CN108304818A - A kind of mosquito matchmaker automatic distinguishing method for image - Google Patents
A kind of mosquito matchmaker automatic distinguishing method for image Download PDFInfo
- Publication number
- CN108304818A CN108304818A CN201810136379.7A CN201810136379A CN108304818A CN 108304818 A CN108304818 A CN 108304818A CN 201810136379 A CN201810136379 A CN 201810136379A CN 108304818 A CN108304818 A CN 108304818A
- Authority
- CN
- China
- Prior art keywords
- mosquito
- mosquito matchmaker
- matchmaker
- image
- automatic distinguishing
- 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
- 241000255925 Diptera Species 0.000 title claims abstract description 197
- 238000000034 method Methods 0.000 title claims abstract description 50
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 24
- 238000001514 detection method Methods 0.000 claims abstract description 17
- 238000000605 extraction Methods 0.000 claims abstract description 13
- 239000000284 extract Substances 0.000 claims abstract description 7
- 238000012549 training Methods 0.000 claims description 28
- RTZKZFJDLAIYFH-UHFFFAOYSA-N Diethyl ether Chemical compound CCOCC RTZKZFJDLAIYFH-UHFFFAOYSA-N 0.000 claims description 12
- 238000005286 illumination Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 6
- 241000256186 Anopheles <genus> Species 0.000 claims description 5
- 241000256118 Aedes aegypti Species 0.000 claims description 4
- 241000256054 Culex <genus> Species 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 abstract description 16
- 230000007774 longterm Effects 0.000 abstract description 4
- 230000008859 change Effects 0.000 abstract description 2
- 238000007689 inspection Methods 0.000 abstract description 2
- 239000000463 material Substances 0.000 abstract description 2
- 208000035473 Communicable disease Diseases 0.000 description 8
- 208000015181 infectious disease Diseases 0.000 description 7
- 208000001490 Dengue Diseases 0.000 description 3
- 206010012310 Dengue fever Diseases 0.000 description 3
- 208000025729 dengue disease Diseases 0.000 description 3
- 238000011176 pooling Methods 0.000 description 3
- 102000002322 Egg Proteins Human genes 0.000 description 2
- 108010000912 Egg Proteins Proteins 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000006806 disease prevention Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 210000004681 ovum Anatomy 0.000 description 2
- 230000002265 prevention Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 241000256173 Aedes albopictus Species 0.000 description 1
- 241000256060 Culex tritaeniorhynchus Species 0.000 description 1
- 241000256113 Culicidae Species 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 239000005667 attractant Substances 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000001364 causal effect Effects 0.000 description 1
- 230000031902 chemoattractant activity Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 206010014599 encephalitis Diseases 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000003475 lamination Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000005541 medical transmission Effects 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 230000029264 phototaxis Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012502 risk assessment Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2431—Multiple classes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/80—Geometric correction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of mosquito matchmaker automatic distinguishing method for image, establish Faster R CNN frame models, input mosquito matchmaker's image, feature extraction is carried out to mosquito matchmaker's image in conjunction with 16 models of VGG, it is generated using full convolutional neural networks and extracts window with the one-to-one mosquito matchmaker of mosquito corpse, mosquito matchmaker extraction window is mapped on characteristic layer, each mosquito matchmaker extraction window is divided into fixed-size characteristic layer based on maximum value pond theory, the mosquito corpse in window is extracted to each mosquito matchmaker using detection class probability and is classified；As a result of 16 model theory of Faster R CNN frame models and VGG, change the pure hand inspection monitoring means that traditional mosquito matchmaker monitors system, not only save a large amount of manpower and materials, and it is advantageously implemented the online real-time intelligent monitoring function to mosquito matchmaker, mosquito matchmaker's monitoring result, easy to operate, saving manpower, energy field quick detection can be accurately obtained rapidly, the Site Detection need of work of large area and the long-term demand that efficiently mosquito matchmaker monitors comprehensively can be met, there is very strong practical value.
Description
Technical field
The present invention relates to mosquito matchmaker's monitoring method field more particularly to a kind of mosquito matchmaker automatic distinguishing method for image.
Background technology
With the acceleration of globalization, climate change and urbanization process, dengue fever, datum hole Kenya heat, stockaded village's card etc. in recent years
Mosquito matchmaker infectious disease is in diffusion tendency in the world, and China also faces the huge anti-governor pressure of mosquito matchmaker infectious disease.
For example, the mosquitos matchmaker such as aedes albopictus and Culex tritaeniorhynchus is to propagate the infectious diseases such as dengue fever and Japanese Type-B encephalitis
Causal organism is broadcast to the mankind from host, brings significant damage to human health, seriously threaten the people's by important medium
The security of the lives and property.And it is the key that carry out mosquito matchmaker's infectious disease prevention and control to be monitored to mosquito matchmaker, not only contributes to understand mosquito
Kind constitutes and Density Distribution, can also provide early warning and risk analysis to mosquito matchmaker's disease transmission.
Currently, the common method of China mosquito matchmaker monitoring have it is artificial bite method, people's account mass trapping, ultraviolet lamp mass trapping,
CO2Mass trapping, mosquito magnet method, BGS mosquito trapping apparatus, lures mosquito to lure ovum method etc. at black box approach.Though manually biting method and people's account mass trapping
It is easy to operate, but there is the risk for easily making operator catch；Ultraviolet lamp mass trapping utilizes the phototaxis catching mosquito of mosquito,
Use scope is wide, but itself is not strong to the attraction of mosquito class；CO2Mass trapping is with CO2It is safe and harmless as attractant, and more incline
To in attracting female mosquito, but the capture less effective for thering is research to think the method to culex and anopheles；Black box approach equipment is simple, easily grasps
Make, but there is research to point out not being suitable for capturing anopheles；It is tight to be mainly used in anaphelism because being not necessarily to external power supply for mosquito magnet method
The remote districts of weight；BGS mosquito trapping apparatus needs power resources and cost is higher；And lure mosquito that ovum fado is lured to be suitable for Dengue medium lineae ablicantes
The monitoring of yellow-fever mosquito population.Thus, most mosquito matchmaker monitoring method has its shortcomings and limitations, cannot meet long-term efficient mosquito comprehensively
The demand of matchmaker's monitoring.
In addition, mosquito matchmaker monitoring traditional at present is mostly after first passing through above method catching mosquito, then it is clear by artificial naked eyes
The mosquito value volume and range of product of point record capture.Minority checks measurement mosquito one by one under the microscope, or emerging using some
Detection method, such as Enzymology method and molecular biology method are monitored.But these detection methods all labor intensives, behaviour
Make complex steps, is affected by human factor, not objective enough and science, it is difficult to instruct the prevention and control of infectious disease, and not
Energy field quick detection, it is difficult to meet the field monitoring need of work of large area.
Invention content
In order to solve the above technical problems, the present invention provides a kind of mosquito matchmaker automatic distinguishing method for image, easy to operate, saving people
Power, energy field quick detection can meet the Site Detection need of work of large area and the long-term need that efficiently mosquito matchmaker monitors comprehensively
It asks.
Technical scheme is as follows：A kind of mosquito matchmaker automatic distinguishing method for image, includes the following steps：
A, Faster R-CNN frame models are established, mosquito matchmaker's image is inputted；
B, in conjunction with VGG-16 models, feature extraction is carried out to mosquito matchmaker's image, obtains characteristic layer；
C, it is generated using full convolutional neural networks and extracts window with the one-to-one mosquito matchmaker of mosquito corpse；
D, mosquito matchmaker extraction window is mapped on characteristic layer；
E, each mosquito matchmaker extraction window is divided into fixed-size characteristic layer based on maximum value pond theory；
F, the mosquito corpse in window is extracted to each mosquito matchmaker using detection class probability to classify.
Mosquito matchmaker's automatic distinguishing method for image, wherein the step A includes：
A1, unified standardization processing is carried out to the size of mosquito matchmaker's image；
A2, the regular processing of illumination is carried out to mosquito matchmaker's image after standardization using Gamma correcting algorithms；
A3, denoising is carried out to mosquito matchmaker's image of the illumination after regular using Gassian low-pass filter algorithm.
Mosquito matchmaker's automatic distinguishing method for image, wherein in the step A1, the size of the mosquito matchmaker image is united
One is 3264 × 2368 × 3.
Mosquito matchmaker's automatic distinguishing method for image, wherein, will be in Gamma correcting algorithms in the step A2gamma
Value range be set between 0.4 ~ 2.2.
Mosquito matchmaker's automatic distinguishing method for image, wherein mosquito matchmaker's image in the step A is divided into mosquito matchmaker and trains sample
This library and mosquito matchmaker's image library to be identified instruct Faster R-CNN frame models by the mosquito matchmaker training sample database
Practice, and mosquito matchmaker's training sample database described in the identification statistical result iteration based on Faster R-CNN frames；Separately detected using mosquito matchmaker
Device obtains mosquito matchmaker's image library to be identified.
Mosquito matchmaker's automatic distinguishing method for image, wherein the mosquito matchmaker training sample database is divided into initial mosquito matchmaker and is trained
Sample and follow-up mosquito matchmaker training sample, by being captured to Laboratory culture or field and the mosquito corpse through ether execution is clapped
According to the mosquito matchmaker's image for obtaining the initial mosquito matchmaker training sample, and according to described in the continuous renewal of the result of follow-up mosquito matchmaker image procossing
Mosquito matchmaker's image of follow-up mosquito matchmaker's training sample.
Mosquito matchmaker's automatic distinguishing method for image, wherein put to death to Laboratory culture or field capture and through ether
The number that mosquito corpse is taken pictures is more than 100 times, to obtain the initial mosquito matchmaker training sample, and using manual type to first
Mosquito matchmaker's image of beginning mosquito matchmaker training sample is marked, and marks the mosquito corpse quantity more than 5000.
Mosquito matchmaker's automatic distinguishing method for image, wherein, will be in the full convolutional neural networks in the step C
The initial window size of each candidate region is set as 64 × 64,48 × 84,84 × 48.
Mosquito matchmaker's automatic distinguishing method for image, wherein in the step E, each mosquito matchmaker is extracted into window and is divided
At the characteristic layer of fixed dimension 7 × 7.
Mosquito matchmaker's automatic distinguishing method for image, wherein in the step F, the type of mosquito corpse is divided into library
Three mosquito, yellow-fever mosquito and anopheles classifications.
A kind of mosquito matchmaker automatic distinguishing method for image provided by the present invention, as a result of Faster R-CNN frame models
With VGG-16 model theories, the pure hand inspection monitoring means that traditional mosquito matchmaker monitors system is changed, a large amount of manpowers are not only saved
Material resources, and it is advantageously implemented the online real-time intelligent monitoring function to mosquito matchmaker, mosquito matchmaker's monitoring result can be accurately obtained rapidly, operated
Simply, manpower is saved, energy field quick detection can meet the Site Detection need of work of large area and long-term efficient mosquito comprehensively
The demand of matchmaker's monitoring has very strong practical value, and human factor influence is smaller, and more objective and science is suitble to instruct mosquito
The prevention and control of matchmaker's infectious disease.
Description of the drawings
Fig. 1 is the flow chart of mosquito matchmaker automatic distinguishing method for image embodiment of the present invention；
Fig. 2 is Faster R-CNN frame model structural schematic diagrams used in mosquito matchmaker automatic distinguishing method for image of the present invention.
Specific implementation mode
Below with reference to attached drawing, the specific implementation mode and embodiment of the present invention are described in detail, described tool
Body embodiment only to explain the present invention, is not intended to limit the specific implementation mode of the present invention.
As shown in Figure 1, Fig. 1 is the flow chart of mosquito matchmaker automatic distinguishing method for image embodiment of the present invention, the mosquito matchmaker image is certainly
Dynamic recognition methods, is used in background server, includes the following steps：
Step S110, Faster R-CNN frame models are established, mosquito matchmaker's image is inputted；Faster R-CNN frame moulds herein
Type 200 refers to deep learning Faster R-CNN frame models, uses for reference and with reference to S. Ren, et al. " Faster R-
CNN: Towards real-time object detection with region proposal networks.”
2015 documents of Advances in Neural Information Processing Systems., and use it for mosquito matchmaker figure
In the recognition methods of picture；In conjunction with shown in Fig. 2, Fig. 2 is Faster R-CNN frames used in mosquito matchmaker automatic distinguishing method for image of the present invention
Frame model structure schematic diagram；
Step S120, mosquito matchmaker image 201 is inputted into VGG-16 models 210, carries out feature extraction, obtains characteristic layer (Feature
maps)220；VGG-16 models 210 herein are used for reference and with reference to K. Simonyan and A. Zisserman. Very
Deep Convolutional Networks for Large-Scale Image RecognitionVery deep
convolutional networks for large-scale image recognition. ArXiv Technical
Report, 2014 documents, and use it in the recognition methods of mosquito matchmaker's image；
Step S130, it is generated using full convolutional neural networks and extracts window 202 with the one-to-one mosquito matchmaker of mosquito corpse；It is described complete
Convolutional neural networks can receive the input picture of arbitrary dimension, and using warp lamination to the characteristic layer of the last one convolutional layer
Feature map are up-sampled, it is made to be restored to the identical size of input picture, so as to generate one to each pixel
A prediction, while the spatial information in original input picture is also remained, convenient for finally being carried out by picture on the characteristic pattern of sampling
Element classification；RPN in Fig. 1 refer to region suggest network module 240, one as Faster R-CNN frame models 200
Point, which suggests that network module 240 is exactly a kind of full convolutional neural networks, for extracting candidate frame i.e. mosquito matchmaker region；In spy
On the basis of levying layer (Feature maps) 220, generates mosquito matchmaker with RPN and extract window (Proposals) 202a, often in pictures
Every mosquito corpse correspond to mosquito matchmaker and extract window 202a；
Step S140, mosquito matchmaker extraction window 202a is mapped on characteristic layer (Feature maps) 220；
Step S150, it is based on maximum value pond (Max-pooling) theory, each mosquito matchmaker extraction window 202a is divided into fixation
The characteristic layer (Feature maps) 230 of size；So-called maximum value pond (Max-pooling) refers to choosing image-region
Maximum value as the value after the pool area (Pooling)；
Step S160, using detection class probability (Softmax Loss), the mosquito corpse in window 202a is extracted to each mosquito matchmaker
Body is classified, and exports the mosquito matchmaker image 203 after identification；So-called detection class probability refers to Softmax loss functions.
In the preferred embodiment of mosquito matchmaker automatic distinguishing method for image of the present invention, to improve the discrimination of mosquito matchmaker's image,
Step S110 further include the steps that mosquito matchmaker image 201 is standardized, illumination correction and denoising, specifically include：
Step S111, unified standardization processing is carried out to the size of mosquito matchmaker image 201；Preferably, by the mosquito matchmaker image 201
Size is unified for 3264 × 2368 × 3；
Step S112, the regular processing of illumination is carried out to the mosquito matchmaker image 201 after standardization using Gamma correcting algorithms；Herein
Gamma correcting algorithm formula be Iout=Iin gamma, the wherein value range of gamma is set between 0.4~2.2；Step
S113, denoising is carried out to mosquito matchmaker image 201 of the illumination after regular using Gassian low-pass filter algorithm.
In the preferred embodiment of mosquito matchmaker automatic distinguishing method for image of the present invention, it is preferred that by the step S110
In mosquito matchmaker image 201 be divided into mosquito matchmaker training sample database and mosquito matchmaker's image library to be identified, pass through the mosquito matchmaker training sample database
Faster R-CNN frame models are trained, and mosquito described in the identification statistical result iteration based on Faster R-CNN frames
Matchmaker's training sample database, to adapt to the diversity of mosquito matchmaker type and its corpse posture；In addition it is obtained using mosquito matchmaker's detection device and waits knowing
Other mosquito matchmaker image library.
Further, the mosquito matchmaker training sample database is divided into initial mosquito matchmaker training sample and follow-up mosquito matchmaker training sample,
By the mosquito corpse for capturing Laboratory culture or field and putting to death through ether take pictures and obtains the initial mosquito matchmaker training
Mosquito matchmaker's image of sample, and constantly update according to the result of follow-up mosquito matchmaker image procossing the mosquito matchmaker of the follow-up mosquito matchmaker training sample
Image.
Specifically, being more than to Laboratory culture or field capture and the number taken pictures of mosquito corpse put to death through ether
100 times, to obtain the initial mosquito matchmaker training sample, and using manual type to the mosquito matchmaker image of initial mosquito matchmaker training sample into
Line flag, and mark the mosquito corpse quantity more than 5000.
In the step S130, it is preferable that for the shapes and sizes of mosquito in mosquito matchmaker image 201, by the full volume
The initial window size of each candidate region is set as 64 × 64,48 × 84,84 × 48 in product neural network, it is possible thereby to save
It is a large amount of to calculate time and memory capacity.
In the step S150, it is preferable that each mosquito matchmaker extraction window 202a is divided into the spy of fixed dimension 7 × 7
Levy layer (Feature maps) 230.
It, can be by the type of mosquito and its corpse point in the preferred embodiment of mosquito matchmaker automatic distinguishing method for image of the present invention
For three culex, yellow-fever mosquito and anopheles classifications.
Mosquito matchmaker's automatic distinguishing method for image of the present invention, the problem of preferably overcoming current mosquito matchmaker monitoring
And technological difficulties, monitoring capability of the China to mosquito matchmaker's infectious disease can be greatly improved, China emphasis is further grasped and passes
The occurrence regularity and fashion trend caught an illness are helped to form scientific and effective infectious disease risk evaluation system, and are carried for control and prevention of disease
For scientific theory foundation.
It should be understood that the foregoing is merely illustrative of the preferred embodiments of the present invention, it is not sufficient to the limitation present invention's
Technical solution within the spirit and principles in the present invention, can add according to the above description for those of ordinary skills
It with increase and decrease, replacement, transformation or improvement, and all these increases and decreases, replaces, transformation or improved technical solution, should all belong to this
The protection domain of invention appended claims.
Claims (10)
1. a kind of mosquito matchmaker automatic distinguishing method for image, which is characterized in that include the following steps：
A, Faster R-CNN frame models are established, mosquito matchmaker's image is inputted；
B, in conjunction with VGG-16 models, feature extraction is carried out to mosquito matchmaker's image, obtains characteristic layer；
C, it is generated using full convolutional neural networks and extracts window with the one-to-one mosquito matchmaker of mosquito corpse；
D, mosquito matchmaker extraction window is mapped on characteristic layer；
E, each mosquito matchmaker extraction window is divided into fixed-size characteristic layer based on maximum value pond theory；
F, the mosquito corpse in window is extracted to each mosquito matchmaker using detection class probability to classify.
2. mosquito matchmaker automatic distinguishing method for image according to claim 1, which is characterized in that the step A includes：
A1, unified standardization processing is carried out to the size of mosquito matchmaker's image；
A2, the regular processing of illumination is carried out to mosquito matchmaker's image after standardization using Gamma correcting algorithms；
A3, denoising is carried out to mosquito matchmaker's image of the illumination after regular using Gassian low-pass filter algorithm.
3. mosquito matchmaker automatic distinguishing method for image according to claim 2, it is characterised in that：In the step A1, by institute
The size for stating mosquito matchmaker's image is unified for 3264 × 2368 × 3.
4. mosquito matchmaker automatic distinguishing method for image according to claim 2, it is characterised in that：It, will in the step A2
In Gamma correcting algorithmsgammaValue range be set between 0.4 ~ 2.2.
5. mosquito matchmaker automatic distinguishing method for image according to claim 1, it is characterised in that：By the mosquito matchmaker in the step A
Image is divided into mosquito matchmaker training sample database and mosquito matchmaker's image library to be identified, by the mosquito matchmaker training sample database to Faster R-
CNN frame models are trained, and mosquito matchmaker's training sample described in the identification statistical result iteration based on Faster R-CNN frames
Library；Mosquito matchmaker's image library to be identified is separately obtained using mosquito matchmaker's detection device.
6. mosquito matchmaker automatic distinguishing method for image according to claim 5, it is characterised in that：By the mosquito matchmaker training sample database
It is divided into initial mosquito matchmaker training sample and follow-up mosquito matchmaker training sample, by capturing to Laboratory culture or field and being put to death through ether
Mosquito corpse take pictures and obtain mosquito matchmaker's image of the initial mosquito matchmaker training sample, and according to follow-up mosquito matchmaker image procossing
As a result mosquito matchmaker's image of the follow-up mosquito matchmaker training sample is constantly updated.
7. mosquito matchmaker automatic distinguishing method for image according to claim 6, it is characterised in that：Laboratory culture or field are caught
It obtains and number that the mosquito corpse put to death through ether is taken pictures is more than 100 times, to obtain the initial mosquito matchmaker training sample, and
Mosquito matchmaker's image of initial mosquito matchmaker training sample is marked using manual type, and marks the mosquito body count more than 5000
Amount.
8. mosquito matchmaker automatic distinguishing method for image according to claim 1, it is characterised in that：It, will be described in the step C
The initial window size of each candidate region is set as 64 × 64,48 × 84,84 × 48 in full convolutional neural networks.
9. mosquito matchmaker automatic distinguishing method for image according to claim 1, it is characterised in that：It, will be each in the step E
Mosquito matchmaker extracts the characteristic layer that window is divided into fixed dimension 7 × 7.
10. mosquito matchmaker automatic distinguishing method for image according to claim 1, it is characterised in that：In the step F, by mosquito
The type of sub- corpse is divided into three culex, yellow-fever mosquito and anopheles classifications.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810136379.7A CN108304818A (en) | 2018-02-09 | 2018-02-09 | A kind of mosquito matchmaker automatic distinguishing method for image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810136379.7A CN108304818A (en) | 2018-02-09 | 2018-02-09 | A kind of mosquito matchmaker automatic distinguishing method for image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108304818A true CN108304818A (en) | 2018-07-20 |
Family
ID=62865153
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810136379.7A Pending CN108304818A (en) | 2018-02-09 | 2018-02-09 | A kind of mosquito matchmaker automatic distinguishing method for image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108304818A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10963742B2 (en) | 2018-11-02 | 2021-03-30 | University Of South Florida | Leveraging smart-phone cameras and image processing techniques to classify mosquito genus and species |
| US11048928B1 (en) | 2020-02-04 | 2021-06-29 | University Of South Florida | Systems and methods of entomology classification based on extracted anatomies |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102332086A (en) * | 2011-06-15 | 2012-01-25 | 夏东 | Facial identification method based on dual threshold local binary pattern |
| CN106548192A (en) * | 2016-09-23 | 2017-03-29 | 北京市商汤科技开发有限公司 | Based on the image processing method of neutral net, device and electronic equipment |
| CN107247955A (en) * | 2016-07-29 | 2017-10-13 | 竹间智能科技（上海）有限公司 | Accessory recognition methods and device |
| CN107665336A (en) * | 2017-09-20 | 2018-02-06 | 厦门理工学院 | Multi-target detection method based on Faster RCNN in intelligent refrigerator |
-
2018
- 2018-02-09 CN CN201810136379.7A patent/CN108304818A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102332086A (en) * | 2011-06-15 | 2012-01-25 | 夏东 | Facial identification method based on dual threshold local binary pattern |
| CN107247955A (en) * | 2016-07-29 | 2017-10-13 | 竹间智能科技（上海）有限公司 | Accessory recognition methods and device |
| CN106548192A (en) * | 2016-09-23 | 2017-03-29 | 北京市商汤科技开发有限公司 | Based on the image processing method of neutral net, device and electronic equipment |
| CN107665336A (en) * | 2017-09-20 | 2018-02-06 | 厦门理工学院 | Multi-target detection method based on Faster RCNN in intelligent refrigerator |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10963742B2 (en) | 2018-11-02 | 2021-03-30 | University Of South Florida | Leveraging smart-phone cameras and image processing techniques to classify mosquito genus and species |
| US11501113B2 (en) | 2018-11-02 | 2022-11-15 | University Of South Florida | Leveraging smart-phone cameras and image processing techniques to classify mosquito genus and species |
| US11989936B2 (en) | 2018-11-02 | 2024-05-21 | University Of South Florida | Leveraging smart-phone cameras and image processing techniques to classify mosquito genus and species |
| US11048928B1 (en) | 2020-02-04 | 2021-06-29 | University Of South Florida | Systems and methods of entomology classification based on extracted anatomies |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109726652B (en) | Method for detecting sleeping behavior of person on duty based on convolutional neural network | |
| CN104992223B (en) | Intensive population estimation method based on deep learning | |
| CN109543679A (en) | A kind of dead fish recognition methods and early warning system based on depth convolutional neural networks | |
| CN108334938A (en) | A kind of mosquito matchmaker's automatic monitoring system based on image recognition | |
| Qing et al. | Development of an automatic monitoring system for rice light-trap pests based on machine vision | |
| CN110046631A (en) | System and method for inferring the variation of time-space image automatically | |
| CN108921039A (en) | The forest fire detection method of depth convolution model based on more size convolution kernels | |
| WO2019126971A1 (en) | Pest disaster monitoring method based on machine vision | |
| US10963742B2 (en) | Leveraging smart-phone cameras and image processing techniques to classify mosquito genus and species | |
| CN105701467A (en) | Many-people abnormal behavior identification method based on human body shape characteristic | |
| EP4161258A1 (en) | Insect monitoring system and method | |
| CN114359727A (en) | Tea disease identification method and system based on lightweight optimization Yolo v4 | |
| CN114049550A (en) | Cage-rearing dead broiler automatic identification method based on deep learning | |
| CN107464260A (en) | A kind of rice canopy image processing method using unmanned plane | |
| CN111352349A (en) | System and method for collecting and adjusting information of old people living environment | |
| CN108304818A (en) | A kind of mosquito matchmaker automatic distinguishing method for image | |
| CN108734277A (en) | A kind of targeted drug spray method based on convolutional neural networks | |
| CN107610122A (en) | Insect pest detection method inside a kind of single grain cereal based on Micro CT | |
| CN104007733B (en) | It is a kind of that the system and method being monitored is produced to intensive agriculture | |
| CN107392251A (en) | A kind of method that target detection network performance is lifted using category images | |
| CN110634557A (en) | Medical care resource auxiliary allocation method and system based on deep neural network | |
| CN116824626A (en) | Artificial intelligent identification method for abnormal state of animal | |
| CN110245604A (en) | Mosquito recognition methods based on convolutional neural networks | |
| CN114511801A (en) | Insect pest recognition model training method, insect pest monitoring method and electronic equipment | |
| CN117315487B (en) | Wild animal image cross-domain identification method based on deep migration learning |
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: 20180720 |