CN110472492A - Target organism detection method, device, computer equipment and storage medium - Google Patents
Target organism detection method, device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN110472492A CN110472492A CN201910603620.7A CN201910603620A CN110472492A CN 110472492 A CN110472492 A CN 110472492A CN 201910603620 A CN201910603620 A CN 201910603620A CN 110472492 A CN110472492 A CN 110472492A
- Authority
- CN
- China
- Prior art keywords
- video
- target organism
- detected
- target
- monitoring end
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Image Analysis (AREA)
Abstract
This application involves machine learning, a kind of target organism detection method, device, computer equipment and storage medium are provided.The described method includes: sending enabled instruction to video monitoring end, video monitoring end starts according to enabled instruction to be monitored when reaching preset condition;The video information for obtaining the monitoring of video monitoring end, video frame to be detected is extracted from video information, obtains video pictures to be detected;The rgb value for calculating video pictures to be detected intercepts the picture region that rgb value in video pictures to be detected is greater than preset threshold, obtains picture region to be detected;Picture region to be detected is input in the target organism identification model trained and is identified, target organism recognition result is obtained;When target organism recognition result is there are when target organism, from the video data obtained in preset time range in video monitoring end, video data to be saved and returns to prompt messages to management terminal.It can be improved the accuracy of detection target organism using this method.
Description
Technical field
This application involves Internet technical fields, set more particularly to a kind of target organism detection method, device, computer
Standby and storage medium.
Background technique
With the raising of social standard of living, people increasingly focus on influence of the target organism to ambient enviroment, for example, In
After food and drink in kitchen, the sanitary condition in kitchen after food and drink can be monitored by monitoring objective biology such as mouse, fly etc., guarantee food
Product safety and sanitation.Currently, being all to be detected by artificial detection or by video monitoring using machine learning algorithm.Manually
Under the efficiency of detection is very low.By video monitoring using machine learning algorithm by larger objects such as detection human bodies, to mesh
The accurate lower of biological detection is marked, the situation of target organism can not be accurately reflected.
Summary of the invention
Based on this, it is necessary in view of the above technical problems, provide a kind of target organism inspection that can be improved accuracy of detection
Survey method, apparatus, computer equipment and storage medium.
A kind of target organism detection method, which comprises
When reaching preset condition, enabled instruction is sent to video monitoring end, video monitoring end starts according to enabled instruction
Monitoring;
The video information for obtaining the monitoring of video monitoring end, video frame to be detected is extracted from video information, is obtained to be detected
Video pictures;
The rgb value for calculating video pictures to be detected intercepts the picture that rgb value in video pictures to be detected is greater than preset threshold
Region obtains picture region to be detected;
Picture region to be detected is input in the target organism identification model trained and is identified, target organism is obtained
Recognition result;
When target organism recognition result is to obtain in preset time range there are when target organism from video monitoring end
Video data is saved and returns to prompt messages to management terminal by video data.
In one of the embodiments, when reaching preset condition, enabled instruction is sent to video monitoring end, comprising:
Monitoring picture is obtained, monitoring picture is input in the human testing model trained and is detected, is detected
As a result;
When testing result is not there is no human body, enabled instruction is sent to video monitoring end.
In one of the embodiments, when reaching preset condition, enabled instruction is sent to video monitoring end, comprising:
The current time in system is obtained, when the current time in system is consistent with preset time, sends and starts to video monitoring end
Instruction.
In one of the embodiments, by picture region to be detected be input in the target organism identification model trained into
Row identification, obtains target organism recognition result, comprising:
Picture is input in convolutional neural networks model by the picture that picture region to be detected is converted to target resolution
It is identified, obtains model output result;
Target organism classification, quantity and location information are obtained from model output result, according to target organism classification, quantity
Target organism recognition result is obtained with location information.
The generation step for the target organism identification model trained in one of the embodiments, comprising:
Obtain history monitor video and corresponding target organism markup information;
Using history monitor video as the input of convolutional neural networks algorithm, using target organism markup information as label into
Row training, when reaching preset condition, the target organism identification model trained.
In one of the embodiments, when target organism recognition result is there are when target organism, from video monitoring end
In end obtain preset time range in video data, by video data save and to management terminal return prompt messages it
Afterwards, further includes:
Alarm times are recorded, when alarm times are greater than preset times, obtain the corresponding target identification in video monitoring end, it will
Target identification and video data send supervisory terminal.
Method in one of the embodiments, further include:
The each video information and the corresponding target identification of each video information that multiple video monitoring ends return are obtained, and is counted
Calculate the cryptographic Hash of target identification;
Each video information is assigned to from node server according to cryptographic Hash, is known from node server by target organism
The video information of other model identification distribution obtains target organism recognition result;
It obtains from node server to the target organism recognition result of each video information, it will be corresponding according to target identification
Target organism recognition result returns to management terminal.
A kind of target organism detection device, device include:
Starting module is monitored, for sending enabled instruction, video monitoring end to video monitoring end when reaching preset condition
Started according to enabled instruction and is monitored;
Frame module is taken out, for obtaining the video information of video monitoring end monitoring, video to be detected is extracted from video information
Frame obtains video pictures to be detected;
Interception module intercepts rgb value in video pictures to be detected and is greater than for calculating the rgb value of video pictures to be detected
The picture region of preset threshold obtains picture region to be detected;
Identification module is known for picture region to be detected to be input in the target organism identification model trained
Not, target organism recognition result is obtained;
Video acquiring module, for when target organism recognition result be there are when target organism, obtained from video monitoring end
The video data in preset time range is taken, video data is saved and returns to prompt messages to management terminal.
A kind of computer equipment, including memory and processor, the memory are stored with computer program, the processing
Device performs the steps of when executing the computer program
When reaching preset condition, enabled instruction is sent to video monitoring end, video monitoring end starts according to enabled instruction
Monitoring;
The video information for obtaining the monitoring of video monitoring end, video frame to be detected is extracted from video information, is obtained to be detected
Video pictures;
The rgb value for calculating video pictures to be detected intercepts the picture that rgb value in video pictures to be detected is greater than preset threshold
Region obtains picture region to be detected;
Picture region to be detected is input in the target organism identification model trained and is identified, target organism is obtained
Recognition result;
When target organism recognition result is to obtain in preset time range there are when target organism from video monitoring end
Video data is saved and returns to prompt messages to management terminal by video data.
A kind of computer readable storage medium, is stored thereon with computer program, and the computer program is held by processor
It is performed the steps of when row
When reaching preset condition, enabled instruction is sent to video monitoring end, video monitoring end starts according to enabled instruction
Monitoring;
The video information for obtaining the monitoring of video monitoring end, video frame to be detected is extracted from video information, is obtained to be detected
Video pictures;
The rgb value for calculating video pictures to be detected intercepts the picture that rgb value in video pictures to be detected is greater than preset threshold
Region obtains picture region to be detected;
Picture region to be detected is input in the target organism identification model trained and is identified, target organism is obtained
Recognition result;
When target organism recognition result is to obtain in preset time range there are when target organism from video monitoring end
Video data is saved and returns to prompt messages to management terminal by video data.
Above-mentioned target organism detection method, device, computer equipment and storage medium, by when reaching preset condition,
Enabled instruction is sent to video monitoring end, video monitoring end starts according to enabled instruction to be monitored, and when meeting monitoring condition, is carried out
Monitoring can save video monitoring resource consumption.Then by obtaining the video information of video monitoring end monitoring, from video information
It is middle to extract video frame to be detected, obtain video pictures to be detected.The rgb value for calculating video pictures to be detected intercepts view to be detected
Rgb value is greater than the picture region of preset threshold in frequency picture, obtains picture region to be detected.Picture region to be detected is input to
It is identified in the target organism identification model trained, obtains target organism recognition result.I.e. by using picture to be detected
Region carries out the identification of target organism, can further improve the accuracy of recognition result.And when target organism identifies
It as a result is, from the video data obtained in video monitoring end in preset time range, video data to be protected there are when target organism
It deposits, from without saving whole monitor videos, saves storage resource.
Detailed description of the invention
Fig. 1 is the application scenario diagram of target organism detection method in one embodiment;
Fig. 2 is the flow diagram of target organism detection method in one embodiment;
Fig. 3 is the flow diagram that human testing is carried out in one embodiment;
Fig. 4 is to obtain the flow diagram of target organism recognition result in one embodiment;
Fig. 5 is the flow diagram of training objective bio-identification model in one embodiment;
Fig. 6 is the flow diagram of target organism detection method in another embodiment;
Fig. 7 is the structural block diagram of target organism detection device in one embodiment;
Fig. 8 is the internal structure chart of computer equipment in one embodiment.
Specific embodiment
It is with reference to the accompanying drawings and embodiments, right in order to which the objects, technical solutions and advantages of the application are more clearly understood
The application is further elaborated.It should be appreciated that specific embodiment described herein is only used to explain the application, not
For limiting the application.
Target organism detection method provided by the present application, can be applied in application environment as shown in Figure 1.Wherein, depending on
Frequency monitoring client 102 is communicated with server 104 by network, and server 104 is led to by network with management terminal 106
Letter.When reaching preset condition, server 104 sends enabled instruction to video monitoring end, and video monitoring end is according to enabled instruction
Starting monitoring.Server 104 obtains the video information that video monitoring end 102 monitors, and video to be detected is extracted from video information
Frame obtains video pictures to be detected.Server 104 calculates the rgb value of video pictures to be detected, intercepts in video pictures to be detected
Rgb value is greater than the picture region of preset threshold, obtains picture region to be detected;Server 104 inputs picture region to be detected
It is identified into the target organism identification model trained, obtains target organism recognition result;When target organism recognition result
For there are when target organism, server 104 will be regarded from the video data obtained in preset time range in video monitoring end 102
Frequency returns to prompt messages according to preservation and to management terminal 106.Wherein, terminal 102 can be, but not limited to be various individuals
Computer, laptop, smart phone, tablet computer and portable wearable device, server 104 can use independent clothes
The server cluster of business device either multiple servers composition is realized.
In one embodiment, as shown in Fig. 2, providing a kind of target organism detection method, it is applied to Fig. 1 in this way
In server for be illustrated, comprising the following steps:
S202 sends enabled instruction to video monitoring end, video monitoring end is according to enabled instruction when reaching preset condition
Starting monitoring.
Wherein, preset condition refer to pre-set video monitoring end starting condition, may include time conditions or
Person is the status condition in target monitoring region.For example, starting monitoring objective biology in video monitoring end at night 7.Video prison
Control end is for being monitored to obtain the photographic device of thermal imaging video, such as thermal imaging camera.
Specifically, server detects whether to meet whether the time conditions either target monitoring region pre-set is located
In unmanned state, when meeting preset condition for the moment, server sends enabled instruction to video monitoring end, video monitoring end according to
Enabled instruction starting monitoring, carrys out monitoring objective monitoring area.
S204, obtain video monitoring end monitoring video information, video frame to be detected is extracted from video information, obtain to
Detect video pictures.
Specifically, server gets the video information of video monitoring end monitoring, according to presetting from video information
Good time interval extracts video frame, obtains video pictures to be detected.Wherein, time interval can be set to 10 seconds etc..
S206 calculates the rgb value of video pictures to be detected, intercepts rgb value in video pictures to be detected and is greater than preset threshold
Picture region, obtain picture region to be detected.
Wherein, rgb value refers to the brightness value of the video pictures to be detected.Picture region to be detected is video pictures to be detected
The higher part of rgb value, a possibility that there are target organisms in the picture region to be detected is higher.Target organism refers to figure
It is smaller, pollution and the harmful biology of human body may cause to ambient enviroment health, for example, fly, moth buffalo gnat, termite, cockroach, mosquito
With mouse etc.,
Specifically, server calculates the rgb value of video pictures to be detected, and it is big that rgb value is intercepted from video pictures to be detected
In the picture region of preset threshold, each picture region to be detected is obtained.
Picture region to be detected is input in the target organism identification model trained and identifies, obtains mesh by S208
Mark biometric.
Wherein, target organism identification model be for identifying the target organism in picture region to be detected, for example, mouse,
Moth buffalo gnat, termite, cockroach, fly, mosquito etc., which is convolutional neural networks model.
Specifically, each picture region to be detected is input in the target organism identification model trained and is identified,
Obtain the corresponding target organism recognition result of each picture region to be detected.The target organism recognition result be used to indicate that this to
Picture region centering is detected with the presence or absence of target organism.
S210, when target organism recognition result is to obtain preset time model from video monitoring end there are when target organism
Interior video data is enclosed, video data is saved and returns to prompt messages to management terminal.
Specifically, when in the corresponding target organism recognition result of each picture region to be detected there are when target organism, clothes
Business device gets video data in preset time range from video monitoring end, and the video data of the preset time range just wraps
Include video frame to be detected.For example, available to two minutes before the video frame and rear two minutes video datas.Server will obtain
The video data obtained saves.Then prompt messages are returned to management terminal.The management terminal is to manage video prison
End is controlled, and the corresponding relationship of configured in the server management terminal and video monitoring end in advance, and the correspondence is closed
System's storage is into mapping table.
Above-mentioned target organism detection method, device, computer equipment and storage medium, by when reaching preset condition,
Enabled instruction is sent to video monitoring end, video monitoring end starts according to enabled instruction to be monitored, and when meeting monitoring condition, is carried out
Monitoring can save video monitoring resource consumption.Then by obtaining the video information of video monitoring end monitoring, from video information
It is middle to extract video frame to be detected, obtain video pictures to be detected.The rgb value for calculating video pictures to be detected intercepts view to be detected
Rgb value is greater than the picture region of preset threshold in frequency picture, obtains picture region to be detected.Picture region to be detected is input to
It is identified in the target organism identification model trained, obtains target organism recognition result.I.e. by using picture to be detected
Region carries out the identification of target organism, can further improve the accuracy of recognition result.And when target organism identifies
It as a result is, from the video data obtained in video monitoring end in preset time range, video data to be protected there are when target organism
It deposits, from without saving whole monitor videos, saves storage resource.
In one embodiment, as shown in figure 3, step S202 is sent that is, when reaching preset condition to video monitoring end
Enabled instruction, comprising steps of
S302 obtains monitoring picture, and monitoring picture is input in the human testing model trained and is detected, is obtained
Testing result.
Wherein, human testing model refers to that (target detection is calculated using Fast R-CNN algorithm according to existing somatic data
One kind of method) train obtained model.
Specifically, server can get the monitoring picture of target area from monitoring devices such as common cameras, will
The monitoring picture, which is directly inputted in the human testing model trained, carries out human testing, obtains testing result, the detection knot
Fruit is not including having there are human body and two kinds of human body.Monitoring picture can also be normalized to the picture of target resolution, by the figure
Piece, which is input in the human testing model trained, carries out human testing.
S304 sends enabled instruction to video monitoring end when testing result is not there is no human body.
Specifically, when testing result is not there is no human body, server sends enabled instruction to video monitoring end.
In the above-described embodiments, picture is monitored by obtaining, monitoring picture is input to the human testing model trained
In detected, obtain testing result, when testing result be not there is no human body when, to video monitoring end send enabled instruction.I.e.
Starting monitoring objective biology in video monitoring end when not there is no human body in target area, target organism detection accuracy can be improved simultaneously
And save monitoring resource.
In one embodiment, step S202 sends enabled instruction, packet to video monitoring end when reaching preset condition
Include step:
The current time in system is obtained, when the current time in system is consistent with preset time, sends and starts to video monitoring end
Instruction.
Specifically, server obtain current time in system point, judge the current time in system point whether with set it is pre-
If time point is consistent, when the current time in system is consistent with preset time, enabled instruction is sent to video monitoring end, according to setting
Preset time point open monitoring be monitored from without round-the-clock, save monitoring resource.For example, kitchen evening after food and beverage enterprise
10 points are come off duty, and setting at this time opens the target organism situation that video monitoring end monitors kitchen after the food and beverage enterprise at 10 points at night, to protect
Kitchen and bath is raw after demonstrate,proving food and beverage enterprise.
In one embodiment, as shown in figure 4, picture region to be detected, i.e., is input to the mesh trained by step S208
It is identified in mark bio-identification model, obtains target organism recognition result, comprising steps of
Picture region to be detected, is converted to the picture of target resolution by S402, and picture is input to convolutional neural networks
It is identified in model, obtains model output result.
Wherein, target resolution refers to the photo resolution of target organism identification model input, which identifies mould
Type is a kind of convolutional neural networks model.
Specifically, the conversion of resolution of picture region to be detected is to meet what target organism identification model inputted by server
Picture is input in convolutional neural networks model and identifies by the picture of target resolution, obtains model output result.Pass through
The conversion of resolution photo resolution of picture region to be detected can be improved to the accuracy of target organism identification.
S406 obtains target organism classification, quantity and location information from model output result, according to target organism class
Not, quantity and location information obtain target organism recognition result.
Wherein, model output result can be the picture that non-structured data have annotation results, the standard results
For showing identification as a result, may include target organism classification, quantity and location information etc., which refers to that target is raw
Object is in the position of monitoring area.Model output result is also possible to structural data, which includes multiple default
Attribute, each preset attribute have multiple fields, such as: mouse attribute includes: that whether there is or not field, mouse quantity for mouse under the attribute
Field and mouse location field etc..
Specifically, server analytic modell analytical model output result obtains target organism classification, quantity and location information, according to target
Category, quantity and location information obtain monitoring area with the presence or absence of target organism result and there are when target organism, target
Biomass, classification and location information.
In the above-described embodiments, it can identify to obtain the life of target present in monitoring area by convolutional neural networks model
Species not, quantity and location information, facilitate supervision and subsequent processing.
In one embodiment, the generation step for the target organism identification model trained, comprising steps of
S502 obtains history monitor video and corresponding target organism markup information.
Wherein, target organism markup information is for marking monitor video there are the video frame of target organism and there is no target
The video frame of biology.
Specifically, server get in history monitor video there are the corresponding video frame of target organism and with do not deposit target
The corresponding video frame of biology.
S504, using history monitor video as the input of convolutional neural networks algorithm, using target organism markup information as
Label is trained, when reaching preset condition, the target organism identification model trained.
Wherein, which can be YOLO (You Only Look Once) v3 algorithm, wherein activation
Function uses leaky RELU:Wherein, a is preset parameter.Preset condition is for judging that target organism identifies mould
Whether type training is completed, and the value that can be loss function reaches preset threshold, is also possible to that the number of iterations is trained to reach maximum and change
Generation number.Wherein loss function can side and error, consist of three parts: error of coordinate, IOU (hand over and compare) error and classification
Error.
Specifically, server, which will be present the corresponding video frame of target organism and not there is no the corresponding video frame of target organism, is
Corresponding target organism whether there is as label and is trained by the input of convolutional neural networks algorithm.Preset condition can be with
It is when the value of loss function reaches preset threshold, or training the number of iterations reaches maximum number of iterations, when reaching preset condition
When, model will be obtained as the target organism identification model trained.
In the above-described embodiments, by using convolutional Neural net previously according to history video data and corresponding markup information
The training of network algorithm obtains target organism identification model, can directly carry out using more convenient.
In one embodiment, when target organism recognition result is there are when target organism, from Video Monitoring Terminal
The video data in preset time range is obtained, after video data is saved and returns to prompt messages to management terminal,
Further include:
Alarm times are recorded, when alarm times are greater than preset times, obtain the corresponding target identification in video monitoring end, it will
Target identification and video data send supervisory terminal.
Wherein, target identification refers to the corresponding mark in region of video monitoring end monitoring, and server is previously stored with view
Frequency monitoring client and the corresponding tables of data of target identification.For example, kitchen and bath is raw after monitoring enterprise, then the target identification can be the enterprise
The address information in kitchen afterwards, mark of the enterprise etc..The video data refers to that server is saved in the video there are target organism.
Specifically, server records alarm times, when alarm times are more than the number pre-set within a certain period of time
When, server is searched video monitoring end more than the corresponding video monitoring end of preset times according to alarm times from tables of data and is corresponded to
Target identification, supervisory terminal is sent by target identification and video data, so that supervisory terminal is according to the target identification and view
Frequency is according to carrying out supervision processing.
In one embodiment, as shown in fig. 6, the target organism monitoring method further comprises the steps of:
S602 obtains each video information and the corresponding target mark of each video information that multiple video monitoring ends return
Know, and calculates the cryptographic Hash of target identification.
Wherein, target identification is to can be video number, video name etc. for identifying video information.
Specifically, when server connects multiple video monitoring ends, each view that multiple video monitoring ends return is got
Frequency information and the corresponding target identification of each video information, calculate the cryptographic Hash of target identification using hash algorithm at this time.
Each video information is assigned to from node server according to cryptographic Hash, passes through target from node server by S604
The video information of bio-identification model identification distribution obtains target organism recognition result.
Specifically, which is calculated calculated result, is divided each video information according to the calculated result
It is fitted on the progress target organism identification from node server, passes through the identification distribution of target organism identification model from node server
Video information obtains target organism recognition result.In one embodiment, encourage the corresponding video information of the target identification according to
It is distributed from node server according to load balancing.
S606 is obtained from node server to the target organism recognition result of each video information, will according to target identification
Corresponding target organism recognition result returns to management terminal.
Specifically, server is obtained from node server to the target organism recognition result of each video information, according to mesh
Corresponding target organism recognition result is returned to management terminal by mark mark, and carries out warning note.Wherein, which can
Be the corresponding video monitoring end of target identification administrative staff terminal, for example, the terminal etc. of the administrative staff of food and beverage enterprise
Deng.And supervisory terminal can be the terminal that all video monitoring ends are carried out with supervisor, for example, the administrative staff of hygiene department
Terminal etc..
In the above-described embodiments, when server connects multitude of video monitoring client, identification mission can be assigned to from section
Point server carries out target organism identification, subtracts the task processing pressure of server, guarantees the normal operation of server.
It should be understood that although each step in the flow chart of Fig. 2-6 is successively shown according to the instruction of arrow,
These steps are not that the inevitable sequence according to arrow instruction successively executes.Unless expressly stating otherwise herein, these steps
Execution there is no stringent sequences to limit, these steps can execute in other order.Moreover, at least one in Fig. 2-6
Part steps may include that perhaps these sub-steps of multiple stages or stage are not necessarily in synchronization to multiple sub-steps
Completion is executed, but can be executed at different times, the execution sequence in these sub-steps or stage is also not necessarily successively
It carries out, but can be at least part of the sub-step or stage of other steps or other steps in turn or alternately
It executes.
In one embodiment, as shown in fig. 7, providing a kind of target organism detection device 700, comprising: monitoring starting
Module 702 takes out frame module 704, interception module 706, identification module 708 and video acquiring module 710, in which:
Starting module 702 is monitored, for sending enabled instruction, video prison to video monitoring end when reaching preset condition
It controls end and monitoring is started according to enabled instruction;
Frame module 704 is taken out, for obtaining the video information of video monitoring end monitoring, view to be detected is extracted from video information
Frequency frame obtains video pictures to be detected;
It is big to intercept rgb value in video pictures to be detected for calculating the rgb value of video pictures to be detected for interception module 706
In the picture region of preset threshold, picture region to be detected is obtained;
Identification module 708 is carried out for picture region to be detected to be input in the target organism identification model trained
Identification, obtains target organism recognition result;
Video acquiring module 710, for when target organism recognition result is there are when target organism, from video monitoring end
The video data in preset time range is obtained, video data is saved and returns to prompt messages to management terminal.
In one embodiment, starting module 702 is monitored, comprising:
Monitoring picture is input in the human testing model trained by human detection module for obtaining monitoring picture
It is detected, obtains testing result;
Starting module, for sending enabled instruction to video monitoring end when testing result is not there is no human body.
In one embodiment, starting module 702 is monitored, comprising:
Time-obtaining module, for obtaining the current time in system, when the current time in system is consistent with preset time, to view
Frequency monitoring client sends enabled instruction.
In one embodiment, identification module 708, comprising:
Picture is input to convolution for picture region to be detected to be converted to the picture of target resolution by conversion module
It is identified in neural network model, obtains model output result;
Information obtains module, for obtaining target organism classification, quantity and location information from model output result, according to
Target organism classification, quantity and location information obtain target organism recognition result.
In one embodiment, target organism detection device 700, comprising:
Data obtaining module, for obtaining history monitor video and corresponding target organism markup information;
Model training module, for using history monitor video as the input of convolutional neural networks algorithm, by target organism
Markup information is trained as label, when reaching preset condition, the target organism identification model trained.
In one embodiment, target organism detection device 700, further includes:
Logging modle, when alarm times are greater than preset times, it is corresponding to obtain video monitoring end for recording alarm times
Target identification, target identification and video data are sent into supervisory terminal.
In one embodiment, target organism detection device 700, further includes:
Cryptographic Hash computing module, each video information and each video information returned for obtaining multiple video monitoring ends
Corresponding target identification, and calculate the cryptographic Hash of target identification;
Video distribution module takes for each video information to be assigned to from node server according to cryptographic Hash from node
Business device obtains target organism recognition result by the video information of target organism identification model identification distribution;
As a result module is obtained, for obtaining the target organism recognition result from node server to each video information, root
Corresponding target organism recognition result is returned into management terminal according to target identification.
Specific restriction about target organism detection device may refer to the above restriction to target organism detection method,
Details are not described herein.Modules in above-mentioned target organism detection device can be fully or partially through software, hardware and its group
It closes to realize.Above-mentioned each module can be embedded in the form of hardware or independently of in the processor in computer equipment, can also be with
Software form is stored in the memory in computer equipment, executes the corresponding behaviour of the above modules in order to which processor calls
Make.
In one embodiment, a kind of computer equipment is provided, which can be server, internal junction
Composition can be as shown in Figure 8.The computer equipment include by system bus connect processor, memory, network interface and
Database.Wherein, the processor of the computer equipment is for providing calculating and control ability.The memory packet of the computer equipment
Include non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system, computer program and data
Library.The built-in storage provides environment for the operation of operating system and computer program in non-volatile memory medium.The calculating
The database of machine equipment is for storing video data.The network interface of the computer equipment is used to pass through network with external terminal
Connection communication.To realize a kind of target organism detection method when the computer program is executed by processor.
It will be understood by those skilled in the art that structure shown in Fig. 8, only part relevant to application scheme is tied
The block diagram of structure does not constitute the restriction for the computer equipment being applied thereon to application scheme, specific computer equipment
It may include perhaps combining certain components or with different component layouts than more or fewer components as shown in the figure.
In one embodiment, a kind of computer equipment, including memory and processor are provided, which is stored with
Computer program, which performs the steps of when executing computer program when reaching preset condition, to video monitoring end
Enabled instruction is sent, video monitoring end starts according to enabled instruction to be monitored;The video information for obtaining the monitoring of video monitoring end, from view
Video frame to be detected is extracted in frequency information, obtains video pictures to be detected;Calculate the rgb value of video pictures to be detected, interception to
The picture region that rgb value in video pictures is greater than preset threshold is detected, picture region to be detected is obtained;By picture region to be detected
It is input in the target organism identification model trained and is identified, obtain target organism recognition result;When target organism identifies
It as a result is, from the video data obtained in video monitoring end in preset time range, video data to be protected there are when target organism
It deposits and returns to prompt messages to management terminal.
In one embodiment, acquisition monitoring picture is also performed the steps of when processor executes computer program, will be supervised
Control picture, which is input in the human testing model trained, to be detected, and testing result is obtained;When testing result is not the presence of people
When body, enabled instruction is sent to video monitoring end.
In one embodiment, it is also performed the steps of when processor executes computer program and obtains the current time in system,
When the current time in system is consistent with preset time, enabled instruction is sent to video monitoring end.
In one embodiment, it also performs the steps of when processor executes computer program by picture region to be detected
Picture is input in convolutional neural networks model and identifies by the picture for being converted to target resolution, obtains model output knot
Fruit;Target organism classification, quantity and location information are obtained from model output result, according to target organism classification, quantity and position
Confidence ceases to obtain target organism recognition result.
In one embodiment, it is also performed the steps of when processor executes computer program and obtains history monitor video
With corresponding target organism markup information;Using history monitor video as the input of convolutional neural networks algorithm, by target organism
Markup information is trained as label, when reaching preset condition, the target organism identification model trained.
In one embodiment, record alarm times are also performed the steps of when processor executes computer program, work as report
When alert number is greater than preset times, the corresponding target identification in video monitoring end is obtained, target identification and video data are sent into prison
Tube terminal.
In one embodiment, it is also performed the steps of when processor executes computer program and obtains multiple video monitorings
The each video information returned and the corresponding target identification of each video information are held, and calculates the cryptographic Hash of target identification;According to
Each video information is assigned to from node server by cryptographic Hash, passes through the identification point of target organism identification model from node server
The video information matched obtains target organism recognition result;It obtains and is identified from target organism of the node server to each video information
As a result, corresponding target organism recognition result is returned to management terminal according to target identification.
In one embodiment, a kind of computer readable storage medium is provided, computer program is stored thereon with, is calculated
Machine program is performed the steps of when being executed by processor when reaching preset condition, sends enabled instruction to video monitoring end, depending on
Frequency monitoring client starts according to enabled instruction to be monitored;Obtain video monitoring end monitoring video information, from video information extract to
Video frame is detected, video pictures to be detected are obtained;The rgb value for calculating video pictures to be detected intercepts in video pictures to be detected
Rgb value is greater than the picture region of preset threshold, obtains picture region to be detected;Picture region to be detected is input to and has been trained
It is identified in target organism identification model, obtains target organism recognition result;When target organism recognition result is that there are targets
Biochron saves video data and from the video data obtained in preset time range in video monitoring end to management terminal
Return to prompt messages.
In one embodiment, acquisition monitoring picture is also performed the steps of when computer program is executed by processor, it will
Monitoring picture, which is input in the human testing model trained, to be detected, and testing result is obtained;When testing result is not exist
When human body, enabled instruction is sent to video monitoring end.
In one embodiment, also performed the steps of when computer program is executed by processor acquisition system it is current when
Between, when the current time in system is consistent with preset time, enabled instruction is sent to video monitoring end.
In one embodiment, it is also performed the steps of when computer program is executed by processor by picture region to be detected
Domain is converted to the picture of target resolution, and picture is input in convolutional neural networks model and is identified, obtains model output
As a result;From model output result in obtain target organism classification, quantity and location information, according to target organism classification, quantity and
Location information obtains target organism recognition result.
In one embodiment, it is also performed the steps of when computer program is executed by processor and obtains history monitoring view
Frequency and corresponding target organism markup information;It is using history monitor video as the input of convolutional neural networks algorithm, target is raw
Object markup information is trained as label, when reaching preset condition, the target organism identification model trained.
In one embodiment, record alarm times are also performed the steps of when computer program is executed by processor, when
When alarm times are greater than preset times, the corresponding target identification in video monitoring end is obtained, target identification and video data are sent
Supervisory terminal.
In one embodiment, it is also performed the steps of when computer program is executed by processor and obtains multiple video prisons
The each video information and the corresponding target identification of each video information that end returns are controlled, and calculates the cryptographic Hash of target identification;Root
Each video information is assigned to from node server according to cryptographic Hash, is identified from node server by target organism identification model
The video information of distribution obtains target organism recognition result;It obtains and knows from target organism of the node server to each video information
Not as a result, corresponding target organism recognition result is returned to management terminal according to target identification
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
To any reference of memory, storage, database or other media used in each embodiment provided herein,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
Each technical characteristic of above embodiments can be combined arbitrarily, for simplicity of description, not to above-described embodiment
In each technical characteristic it is all possible combination be all described, as long as however, the combination of these technical characteristics be not present lance
Shield all should be considered as described in this specification.
The several embodiments of the application above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, without departing from the concept of this application, various modifications and improvements can be made, these belong to the protection of the application
Range.Therefore, the scope of protection shall be subject to the appended claims for the application patent.
Claims (10)
1. a kind of target organism detection method, which comprises
When reaching preset condition, enabled instruction is sent to video monitoring end, the video monitoring end is according to the enabled instruction
Starting monitoring;
The video information for obtaining the monitoring of video monitoring end, video frame to be detected is extracted from the video information, is obtained to be detected
Video pictures;
The rgb value of the video pictures to be detected is calculated, rgb value described in the video pictures to be detected is intercepted and is greater than default threshold
The picture region of value obtains picture region to be detected;
The picture region to be detected is input in the target organism identification model trained and is identified, target organism is obtained
Recognition result;
When the target organism recognition result is to obtain preset time range from the video monitoring end there are when target organism
The video data is saved and returns to prompt messages to management terminal by interior video data.
2. being sent out to video monitoring end the method according to claim 1, wherein described when reaching preset condition
Send enabled instruction, comprising:
Monitoring picture is obtained, the monitoring picture is input in the human testing model trained and is detected, is detected
As a result;
When the testing result is not there is no human body, enabled instruction is sent to video monitoring end.
3. being sent out to video monitoring end the method according to claim 1, wherein described when reaching preset condition
Send enabled instruction, comprising:
The current time in system is obtained, when the current time in system is with presetting consistent, sends enabled instruction to video monitoring end.
4. the method according to claim 1, wherein the picture region to be detected is input to the mesh trained
It is identified in mark bio-identification model, obtains target organism recognition result, comprising:
The picture is input to convolutional neural networks mould by the picture that the picture region to be detected is converted to target resolution
It is identified in type, obtains model output result;
From the model output result in obtain target organism classification, quantity and location information, according to the target organism classification,
Quantity and location information obtain target organism recognition result.
5. the method according to claim 1, wherein the generation of the target organism identification model trained walks
Suddenly, comprising:
Obtain history monitor video and corresponding target organism markup information;
Using the history monitor video as the input of convolutional neural networks algorithm, using the target organism markup information as mark
Label are trained, and when reaching preset condition, obtain the target organism identification model trained.
6. the method according to claim 1, wherein described is being that there are mesh when the target organism recognition result
The biochron is marked, from the video data obtained in the Video Monitoring Terminal in preset time range, the video data is saved
And to management terminal return prompt messages after, further includes:
Alarm times are recorded, when the alarm times are greater than preset times, obtain the corresponding target mark in the video monitoring end
Know, the target identification and the video data are sent into supervisory terminal.
7. the method according to claim 1, wherein the method also includes:
The each video information and the corresponding target identification of each video information that multiple video monitoring ends return are obtained, and calculates institute
State the cryptographic Hash of target identification;
Each video information is assigned to from node server according to the cryptographic Hash, it is described to pass through institute from node server
The video information for stating the identification distribution of target organism identification model obtains target organism recognition result;
It obtains from node server to the target organism recognition result of each video information, it will be corresponding according to the target identification
Target organism recognition result returns to management terminal.
8. a kind of target organism detection device, which is characterized in that described device includes:
Starting module is monitored, for sending enabled instruction, the video monitoring end to video monitoring end when reaching preset condition
Started according to the enabled instruction and is monitored;
Frame module is taken out, for obtaining the video information of video monitoring end monitoring, video to be detected is extracted from the video information
Frame obtains video pictures to be detected;
Interception module intercepts described in the video pictures to be detected for calculating the rgb value of the video pictures to be detected
Rgb value is greater than the picture region of preset threshold, obtains picture region to be detected;
Identification module is known for the picture region to be detected to be input in the target organism identification model trained
Not, target organism recognition result is obtained;
Video acquiring module, for when the target organism recognition result is there are when target organism, from the video monitoring end
The video data is saved and returns to warning note letter to management terminal by the video data in middle acquisition preset time range
Breath.
9. a kind of computer equipment, including memory and processor, the memory are stored with computer program, feature exists
In the step of processor realizes any one of claims 1 to 7 the method when executing the computer program.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of method described in any one of claims 1 to 7 is realized when being executed by processor.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910603620.7A CN110472492A (en) | 2019-07-05 | 2019-07-05 | Target organism detection method, device, computer equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910603620.7A CN110472492A (en) | 2019-07-05 | 2019-07-05 | Target organism detection method, device, computer equipment and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110472492A true CN110472492A (en) | 2019-11-19 |
Family
ID=68506800
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910603620.7A Pending CN110472492A (en) | 2019-07-05 | 2019-07-05 | Target organism detection method, device, computer equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110472492A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111126153A (en) * | 2019-11-25 | 2020-05-08 | 北京锐安科技有限公司 | Safety monitoring method, system, server and storage medium based on deep learning |
CN111160123A (en) * | 2019-12-11 | 2020-05-15 | 桂林长海发展有限责任公司 | Airplane target identification method and device and storage medium |
CN111552837A (en) * | 2020-05-08 | 2020-08-18 | 深圳市英威诺科技有限公司 | Animal video tag automatic generation method based on deep learning, terminal and medium |
CN113365018A (en) * | 2020-03-06 | 2021-09-07 | 杭州海康威视数字技术股份有限公司 | Period adjusting method, device and equipment |
JP2021152966A (en) * | 2020-06-29 | 2021-09-30 | ベイジン バイドゥ ネットコム サイエンス アンド テクノロジー カンパニー リミテッド | Seat belt wearing state recognition method, recognition apparatus, electronic device, and recording medium |
CN113591589A (en) * | 2021-07-02 | 2021-11-02 | 北京百度网讯科技有限公司 | Product missing detection identification method and device, electronic equipment and storage medium |
CN113593142A (en) * | 2021-07-26 | 2021-11-02 | 中国工商银行股份有限公司 | Automatic Teller Machine (ATM) patrolling method and device |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102693537A (en) * | 2011-01-17 | 2012-09-26 | 三星泰科威株式会社 | Image surveillance system and method of detecting whether object is left behind or taken away |
CN108829762A (en) * | 2018-05-28 | 2018-11-16 | 思百达物联网科技(北京)有限公司 | The Small object recognition methods of view-based access control model and device |
CN109640113A (en) * | 2018-12-28 | 2019-04-16 | 网宿科技股份有限公司 | A kind of processing method and proxy server of dilatory video data |
CN109684946A (en) * | 2018-12-10 | 2019-04-26 | 成都睿码科技有限责任公司 | A kind of kitchen mouse detection method based on the modeling of single Gaussian Background |
CN109815851A (en) * | 2019-01-03 | 2019-05-28 | 深圳壹账通智能科技有限公司 | Kitchen hygiene detection method, device, computer equipment and storage medium |
CN109919009A (en) * | 2019-01-24 | 2019-06-21 | 北京明略软件***有限公司 | The monitoring method of target object, apparatus and system |
-
2019
- 2019-07-05 CN CN201910603620.7A patent/CN110472492A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102693537A (en) * | 2011-01-17 | 2012-09-26 | 三星泰科威株式会社 | Image surveillance system and method of detecting whether object is left behind or taken away |
CN108829762A (en) * | 2018-05-28 | 2018-11-16 | 思百达物联网科技(北京)有限公司 | The Small object recognition methods of view-based access control model and device |
CN109684946A (en) * | 2018-12-10 | 2019-04-26 | 成都睿码科技有限责任公司 | A kind of kitchen mouse detection method based on the modeling of single Gaussian Background |
CN109640113A (en) * | 2018-12-28 | 2019-04-16 | 网宿科技股份有限公司 | A kind of processing method and proxy server of dilatory video data |
CN109815851A (en) * | 2019-01-03 | 2019-05-28 | 深圳壹账通智能科技有限公司 | Kitchen hygiene detection method, device, computer equipment and storage medium |
CN109919009A (en) * | 2019-01-24 | 2019-06-21 | 北京明略软件***有限公司 | The monitoring method of target object, apparatus and system |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111126153A (en) * | 2019-11-25 | 2020-05-08 | 北京锐安科技有限公司 | Safety monitoring method, system, server and storage medium based on deep learning |
CN111160123A (en) * | 2019-12-11 | 2020-05-15 | 桂林长海发展有限责任公司 | Airplane target identification method and device and storage medium |
CN111160123B (en) * | 2019-12-11 | 2023-06-09 | 桂林长海发展有限责任公司 | Aircraft target identification method, device and storage medium |
CN113365018A (en) * | 2020-03-06 | 2021-09-07 | 杭州海康威视数字技术股份有限公司 | Period adjusting method, device and equipment |
CN113365018B (en) * | 2020-03-06 | 2022-07-01 | 杭州海康威视数字技术股份有限公司 | Period adjusting method, device and equipment |
CN111552837A (en) * | 2020-05-08 | 2020-08-18 | 深圳市英威诺科技有限公司 | Animal video tag automatic generation method based on deep learning, terminal and medium |
JP2021152966A (en) * | 2020-06-29 | 2021-09-30 | ベイジン バイドゥ ネットコム サイエンス アンド テクノロジー カンパニー リミテッド | Seat belt wearing state recognition method, recognition apparatus, electronic device, and recording medium |
CN113591589A (en) * | 2021-07-02 | 2021-11-02 | 北京百度网讯科技有限公司 | Product missing detection identification method and device, electronic equipment and storage medium |
CN113593142A (en) * | 2021-07-26 | 2021-11-02 | 中国工商银行股份有限公司 | Automatic Teller Machine (ATM) patrolling method and device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110472492A (en) | Target organism detection method, device, computer equipment and storage medium | |
CN108922622B (en) | Animal health monitoring method, device and computer readable storage medium | |
US7617167B2 (en) | Machine vision system for enterprise management | |
CN109766766A (en) | Employee work condition monitoring method, device, computer equipment and storage medium | |
CN109767261A (en) | Products Show method, apparatus, computer equipment and storage medium | |
CN109344682A (en) | Classroom monitoring method, device, computer equipment and storage medium | |
US20150327518A1 (en) | Method of monitoring infectious disease, system using the same, and recording medium for performing the same | |
US20200226360A1 (en) | System and method for automatically detecting and classifying an animal in an image | |
CN110197107B (en) | Micro-expression recognition method, micro-expression recognition device, computer equipment and storage medium | |
US11875569B2 (en) | Smart video surveillance system using a neural network engine | |
CN110458008A (en) | Method for processing video frequency, device, computer equipment and storage medium | |
CN109376237A (en) | Prediction technique, device, computer equipment and the storage medium of client's stability | |
US11429820B2 (en) | Methods for inter-camera recognition of individuals and their properties | |
US20180197287A1 (en) | Process of using machine learning for cannabis plant health diagnostics | |
US20200225820A1 (en) | Minimally invasive user metadata | |
US10891480B2 (en) | Image zone processing | |
CN111159476B (en) | Target object searching method and device, computer equipment and storage medium | |
WO2023279716A1 (en) | Device linkage method and apparatus, and device, storage medium, program product and computer program | |
CN110298239B (en) | Target monitoring method, device, computer equipment and storage medium | |
CN108874910A (en) | The Small object identifying system of view-based access control model | |
US20220094785A1 (en) | Systems and methods for telephone call regulation based on spam factor and user input | |
WO2008127235A2 (en) | Machine vision system for enterprise management | |
Zambare et al. | Deep Learning Model for Disease Identification of Cotton Plants | |
CN116311036A (en) | Behavior recognition method, behavior recognition device, computer equipment and storage medium | |
Alexander et al. | A novel camera trapping method for individually identifying pumas by facial features |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191119 |