CN112668375A (en) * | 2019-10-16 | 2021-04-16 | 广东轻工职业技术学院 | System and method for analyzing tourist distribution in scenic spot |
CN111210047A (en) * | 2019-11-12 | 2020-05-29 | 恒大智慧科技有限公司 | Scenic spot service time estimation method, device and storage medium |
CN112824833A (en) * | 2019-11-20 | 2021-05-21 | 深圳云天励飞技术有限公司 | Route planning method and related product |
CN112824833B (en) * | 2019-11-20 | 2024-03-22 | 深圳云天励飞技术有限公司 | Route planning method and related product |
CN111198903A (en) * | 2019-11-26 | 2020-05-26 | 恒大智慧科技有限公司 | Poor trip route recommendation method and device and storage medium |
CN111191149A (en) * | 2019-11-26 | 2020-05-22 | 恒大智慧科技有限公司 | Shortest time-based travel route recommendation method and device and storage medium |
CN111191150A (en) * | 2019-12-03 | 2020-05-22 | 恒大智慧科技有限公司 | Tour route recommendation method and device and storage medium |
CN111144652A (en) * | 2019-12-26 | 2020-05-12 | 浙江力石科技股份有限公司 | Tour comfort degree algorithm and trend prediction method, system and device |
CN111144652B (en) * | 2019-12-26 | 2023-08-08 | 浙江力石科技股份有限公司 | Tour comfort algorithm and trend prediction based method, system and device |
CN111063091A (en) * | 2019-12-27 | 2020-04-24 | 惠州视维新技术有限公司 | Play yard management method, system, terminal equipment and storage medium |
CN111126715B (en) * | 2020-01-03 | 2023-05-12 | 成都中科大旗软件股份有限公司 | Scenic spot passenger flow control system |
CN111126715A (en) * | 2020-01-03 | 2020-05-08 | 成都中科大旗软件股份有限公司 | Scenic spot passenger flow volume management and control system |
CN110941776A (en) * | 2020-01-03 | 2020-03-31 | 成都中科大旗软件股份有限公司 | Scenic spot guide system based on mobile intelligent terminal |
CN110941776B (en) * | 2020-01-03 | 2023-05-12 | 成都中科大旗软件股份有限公司 | Scenic spot guiding system based on mobile intelligent terminal |
CN111428858A (en) * | 2020-03-04 | 2020-07-17 | 东软睿驰汽车技术(沈阳)有限公司 | Method and device for determining number of samples, electronic equipment and storage medium |
CN113298452A (en) * | 2020-06-01 | 2021-08-24 | 阿里巴巴集团控股有限公司 | Commodity inventory determining method and device, and commodity processing method and device |
CN112200956A (en) * | 2020-09-27 | 2021-01-08 | 北京百度网讯科技有限公司 | Access control method, system, electronic device and storage medium |
CN113780613A (en) * | 2020-12-15 | 2021-12-10 | 京东城市(北京)数字科技有限公司 | Path recommendation method and device |
CN113077102A (en) * | 2021-04-16 | 2021-07-06 | 合肥工业大学 | Landscape route optimization method |
CN113077102B (en) * | 2021-04-16 | 2022-11-08 | 合肥工业大学 | Landscape route optimization method |
CN113536090A (en) * | 2021-07-07 | 2021-10-22 | 中国电子科技集团公司第五十四研究所 | Tourist group-following line people number management and control method |
CN113472902A (en) * | 2021-09-02 | 2021-10-01 | 环球数科集团有限公司 | Scenic spot visitor flow monitoring and auxiliary system based on cloud technology |
CN114330796B (en) * | 2022-03-04 | 2022-05-27 | 环球数科集团有限公司 | Scenic spot queuing time prediction method and device and computer equipment |
CN114330796A (en) * | 2022-03-04 | 2022-04-12 | 环球数科集团有限公司 | Scenic spot queuing time prediction method and device and computer equipment |
CN115470418A (en) * | 2022-09-15 | 2022-12-13 | 安徽大学 | Queuing point recommendation method and system based on unmanned aerial vehicle aerial photography |
CN115470418B (en) * | 2022-09-15 | 2023-07-25 | 安徽大学 | Queuing point recommendation method and system based on unmanned aerial vehicle aerial photography |
Also Published As
Publication number | Publication date |
---|---|
CN108629323B (en) | 2021-10-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108629323A (en) | A kind of integrated providing method of scenic spot tourist chain type trip | |
CN109858388A (en) | A kind of intelligent tourism management system | |
CN101010560B (en) | Information processing device and method, program, and information processing system | |
CN112074845A (en) | Deep reinforcement learning for optimizing car pooling strategies | |
CN107167156B (en) | A kind of multimode Trip chain preferred method and system towards integration trip | |
CN108133619A (en) | Method, apparatus, storage medium and the terminal device of Parking prediction | |
CN108151756A (en) | The travel time computation method and device of a kind of urban public tranlport system | |
CN107036617A (en) | The traffic path method and system for planning that a kind of taxi is combined with subway | |
CN108681586A (en) | Tourism route personalized recommendation method based on intelligent perception | |
CN109059934A (en) | Paths planning method, device, terminal and storage medium | |
CN108537691A (en) | A kind of region visit intelligent management system and method | |
CN105300398B (en) | The methods, devices and systems of gain location information | |
CN112668375B (en) | Tourist distribution analysis system and method in scenic spot | |
CN104636457B (en) | A kind of method and device of location finding cognition | |
CN113935547A (en) | Optimal-utility customized scenic spot route planning system | |
CN110472999A (en) | Passenger flow pattern analysis method and device based on subway and shared bicycle data | |
CN111859183A (en) | Method, system and device for recommending boarding points and storage medium | |
CN110245377A (en) | A kind of travel plan recommended method and recommender system | |
CN111737826B (en) | Rail transit automatic simulation modeling method and device based on reinforcement learning | |
CN111161545B (en) | Intersection region traffic parameter statistical method based on video | |
CN109472416B (en) | Indoor path planning method and device based on automatic road network data extraction, and client | |
CN117726369A (en) | Commercial area address selection method and system based on mobile phone signaling data | |
CN111898043B (en) | Urban travel route planning method | |
Castagna et al. | Demand-responsive rebalancing zone generation for reinforcement learning-based on-demand mobility | |
CN111241941A (en) | Public water-saving control method and system based on artificial intelligence |
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 |