CN110139067A - A kind of wild animal monitoring data management information system - Google Patents
A kind of wild animal monitoring data management information system Download PDFInfo
- Publication number
- CN110139067A CN110139067A CN201910241710.6A CN201910241710A CN110139067A CN 110139067 A CN110139067 A CN 110139067A CN 201910241710 A CN201910241710 A CN 201910241710A CN 110139067 A CN110139067 A CN 110139067A
- Authority
- CN
- China
- Prior art keywords
- wild animal
- information
- data
- monitoring
- wild
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/08—Network architectures or network communication protocols for network security for authentication of entities
- H04L63/0815—Network architectures or network communication protocols for network security for authentication of entities providing single-sign-on or federations
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/10—Network architectures or network communication protocols for network security for controlling access to devices or network resources
- H04L63/105—Multiple levels of security
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/02—Protocols based on web technology, e.g. hypertext transfer protocol [HTTP]
- H04L67/025—Protocols based on web technology, e.g. hypertext transfer protocol [HTTP] for remote control or remote monitoring of applications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
- H04L69/16—Implementation or adaptation of Internet protocol [IP], of transmission control protocol [TCP] or of user datagram protocol [UDP]
- H04L69/161—Implementation details of TCP/IP or UDP/IP stack architecture; Specification of modified or new header fields
- H04L69/162—Implementation details of TCP/IP or UDP/IP stack architecture; Specification of modified or new header fields involving adaptations of sockets based mechanisms
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
- H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Computer Security & Cryptography (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computer Hardware Design (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of wild animal monitoring data management information systems, including monitoring information, and subsystem, monitoring information management subsystem, Login Register subsystem and user management subsystem etc. is presented.The present invention establish one can understand wild animal comprehensively, in real time inhabit situation, species information, identify simultaneously classification storage management to monitoring information according to picture feature, statistical analysis is intelligently made according to geographical location information, environmental information, the wild animal monitoring data management information system that policy provides the data foundation of science can be monitored for wild animal.
Description
Technical field
The present invention relates to wild animal monitoring technical fields, particularly relate to a kind of wild animal monitoring data agrment information system
System.
Background technique
Wild animal inhabites as the indispensable component part in terrestrial ecosystem and maintains ecology with procreation
The balance and development of system.With the diversification of wild animal monitoring data, wild animal monitoring data need more convenient, high
The management of effect.The development of information technology improves the working efficiency of information management, so that data management is more efficiently, more just
It is prompt.The foundation of wild animal monitoring data management information system facilitate comprehensively in real time understand wild animal inhabit situation,
Species information provides reliable data for wild animal monitoring and supports.Wild animal monitoring data management information system can incite somebody to action
Wild animal monitoring, identification, classification, analysis are organically together in series, and form a closed loop, for the political affairs of subsequent wild animal monitoring
Plan direction provides reasonable data and supports.
Application has been obtained in terms of natural resources monitoring in management information system at present.But due to towards natural resources
The management information system of supervision has particularity, needs for the information feature of different natural resources supervision using different exploitations
Language and environment design different types of database table, to ensure tightness, safety and the execution of management information system
Efficiency.Therefore to develop towards wild animal monitoring management information system need comprehensive analysis wild animal monitor demand and
The characteristics of wild animal monitoring information.Management information system is mainly used in agriculture and forestry pair in terms of natural resources monitoring at this stage
Scalar information is monitored, and wild animal monitoring information is based on image information.
Summary of the invention
In view of this, it is an object of the invention to propose a kind of wild animal monitoring data management information system, Neng Gouquan
Face, in real time understanding wild animal inhabit situation, species information.
Based on a kind of above-mentioned purpose wild animal monitoring data management information system provided by the invention, the system comprises
Subsystem is presented in monitoring information, and the monitoring information is presented subsystem and includes:
Data communication module, for receiving the wild animal monitoring node information of wild animal monitoring device transmission and depositing
Storage；The wild animal monitoring device includes multiple node devices, and each node device, which is monitored the node, to be obtained
Obtain the wild animal monitoring node information；
Real-time monitoring information module, for obtaining the pictorial information in real time, to the wild animal in the pictorial information
It carries out identification classification and obtains wild animal classification information, the wild animal classification information is subjected to intelligence in real time and is shown；
Geographical location information module, for generating wild animal monitoring node according to the wild animal monitoring node information
Distribution map, the wild animal monitoring node distribution map haunt probability and to being unfavorable for for predicting the wild animal of each node
The node that wild animal inhabites carries out early warning.
Optionally, the wild animal in the pictorial information carries out identifying that classification includes:
Construct system identification model；
According to the historical data in the wild animal monitoring node information, the system identification model is trained,
Wild animal system identification model after being trained；The wild animal monitoring node information includes the temperature of each node device
Spend information, humidity information, Lighting information, location information, pictorial information；What the wild animal system identification model can identify
Information include wild animal information, wild animal gender information, wild animal the range of age information, current context information,
The survival condition information of wild animal under current environment；
Using the pictorial information obtained in real time as the input information of the wild animal system identification model, generate wild
Raw animal class information；The wild animal classification information includes wild animal information, wild animal gender information, wild
The survival condition information of wild animal under animal age range information, current context information, current environment.
Optionally, the building system identification model includes:
Sample image is input to full convolutional neural networks, lateral output layer is connected to the last one pond layer, obtains sample
The positioning of wild animal image-region in this image；
In the deeper short connection for establishing shiver blocking structure between shallower lateral output layer, and pass through combination different stage
Feature, Analysis On Multi-scale Features figure is supplied to each layer of network structure, while carrying out multi-channel feature fusion, realizes the open country
The Pixel-level segmentation in lively object image region；Using the wild animal image-region after segmentation as the defeated of fine grit classification
Enter, use two feature extractors respectively to the physical feeling of wild animal in the wild animal image-region positioned with
And feature extraction, obtain iamge description operator；
Operator, which is described, according to described image obtains the system identification model.
Optionally, further includes:
Two classification are done using independent logisitc in the full convolutional neural networks to handle the sample image,
Assuming that the input of logistic function is o=f (x；θ), wherein θ be network parameter；Y=h (o) is then exported, h refers to logistic
Activation primitive；
P (y=1 | x)=h (o) (1)
P (y=0 | x)=1-h (o) (2)
The loss function of fused layer is
To i-th of input oiAfter derivation
J=-log L=∑-y log h- (1-y) log (1-h) (4)
According to the derivation property of logistic function, have:
Wherein, hiOutput after as logistic activation, yiFor target.
Optionally, further includes: trained result is assessed using loss function, verifying collection and test set image, is adjusted
Whole network model parameter is repeatedly trained, and the system identification model is obtained, wherein the loss function includes:
Wherein, LclsIt indicates assessment classification cost, is determined by the corresponding probability of the u that really classifies and Lcls=-log Pu；Lbox
It assesses detection block and positions cost, more really classify corresponding Prediction Parameters tuThe difference for being v with true translation zooming parameter and
Optionally, described that wild animal monitoring node distribution map, packet are generated according to the wild animal monitoring node information
It includes:
According to the positional information, the location information of each node device is marked on map；
Obtain wild animal information described in each node device；
According to the collected temperature data of each node device, the humidity data, the photometric data, calculating should
The probability that haunts is compared by the probability that haunts of type wild animal with preset probability threshold value, to meeting the probability
The node of threshold value carries out prominent displaying on the map；
Environmental change is carried out according to the temperature data, the humidity data, the photometric data in a period of time to become
Gesture prediction will be unfavorable for the node that wild animal inhabites on the map and carry out early warning displaying.
Optionally, described according to the collected temperature data of each node device, the humidity data, the illumination
Data calculate the probability that haunts of the type wild animal, comprising:
Obtain the temperature data, the humidity data, the historical data in the photometric data；
Historical data in the temperature data, the humidity data, the photometric data is added as independent variable
Power processing, haunts probability as dependent variable for wild animal, will will have complete display without wild animal Image Acquisition as 0 value
Wild animal Image Acquisition constructs wild animal and haunts prediction model as 1 value；
Using ought for the previous period within the wild animal monitoring node information haunt prediction as the wild animal
The input value of model predicts the probability that haunts of the type wild animal.
Optionally, the temperature data according in a period of time, the humidity data, the photometric data carry out
Trend of Environmental Change prediction, comprising:
Obtain ought for the previous period in the temperature data, the humidity data, the photometric data；
To acquisition ought for the previous period in the temperature data, the humidity data, the photometric data carry out line
Property regression fit, obtain wild animal living environment predicted value；
Living environment threshold value of warning is preset according to the living environment that wild animal is suitable for；
Judge whether wild animal living environment predicted value meets the living environment threshold value of warning, if it is not, then described
It will be unfavorable for the node that wild animal inhabites on map and carry out early warning displaying.
Optionally, the wild animal in the pictorial information carries out identification classification and obtains wild animal classification letter
The wild animal classification information is carried out intelligence in real time and shown by breath, comprising:
It is shown after being classified using the wild animal system identification model to the pictorial information；The classification standard
Including wild animal information, wild animal gender information, wild animal the range of age information, current context information, current
The survival condition information of wild animal under environment；
According to the wild animal information judge the wild animal whether with endangered wildlife preset in system；
If so, being highlighted to the pictorial information with the wild animal.
Optionally, the system also includes monitoring information management subsystem, Login Register subsystem and user management
System, wherein
The monitoring information management subsystem includes:
Wild animal picture uploading module, for the local wild animal monitoring node information to be carried out Classification and Identification
After be uploaded to server；
Monitoring information modified module, exception information, error message for rendering, and can be to the exception information, described
Error message is manually adjusted, marked and is modified；The Login Register subsystem includes:
Login module, for determine user account whether be it is legal, then judge the permission of user if legal, make the user
Account accesses the system according to the permission of setting；
Registration module, for realizing the registration of user account and the storage of user account information；
Password recovery module is given for change for realizing the password of user account；
The user management subsystem includes:
User authority management module, for realizing the management of the operating right to user account；
Subscriber information management module, for realizing the management to user information；
User behavior logging modle, for realizing to user behavior record and inquiry.
From the above it can be seen that wild animal monitoring data management information system provided by the invention, to manage letter
Based on breath system, developed for the characteristics of wild animal monitoring.The wild animal monitoring data management information system
The function of monitoring information classification, identification, inquiry, analysis, monitoring, management may be implemented.
Detailed description of the invention
Fig. 1 is a kind of structure chart of wild animal monitoring data management information system of the embodiment of the present invention；
Fig. 2 is a kind of block schematic illustration of wild animal monitoring data management information system of the embodiment of the present invention；
Fig. 3 is the picture query logic chart of wild animal of embodiment of the present invention monitoring data management information system；
Fig. 4 is the Information Statistics logic chart of wild animal of embodiment of the present invention monitoring data management information system；
Fig. 5 is the picture upload logic figure of wild animal of embodiment of the present invention monitoring data management information system；
Fig. 6 is the login authentication logic chart of wild animal of embodiment of the present invention monitoring data management information system.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with specific embodiment, and reference
Attached drawing, the present invention is described in more detail.
It should be noted that all statements for using " first " and " second " are for differentiation two in the embodiment of the present invention
The non-equal entity of a same names or non-equal parameter, it is seen that " first " " second " only for the convenience of statement, does not answer
It is interpreted as the restriction to the embodiment of the present invention, subsequent embodiment no longer illustrates this one by one.
Fig. 1 is a kind of structure chart of wild animal monitoring data management information system of the embodiment of the present invention.Fig. 2 is the present invention
A kind of block schematic illustration of wild animal monitoring data management information system of embodiment.
In one embodiment of the invention, the system comprises monitoring informations, and subsystem is presented, and the monitoring information is in
Now subsystem includes:
Data communication module, for receiving the wild animal monitoring node information of wild animal monitoring device transmission and depositing
Storage；The wild animal monitoring device includes multiple node devices, and each node device, which is monitored the node, to be obtained
Obtain the wild animal monitoring node information.The wild animal monitoring node information includes the temperature letter of each node device
The transmission mode of breath, humidity information, Lighting information, location information, pictorial information, pictorial information is passed through using progressive transmission
Socket communication is realized, pictorial information is encoded by region of interesting extraction technology and is carried out to image information primary and secondary degree
Sequence carries out prioritised transmission to area-of-interest, and the data communication module of system receives after data according to coding strategy to monitoring
Image is reconstructed.
Real-time monitoring information module, for obtaining the pictorial information in real time, to the wild animal in the pictorial information
It carries out identification classification and obtains wild animal classification information, the wild animal classification information is subjected to intelligence in real time and is shown.
Geographical location information module, for generating wild animal monitoring node according to the wild animal monitoring node information
Distribution map, the wild animal monitoring node distribution map haunt probability and to being unfavorable for for predicting the wild animal of each node
The node that wild animal inhabites carries out early warning.
Further, in real-time monitoring information module, identification classification is carried out to the wild animal in the pictorial information
Step includes:
Step 101, system identification model is constructed.
Step 102, according to the historical data in the wild animal monitoring node information, to the system identification model into
Row training, the wild animal system identification model after being trained；The wild animal monitoring node information includes each node
Temperature information, humidity information, Lighting information, location information, the pictorial information of equipment；The wild animal system identification model energy
The information enough identified includes wild animal information, wild animal gender information, wild animal the range of age information, works as front ring
The survival condition information of wild animal under border information, current environment.
Step 103, believe the pictorial information obtained in real time as the input of the wild animal system identification model
Breath generates wild animal classification information；The wild animal classification information includes wild animal information, wild animal gender
The survival condition information of wild animal under information, wild animal the range of age information, current context information, current environment.
Preferably, the pictorial information obtained in real time can be the pictorial information of current time acquisition, or
The pictorial information obtained in nearest preset time period.The wild animal system identification model uses Fast-RCNN model structure
It builds, historical data is monitored according to wild animal, model is trained, obtain the identification model for being suitble to protection animal, identification model
Using picture collection data as inputting and exporting the fine granularities information such as wild animal type, gender, the range of age in picture, for reality
When monitoring information module present.
In some embodiments of the invention, the building system identification model includes:
Step 201, sample image is input to full convolutional neural networks, using whole-nested edge detector (HED,
Holistically-Nested Edge Detection) framework, lateral output layer is connected to the last one pond layer, obtains sample
The positioning of most significant information, that is, wild animal image-region in this image.
Optionally, two classification are done to described using independent logisitc (S type curve) in the full convolutional neural networks
Sample image is handled, and the multi-tag problem of overlapping can be preferably handled.Assuming that the input of logistic function is o=f
(x；θ), wherein θ be network parameter；Y=h (o) is then exported, h refers to logistic activation primitive or sigmoid (logarithm probability)
Function；
P (y=1 | x)=h (o) (1)
P (y=0 | x)=1-h (o) (2)
The loss function of fused layer is
To i-th of input oiAfter derivation
J=-log L=∑-y log h- (1-y) log (1-h) (4)
According to the derivation property of logistic function, have:
Wherein, hiOutput after as logistic activation, yiFor target (result of label).
Step 202, on the basis of conspicuousness positioning stage, the details first stage is carried out.Using a kind of top-down
Method in the deeper short connection for establishing shiver blocking structure between shallower lateral output layer, and passes through combination different stage
Relatively rich scale feature figure, is supplied to each layer of network structure by feature, while carrying out multi-channel feature fusion, realizes institute
State the Pixel-level segmentation of wild animal image-region.
Step 203, using the wild animal image-region after segmentation as the input of fine grit classification, using two spies
Sign extractor carries out positioning and feature extraction to the physical feeling of wild animal in the wild animal image-region respectively, obtains
To iamge description operator.
By the output of previous step, input of the wild animal image-region as fine grit classification after dividing after testing,
Proposed adoption bilinear model (bilinear models) designs two feature extractors, and two outputs are each positions of image
Apposition (outer product), then carry out pool, obtain final iamge description operator.The effect of network A is to wild
Animal bodies different parts are positioned, and network B is then the object space progress feature extraction for detecting to network A.
Two mutually coordinated effects of network complete two most important tasks during fine granularity image classification, wild animal part
The detection and feature extraction in region.
Step 204, operator is described according to described image and obtains the system identification model.
Training set image is trained using above-mentioned network model and Training strategy, and utilizes verifying collection and test set
Image assesses trained result, and adjustment network model parameter is repeatedly trained, and obtains optimal network model and carries out
It saves.
Specifically, further include: trained result is assessed using loss function, verifying collection and test set image, is adjusted
Whole network model parameter is repeatedly trained, and the system identification model is obtained, wherein the loss function includes:
Wherein, LclsIt indicates assessment classification cost, is determined by the corresponding probability of the u that really classifies and Lcls=-log Pu；Lbox
It assesses detection block and positions cost, more really classify corresponding Prediction Parameters tuThe difference for being v with true translation zooming parameter and
In a specific embodiment, using the storage address information of MySQL storage wild animal monitoring image, special
Under the guidance of industry wild animal scholar, the staking-out work of wild animal monitoring image is completed, variety classes in database are wild
Animal is stored under different paths.And wild animal monitoring image in database is divided by training set with the ratio of 6:2:2, is tested
Card collection, test set.
Tensorflow deep learning frame is built in server end, using python as programming language, firstly, all instructions
Practice image and readjust 224*224, to adapt to the input of convolutional neural networks topology.Then enhance training using data enhancing
Collect and overcomes the problems, such as overfitting.Using Random-Rotation, shearing, the methods of scaling and overturning extend to raw data set originally
5 times.
Convolutional neural networks are built using tensorflow, using area is suggested that network (RPN) is generated and suggested, each suggestion
Generate 300 suggestion windows.It will suggest that window is mapped to the last one CNN convolution Feature Mapping, and pass through area-of-interest
(RoI) pond layer generates the Feature Mapping of fixed size for each RoI.The pond RoI using maximum value pond method with carrying out processing feature
Area-of-interest in figure.This method avoid the feature extractions to the candidate region in single image with many overlapping regions
Compute repeatedly.Map operation executes convolution feature extraction and joint account.The complete forward direction that it will create image is transmitted, and from
The converting characteristic of each RoI, and shared positive transmitting network are extracted in the positive transmitting of acquisition.In the back-propagating stage, use
Stochastic gradient descent (SGD) algorithm is trained model parameter, using cross entropy loss function of classifying as optimization aim.
Training set image is trained using above-mentioned network model and Training strategy, and utilizes verifying collection and test set
Image assesses trained result, and adjustment network model parameter is repeatedly trained, and obtains optimal network model and carries out
It saves, it is automatic to call network model file when the wild animal monitoring image newly taken enters in database, to wild dynamic
Object is identified.
According to recognition result, image is separately deposited to the file where the species, if without it is any detect it is wild
These images of animal will be excluded automatically, finally count the number of species of every kind of wild animal, be completed wild animal and be known automatically
The function of other differential counting, the animal species to save the wild animals provide effective assessment foundation.
Fig. 3 is the picture query logic chart of wild animal of embodiment of the present invention monitoring data management information system.In this hair
In bright one embodiment, obtaining the process that picture inquires picture includes:
Step 1 obtains screening conditions, which includes space-time screening conditions and environmental screening condition；If selection
Space-time screening conditions then jump to step 2, and environmental screening condition is selected then to jump to step 5.
Step 2 selects picture according to the sequencing of time, can choose regular hour section, when can also skip
Between select, directly carry out in next step；
The node checked is wanted in step 3, selection, can select the node for needing to check according to the demand of user, can also be with
Node is not selected, is directly carried out in next step；
Screening conditions are arranged according to the demand of user in step 4, traverse image array according to the screening conditions of setting, or
Then meet the data of user demand, and final data are presented；
Step 5, user demand select corresponding environmental data, can not also select to be directly entered step 6；
Step 6 returns to environmental information into comparison database, return to data splitting and image is carried out final be in
It is existing；Or enter comparison database, image array is traversed, image data is returned, returns to combination array, finally carry out data is in
It is existing.
Further, in geographical location information module, wild animal is generated according to the wild animal monitoring node information
The step of monitoring node distribution map includes:
Step 301, according to the positional information, the location information of each node device is marked on map.
Step 302, wild animal information described in each node device is obtained.
Step 303, according to the collected temperature data of each node device, the humidity data, the illumination number
According to calculating the probability that haunts of the type wild animal, the probability that haunts be compared with preset probability threshold value, to satisfaction
The node of the probability threshold value carries out prominent displaying on the map.Preferably, exist to the node for meeting the probability threshold value
Prominent show is carried out on the map can show to be highlighted.
Step 304, ring is carried out according to the temperature data, the humidity data, the photometric data in a period of time
Border trend will be unfavorable for the node that wild animal inhabites on the map and carry out early warning displaying.It preferably, can be with
It carries out continuing early warning to the node that wild animal inhabites is unfavorable for, until the node, which is conducive to wild animal, inhabites.
By step 301-304, wild animal monitoring node distribution map is generated.
Further, described according to the collected temperature data of each node device, the humidity data, the illumination
Data calculate the probability that haunts of the type wild animal, comprising:
Step 401, the temperature data, the humidity data, the historical data in the photometric data are obtained.
Step 402, using the historical data in the temperature data, the humidity data, the photometric data as change certainly
Amount is weighted processing, and wild animal is haunted probability as dependent variable, will will be had without wild animal Image Acquisition as 0 value
Complete display wild animal Image Acquisition constructs wild animal and haunts prediction model as 1 value.
Step 403, using ought for the previous period within the wild animal monitoring node information as the wild animal
It haunts the input value of prediction model, predicts the probability that haunts of the type wild animal.
Further, the temperature data according in a period of time, the humidity data, the photometric data carry out
Trend of Environmental Change prediction, comprising:
Step 501, obtain ought for the previous period in the temperature data, the humidity data, the photometric data.
Step 502, to acquisition ought for the previous period in the temperature data, the humidity data, the illumination number
According to linear regression fit is carried out, wild animal living environment predicted value is obtained.
Step 503, living environment threshold value of warning is preset according to the living environment that wild animal is suitable for.
Step 504, judge whether wild animal living environment predicted value meets the living environment threshold value of warning, if it is not,
The node that wild animal inhabites will be then unfavorable on the map and carry out early warning displaying.
Further, the wild animal in the pictorial information carries out identification classification and obtains wild animal classification letter
The wild animal classification information is carried out intelligence in real time and shown by breath, comprising:
Step 601, it is shown after being classified using the wild animal system identification model to the pictorial information；It is described
Classification standard includes wild animal information, wild animal gender information, wild animal the range of age information, current environment letter
The survival condition information of wild animal under breath, current environment.
Step 602, according to the wild animal information judge the wild animal whether with it is preset in imminent danger in system
Wild animal.
Step 603, if so, being highlighted to the pictorial information with the wild animal.
In another embodiment of the present invention, the system also includes historical information enquiry module, the historical informations
Enquiry module is used to carry out multidimensional presentation to historical data.The historical data include data communication module receive it is wild
Animal monitoring nodal information, and the every statistical data obtained in the above-described embodiments, including wild animal classification information, open country
Lively object monitoring node distribution map etc..User can be to monitoring information according to title, time, geographical location, frequency in browser end
Rate carries out multidimensional information inquiry, statistics and downloading.
Fig. 4 is the Information Statistics logic chart of wild animal monitoring data management information system of the present invention of the embodiment of the present invention.
In one embodiment of the invention, the logical process counted to information includes:
Step 2 one selects picture according to the sequencing of time, can choose regular hour section, can also skip
Selection of time directly carries out in next step；
The node checked is wanted in step 2 two, selection, can select the node for needing to check according to the demand of user, can also
Not select node, directly carry out in next step；
Screening conditions are arranged according to the demand of user in step 2 three, filter out the data for meeting user demand, and selection is used
The data type that family needs, returned data combination, and final data are presented.
In some embodiments of the invention, the system also includes monitoring information management subsystem, the monitoring informations
Management subsystem includes:
Wild animal picture uploading module, for the local wild animal monitoring node information to be carried out Classification and Identification
After be uploaded to server.The wild animal data locally arranged can be carried out Classification and Identification by administrator, and be uploaded to cloud
End.
Fig. 5 is the picture upload logic figure of wild animal of embodiment of the present invention monitoring data management information system.In this hair
In bright one embodiment, the logic judgment relationship of wild animal picture uploading module includes:
Step 3 one opens picture upload interface；
Step 3 two adds picture name, adds shooting time, adds description information, carries out canonical filtering to picture, sentences
Break the legitimacy of the picture, the picture is then stored in system if legal, and be recorded in operation log, if it is illegal then to the figure
Piece carries out error prompting.
Step 3 three after opening picture upload interface, while judging whether the picture is post.If it is not, then directly more
The picture carries out error prompting；If so, the file type of the picture is further judged, if file type is legal, by the figure
Piece deposit database is stored in operation log simultaneously, then directly carries out error prompting if it is illegal.
Monitoring information modified module, exception information, error message, administrator can be to the abnormal letters for rendering
Breath, the error message are manually adjusted, marked and are modified.
In further embodiments, the system also includes Login Register subsystem, the Login Register subsystem includes:
Login module, for determine user account whether be it is legal, then judge the permission of user if legal, make the user
Account accesses the system according to the permission of setting.Login module is the unique passage that user enters system, determines whether conjunction
Method user determines if legitimate user and to user right, the power that user is endowed according to the user account
Limit the correlation function of Range Access system.
Registration module, for realizing the registration of user account and the storage of user account information.Registration module is common
The unique passage that user is registered, reserves user information and realization increases user information in database.
Password recovery module, the password for realizing the user account of ordinary user are given for change.
Fig. 6 is the login authentication logic chart of wild animal of embodiment of the present invention monitoring data management information system.In this hair
In bright one embodiment, the logic checking step of Login Register subsystem includes:
Step 4 one inputs user information, judges whether user name is empty, if otherwise carrying out in next step, otherwise prompt is used
Name in an account book cannot be sky；
Step 4 two judges that user name whether there is, and then carries out in next step, otherwise user name being prompted to be not present if it exists；
Step 4 three judges whether password is empty, if otherwise carrying out in next step, otherwise prompt cipher cannot be sky；
Step 4 four, judges whether identifying code matches, if then carrying out in next step, otherwise knowing from experience identifying code mistake；
Step 4 five, judges whether password matches, if so, carrying out in next step, otherwise prompting user name or password mistake；
Step 4 six, judges user right grade, and user information is stored in session, according to user right jump page,
Login system.
In further embodiments, the system also includes user management subsystem, the user management subsystem includes:
User authority management module, for realizing the management of the operating right to user account, administrator can be right
Corresponding ordinary user carries out the inquiry and modification of operating right.
Subscriber information management module, for realizing the management to user information, administrator can to user information into
Row inquiry, mark and modification.
User behavior logging modle, for realizing to user behavior record and inquiry, administrator can be to common
The behavior record of user is inquired, and influence of the human factor to management information system is understood.
The system structure of wild animal monitoring data management information system includes Cloud Server, database server and visitor
Family end browser, Cloud Server are used for the associated profile of storage system, and database server is received for storage system
Wild animal monitoring node information and the system generate all data information, client browser is for realizing user's
Sign-on access.
Wild animal monitoring data management information system of the present invention, based on management information system, for wild
The characteristics of animal monitoring, is developed, and is combined using Web technology with depth learning technology, and using B/S framework, user passes through mutual
Networking accesses to management information system；Monitoring information point may be implemented in the wild animal monitoring data management information system
Class, the function of identification, inquiry, analysis, monitoring, management；A kind of visual data supervising platform is realized, there is perfect number
According to analysis and prediction function；This system is developed according to the actual features that wild animal monitors, and function is supervised close to wild animal
Survey the demand of work, it is ensured that the timeliness of wild animal monitoring enhances the utilizability of wild animal monitoring data, improves
The working efficiency of wild animal monitoring；There is simultaneity factor stringent security permission to be arranged, and user rights at different levels are clear, guarantee
Data it is authentic and valid, ensure that system is safely and smoothly run.
It should be understood by those ordinary skilled in the art that: the discussion of any of the above embodiment is exemplary only, not
It is intended to imply that the scope of the present disclosure (including claim) is limited to these examples；Under thinking of the invention, above embodiments
Or can also be combined between the technical characteristic in different embodiments, step can be realized with random order, and be existed such as
Many other variations of the upper different aspect of the invention, for simplicity, they are not provided in details.
The embodiment of the present invention be intended to cover fall into all such replacements within the broad range of appended claims,
Modifications and variations.Therefore, all within the spirits and principles of the present invention, any omission, modification, equivalent replacement, the improvement made
Deng should all be included in the protection scope of the present invention.
Claims (10)
1. a kind of wild animal monitoring data management information system, which is characterized in that the system comprises monitoring informations, and son is presented
System, the monitoring information are presented subsystem and include:
Data communication module, for receiving the wild animal monitoring node information of wild animal monitoring device transmission and storing；Institute
Stating wild animal monitoring device includes multiple node devices, and each node device is monitored described in acquisition the node
Wild animal monitoring node information；
Real-time monitoring information module carries out the wild animal in the pictorial information for obtaining the pictorial information in real time
Identification classification obtains wild animal classification information, and the wild animal classification information is carried out intelligence in real time and is shown；
Geographical location information module, for generating the distribution of wild animal monitoring node according to the wild animal monitoring node information
Figure, the wild animal monitoring node distribution map haunt probability and wild to being unfavorable for for predicting the wild animal of each node
The node of animal habitat carries out early warning.
2. wild animal monitoring data management information system according to claim 1, which is characterized in that described to the figure
Wild animal in piece information carries out identification classification
Construct system identification model；
According to the historical data in the wild animal monitoring node information, the system identification model is trained, is obtained
Wild animal system identification model after training；The wild animal monitoring node information includes the temperature letter of each node device
Breath, humidity information, Lighting information, location information, pictorial information；The information that the wild animal system identification model can identify
Including wild animal information, wild animal gender information, wild animal the range of age information, current context information, current
The survival condition information of wild animal under environment；
Using the pictorial information obtained in real time as the input information of the wild animal system identification model, generate wild dynamic
Object classification information；The wild animal classification information includes wild animal information, wild animal gender information, wild animal
The survival condition information of wild animal under the range of age information, current context information, current environment.
3. wild animal monitoring data management information system according to claim 2, which is characterized in that the building system
Identification model includes:
Sample image is input to full convolutional neural networks, lateral output layer is connected to the last one pond layer, obtains sample graph
The positioning of wild animal image-region as in；
In the deeper short connection for establishing shiver blocking structure between shallower lateral output layer, and pass through the spy of combination different stage
Analysis On Multi-scale Features figure, is supplied to each layer of network structure, while carrying out multi-channel feature fusion by sign, is realized described wild dynamic
The Pixel-level in object image region is divided；Using the wild animal image-region after segmentation as the input of fine grit classification, adopt
Positioning and spy are carried out to the physical feeling of wild animal in the wild animal image-region respectively with two feature extractors
Sign is extracted, and iamge description operator is obtained；
Operator, which is described, according to described image obtains the system identification model.
4. wild animal monitoring data management information system according to claim 3, which is characterized in that further include:
Two classification are done using independent logisitc in the full convolutional neural networks to handle the sample image, it is assumed that
The input of logistic function is o=f (x；θ), wherein θ be network parameter；
Y=h (o) is then exported, h refers to logistic activation primitive；
P (y=1 | x)=h (o) (1)
P (y=0 | x)=1-h (o) (2)
The loss function of fused layer is
To i-th of input oiAfter derivation
J=-logL=∑-ylogh- (1-y) log (1-h) (4)
According to the derivation property of logistic function, have:
Wherein, hiOutput after as logistic activation, yiFor target.
5. wild animal monitoring data management information system according to claim 3, which is characterized in that further include: it uses
Loss function, verifying collection and test set image assess trained result, and adjustment network model parameter is repeatedly trained,
The system identification model is obtained, wherein the loss function includes:
Wherein, LclsIt indicates assessment classification cost, is determined by the corresponding probability of the u that really classifies and Lcls=-logPu；LboxAssessment inspection
A cost is confined in survey, and more really classify corresponding Prediction Parameters tuThe difference for being v with true translation zooming parameter and
6. wild animal monitoring data management information system according to claim 2, which is characterized in that described according to
Wild animal monitoring node information generates wild animal monitoring node distribution map, comprising:
According to the positional information, the location information of each node device is marked on map；
Obtain wild animal information described in each node device；
According to the collected temperature data of each node device, the humidity data, the photometric data, the type is calculated
The probability that haunts is compared by the probability that haunts of wild animal with preset probability threshold value, to meeting the probability threshold value
Node carry out prominent displaying on the map；
It is pre- according to the temperature data, the humidity data, the photometric data progress Trend of Environmental Change in a period of time
It surveys, the node that wild animal inhabites will be unfavorable on the map and carry out early warning displaying.
7. wild animal monitoring data management information system according to claim 3, which is characterized in that the basis is each
The collected temperature data of node device, the humidity data, the photometric data calculate going out for the type wild animal
No probability, comprising:
Obtain the temperature data, the humidity data, the historical data in the photometric data；
Historical data in the temperature data, the humidity data, the photometric data is weighted place as independent variable
Reason, haunts probability as dependent variable for wild animal, will will have complete display wild without wild animal Image Acquisition as 0 value
Animal painting acquisition is used as 1 value, and building wild animal haunts prediction model；
Using ought for the previous period within the wild animal monitoring node information haunt prediction model as the wild animal
Input value, predict the probability that haunts of the type wild animal.
8. wild animal monitoring data management information system according to claim 3, which is characterized in that described according to one section
The temperature data, the humidity data, the photometric data in time carry out Trend of Environmental Change prediction, comprising:
Obtain ought for the previous period in the temperature data, the humidity data, the photometric data；
To acquisition ought for the previous period in the temperature data, the humidity data, the photometric data linearly returned
Return fitting, obtains wild animal living environment predicted value；
Living environment threshold value of warning is preset according to the living environment that wild animal is suitable for；
Judge whether wild animal living environment predicted value meets the living environment threshold value of warning, if it is not, then in the map
On will be unfavorable for the node that wild animal inhabites and carry out early warning displaying.
9. wild animal monitoring data management information system according to claim 2, which is characterized in that described to the figure
Wild animal in piece information carries out identification classification and obtains wild animal classification information, and the wild animal classification information is carried out
Intelligence is shown in real time, comprising:
It is shown after being classified using the wild animal system identification model to the pictorial information；The classification standard includes
Wild animal information, wild animal gender information, wild animal the range of age information, current context information, current environment
The survival condition information of lower wild animal；
According to the wild animal information judge the wild animal whether with endangered wildlife preset in system；
If so, being highlighted to the pictorial information with the wild animal.
10. wild animal monitoring data management information system according to claim 1, which is characterized in that the system is also
Including monitoring information management subsystem, Login Register subsystem and user management subsystem, wherein
The monitoring information management subsystem includes:
Wild animal picture uploading module, after the local wild animal monitoring node information is carried out Classification and Identification
Reach server；
Monitoring information modified module, exception information, error message for rendering, and can be to the exception information, the mistake
Information is manually adjusted, marked and is modified；The Login Register subsystem includes:
Login module, for determine user account whether be it is legal, then judge the permission of user if legal, make the user account
The system is accessed according to the permission of setting；
Registration module, for realizing the registration of user account and the storage of user account information；
Password recovery module is given for change for realizing the password of user account；
The user management subsystem includes:
User authority management module, for realizing the management of the operating right to user account；
Subscriber information management module, for realizing the management to user information；
User behavior logging modle, for realizing to user behavior record and inquiry.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910241710.6A CN110139067B (en) | 2019-03-28 | 2019-03-28 | Wild animal monitoring data management information system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910241710.6A CN110139067B (en) | 2019-03-28 | 2019-03-28 | Wild animal monitoring data management information system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110139067A true CN110139067A (en) | 2019-08-16 |
| CN110139067B CN110139067B (en) | 2020-12-25 |
Family
ID=67568679
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910241710.6A Active CN110139067B (en) | 2019-03-28 | 2019-03-28 | Wild animal monitoring data management information system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110139067B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110598461A (en) * | 2019-09-27 | 2019-12-20 | 腾讯科技（深圳）有限公司 | Wild animal information management method, device, terminal, system and storage medium |
| CN110672162A (en) * | 2019-12-04 | 2020-01-10 | 中科星图(深圳)数字技术产业研发中心有限公司 | Animal environment monitoring method for natural protected area |
| CN110704646A (en) * | 2019-10-16 | 2020-01-17 | 支付宝(杭州)信息技术有限公司 | Method and device for establishing stored material file |
| CN111078958A (en) * | 2019-12-18 | 2020-04-28 | 广州澳盾智能科技有限公司 | Instant statistics and decision-making system for diversified biological data |
| CN113178060A (en) * | 2021-04-23 | 2021-07-27 | 知晓（北京）通信科技有限公司 | Wild animal AI detection method and detection system |
| CN113645401A (en) * | 2021-07-21 | 2021-11-12 | 甘肃祁连山国家级自然保护区管护中心(大熊猫祁连山国家公园甘肃省管理局张掖分局) | Wild animal living environment monitoring method based on Beidou Internet of things |
| TWI780378B (en) * | 2019-11-26 | 2022-10-11 | 臺中榮民總醫院 | System and method for detecting and classifying animal behavior |
| CN116248830A (en) * | 2022-12-17 | 2023-06-09 | 航天行云科技有限公司 | Wild animal identification method, terminal and system based on space-based Internet of things |
| CN117436533A (en) * | 2023-12-20 | 2024-01-23 | 贵州大学 | Species distribution monitoring method and device based on habitat data analysis |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN204578567U (en) * | 2015-04-20 | 2015-08-19 | 河南师范大学 | Based on the intelligent Safari Park management system of Zigbee |
| CN107770494A (en) * | 2017-10-27 | 2018-03-06 | 深圳市文立科技有限公司 | Wild animal monitoring system |
| CN107809462A (en) * | 2017-09-23 | 2018-03-16 | 南京律智诚专利技术开发有限公司 | A kind of wild animal quantity monitoring and control recording system |
| US20180137338A1 (en) * | 2016-11-16 | 2018-05-17 | The Governing Council Of The University Of Toronto | System and method for classifying and segmenting microscopy images with deep multiple instance learning |
| CN108090904A (en) * | 2018-01-03 | 2018-05-29 | 深圳北航新兴产业技术研究院 | A kind of medical image example dividing method and device |
| US20180240235A1 (en) * | 2017-02-23 | 2018-08-23 | Zebra Medical Vision Ltd. | Convolutional neural network for segmentation of medical anatomical images |
| CN108510000A (en) * | 2018-03-30 | 2018-09-07 | 北京工商大学 | The detection and recognition methods of pedestrian's fine granularity attribute under complex scene |
| CN108920643A (en) * | 2018-06-26 | 2018-11-30 | 大连理工大学 | Weight the fine granularity image retrieval algorithm of multiple features fusion |
| CN109002853A (en) * | 2018-07-18 | 2018-12-14 | 宁夏智启连山科技有限公司 | Crop specie combines recognition methods and device with pest and disease damage type |
-
2019
- 2019-03-28 CN CN201910241710.6A patent/CN110139067B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN204578567U (en) * | 2015-04-20 | 2015-08-19 | 河南师范大学 | Based on the intelligent Safari Park management system of Zigbee |
| US20180137338A1 (en) * | 2016-11-16 | 2018-05-17 | The Governing Council Of The University Of Toronto | System and method for classifying and segmenting microscopy images with deep multiple instance learning |
| US20180240235A1 (en) * | 2017-02-23 | 2018-08-23 | Zebra Medical Vision Ltd. | Convolutional neural network for segmentation of medical anatomical images |
| CN107809462A (en) * | 2017-09-23 | 2018-03-16 | 南京律智诚专利技术开发有限公司 | A kind of wild animal quantity monitoring and control recording system |
| CN107770494A (en) * | 2017-10-27 | 2018-03-06 | 深圳市文立科技有限公司 | Wild animal monitoring system |
| CN108090904A (en) * | 2018-01-03 | 2018-05-29 | 深圳北航新兴产业技术研究院 | A kind of medical image example dividing method and device |
| CN108510000A (en) * | 2018-03-30 | 2018-09-07 | 北京工商大学 | The detection and recognition methods of pedestrian's fine granularity attribute under complex scene |
| CN108920643A (en) * | 2018-06-26 | 2018-11-30 | 大连理工大学 | Weight the fine granularity image retrieval algorithm of multiple features fusion |
| CN109002853A (en) * | 2018-07-18 | 2018-12-14 | 宁夏智启连山科技有限公司 | Crop specie combines recognition methods and device with pest and disease damage type |
Non-Patent Citations (4)
| Title |
|---|
| QIBIN HOU等: "Deeply Supervised Salient Object Detection with Short Connections", 《2017 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION》 * |
| 周海涛等: "近30年松嫩平原西部地区丹顶鹤栖息地适宜性动态变化", 《生态学杂志》 * |
| 唐书培等: "基于MaxEnt 模型的赛罕乌拉国家级自然保护区斑羚", 《北京林业大学学报》 * |
| 李宏群: "陕西黄龙山自然保护区褐马鸡生境选择的研究", 《中国博士学位论文全文数据库 基础科学辑》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110598461A (en) * | 2019-09-27 | 2019-12-20 | 腾讯科技（深圳）有限公司 | Wild animal information management method, device, terminal, system and storage medium |
| CN110704646A (en) * | 2019-10-16 | 2020-01-17 | 支付宝(杭州)信息技术有限公司 | Method and device for establishing stored material file |
| TWI780378B (en) * | 2019-11-26 | 2022-10-11 | 臺中榮民總醫院 | System and method for detecting and classifying animal behavior |
| CN110672162A (en) * | 2019-12-04 | 2020-01-10 | 中科星图(深圳)数字技术产业研发中心有限公司 | Animal environment monitoring method for natural protected area |
| CN110672162B (en) * | 2019-12-04 | 2020-04-10 | 中科星图(深圳)数字技术产业研发中心有限公司 | Animal environment monitoring method for natural protected area |
| CN111078958A (en) * | 2019-12-18 | 2020-04-28 | 广州澳盾智能科技有限公司 | Instant statistics and decision-making system for diversified biological data |
| CN111078958B (en) * | 2019-12-18 | 2024-04-09 | 广州澳盾智能科技有限公司 | Instant statistics and decision system for diversified biological data |
| CN113178060A (en) * | 2021-04-23 | 2021-07-27 | 知晓（北京）通信科技有限公司 | Wild animal AI detection method and detection system |
| CN113645401A (en) * | 2021-07-21 | 2021-11-12 | 甘肃祁连山国家级自然保护区管护中心(大熊猫祁连山国家公园甘肃省管理局张掖分局) | Wild animal living environment monitoring method based on Beidou Internet of things |
| CN116248830A (en) * | 2022-12-17 | 2023-06-09 | 航天行云科技有限公司 | Wild animal identification method, terminal and system based on space-based Internet of things |
| CN117436533A (en) * | 2023-12-20 | 2024-01-23 | 贵州大学 | Species distribution monitoring method and device based on habitat data analysis |
| CN117436533B (en) * | 2023-12-20 | 2024-02-13 | 贵州大学 | Species distribution monitoring method and device based on habitat data analysis |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110139067B (en) | 2020-12-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110139067A (en) | A kind of wild animal monitoring data management information system | |
| CN108921051B (en) | Pedestrian attribute identification network and technology based on cyclic neural network attention model | |
| Höferlin et al. | Inter-active learning of ad-hoc classifiers for video visual analytics | |
| CN106844614A (en) | A kind of floor plan functional area system for rapidly identifying | |
| CN110119446A (en) | Interpretable artificial intelligence | |
| CN105590099B (en) | A kind of more people's Activity recognition methods based on improvement convolutional neural networks | |
| CN107153817A (en) | Pedestrian's weight identification data mask method and device | |
| CN108460362A (en) | A kind of system and method for detection human body | |
| CN105208325B (en) | The land resources monitoring and early warning method captured and compare analysis is pinpointed based on image | |
| CN104134364B (en) | Real-time traffic sign identification method and system with self-learning capacity | |
| CN109063456A (en) | The safety detecting method and system of image-type identifying code | |
| Shi et al. | Amur tiger stripes: Individual identification based on deep convolutional neural network | |
| CN110689093A (en) | Image target fine classification method under complex scene | |
| CN110516791A (en) | A kind of vision answering method and system based on multiple attention | |
| CN110909195A (en) | Picture labeling method and device based on block chain, storage medium and server | |
| CN110929806A (en) | Picture processing method and device based on artificial intelligence and electronic equipment | |
| CN117390499A (en) | Be applied to multiple sample detecting system that food pesticide remained and detected | |
| CN113918526A (en) | Log processing method and device, computer equipment and storage medium | |
| CN111368911A (en) | Image classification method and device and computer readable storage medium | |
| CN111125290B (en) | Intelligent river patrol method and device based on river growth system and storage medium | |
| CN114648680A (en) | Training method, device, equipment, medium and program product of image recognition model | |
| CN117197726B (en) | Important personnel accurate management and control system and method | |
| CN116952654A (en) | Environment monitoring and early warning system for administrative supervision | |
| CN113807462B (en) | Network equipment fault cause positioning method and system based on AI | |
| CN116248346A (en) | Smart city-oriented CPS network security situation awareness establishing method and system |
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 |