CN109325499A - Pest and disease damage recognition methods and device - Google Patents
Pest and disease damage recognition methods and device Download PDFInfo
- Publication number
- CN109325499A CN109325499A CN201810869379.8A CN201810869379A CN109325499A CN 109325499 A CN109325499 A CN 109325499A CN 201810869379 A CN201810869379 A CN 201810869379A CN 109325499 A CN109325499 A CN 109325499A
- Authority
- CN
- China
- Prior art keywords
- pest
- color value
- picture
- feature data
- edge feature
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- 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
Abstract
The invention discloses a kind of pest and disease damage recognition methods and device to belong to computer technology image identification technical field, and the method specifically includes the edge feature data that parsing obtains pest picture to be identified；According to the edge feature data from default sample database, screening obtains the first matching result；Parsing obtains the color value characteristic of the pest picture；According to the color value characteristic from the first matching result, screening obtains the second matching result, and described device includes the first parsing module, the first matching module, the second parsing module and the second matching module.The present invention improves pest identification accuracy by being matched the pest being closer to for distinguishing some forms to edge feature and color value feature.
Description
Technical field
The present invention relates to computer technology image identification technical field more particularly to pest and disease damage recognition methods and realize the party
The device of method.
Background technique
In field of agricultural cultivation, the prevention and control of plant diseases, pest control is unavoidable work forever.It is in the first step of the prevention and control of plant diseases, pest control
Accurate identification and judgement to pest species.General agriculturist because professional knowledge and data shortage, it is difficult to the
The type of one time accurate judgement pest.It is existing based on image recognition technology be based on user shooting pest picture, pass through
The limb recognition of pest goes out pest shape in photo, then matches pest database by pest shape feature, to identify evil
Worm type provides the advisory opinion etc. of control of insect for agriculturist.
The place but this pest identification method comes with some shortcomings: the edge feature of different types of pest itself is more
Close, the features such as especially some pest larva shape postures are very close, be easy to cause biggish identification error, while photo
Background also can further interfere with the order of accuarcy of identification.
Summary of the invention
The present invention is the shortcoming in order to overcome Insect Pest Identification in the prior art, and providing can be by edge spy
Sign and color value feature distinguish the pest that some forms are closer to, to improve a kind of pest and disease damage identification of pest identification accuracy
Method and device.
To achieve the above object, the invention adopts the following technical scheme:
A kind of pest and disease damage recognition methods of the invention, the described method comprises the following steps:
Parsing obtains the edge feature data of pest picture to be identified；
According to the edge feature data from default sample database, screening obtains the first matching result；
Parsing obtains the color value characteristic of the pest picture；
According to the color value characteristic from the first matching result, screening obtains the second matching result.
Preferably, the sample database prestores sample pest picture, and with the sample pest picture pair
The edge feature data and color value characteristic answered.
Preferably, the step of parsing obtains the edge feature data of the pest picture, further comprises: passing through
Canny edge detection algorithm parses and obtains the edge feature data in pest picture.
Preferably, it includes shared by the color value data and each color value of color that the color value characteristic, which includes in pest picture,
The step of ratio, the parsing obtains the color value characteristic of the pest picture, further comprises: obtaining evil to be identified
The background colour color value range of worm photo；Color value section and the shared ratio of each color value of remainder are counted after exclusion background color value range
Example.
Preferably, the pest picture to be identified is the pest picture in solid background photographs.
A kind of pest and disease damage identification device of the invention, described device include:
First parsing module, for parsing the edge feature data for obtaining pest picture to be identified；
First matching module, for from default sample database, screening to obtain first according to the edge feature data
Matching result；
Second parsing module, for parsing the color value characteristic for obtaining the pest picture；
Second matching module, for from the first matching result, screening to obtain second according to the color value characteristic
With result.
Preferably, further include sample database, the sample database prestores sample pest picture, and with institute
State the corresponding edge feature data of sample pest picture and color value characteristic.
Preferably, first parsing module, is parsed by Canny edge detection algorithm and is obtained in pest picture
Edge feature data.
Preferably, second parsing module includes:
Acquiring unit, for obtaining the background colour color value range of pest photo to be identified；
Statistic unit compares shared by the color value section and each color value for excluding to count remainder after background color value range
Example.
Preferably, the pest picture to be identified is the pest picture in solid background photographs.
Technical solution of the present invention passes through edge feature data first and filters out the first matching result from sample database,
First matching result is a series of samples morphologically close with pest picture to be identified.Then schemed again by pest
The color value section of piece and color value proportion, two parameters filter out the second matching result from the first matching result.Due to knot
Color has been closed as identification feature, can the close pest of cog region division aspect, improve recognition efficiency and accuracy.
Detailed description of the invention
Fig. 1 is a kind of flow chart of pest and disease damage recognition methods of the invention.
Fig. 2 is the flow chart of another pest and disease damage recognition methods of the invention.
Fig. 3 is a kind of functional block diagram of pest and disease damage identification device of the invention.
Fig. 4 is the functional block diagram of another pest and disease damage identification device of the invention.
Specific embodiment
The present invention is described further with reference to the accompanying drawings and detailed description.
Embodiment one:
As shown in Figure 1, the embodiment of the present invention provides a kind of pest and disease damage recognition methods, the described method comprises the following steps:
S101 parsing obtains the edge feature data of pest picture to be identified.
It is that pest picture to be identified is parsed by Canny edge detection algorithm in the present embodiment, obtains edge therein
Characteristic, the edge feature data are the character shape datas that can show pest ontology in pest picture.It is described to
The pest picture of identification is the pest picture in solid background photographs.
For S102 according to the edge feature data from default sample database, screening obtains the first matching result.
The default sample database prestores sample pest picture, and side corresponding with the sample pest picture
Edge characteristic and color value characteristic.Edge feature data and the sample pest picture by pest picture to be identified
Edge feature data matched one by one.Filter out a certain number of set of metadata of similar data from high to low according to matching degree, i.e.,
One matching result.First matching result as it is tentatively identifying as a result, be one group of pest in shape with evil to be identified
The approximate sample pest picture of worm picture.
S103 parsing obtains the color value characteristic of the pest picture.
So-called color value is to represent the data of color.Color value characteristic is to be represented in pest picture using color value data
The color characteristic of pest.
For S104 according to the color value characteristic from the first matching result, screening obtains the second matching result.
By counting the color value characteristic of remainder after exclusion background color value range in this step, knot is matched with first
The color value characteristic of sample pest picture in fruit is matched one by one, filters out immediate second matching result.In this way
Secondary sieve can be carried out by color value characteristic on the basis of the first matching result gone out by shape edges Feature Selection
Choosing.So as to effectively identify the pest species distinguished and there is similar profile shape.Due to combining color as identification feature,
It being capable of the close pest of cog region division aspect, raising recognition efficiency and accuracy.
Embodiment two:
As shown in Fig. 2, the embodiment of the present invention is advanced optimizing and supplementing to embodiment one, another disease pest provided
Evil recognition methods, the described method comprises the following steps:
S201 obtains pest picture to be identified.The pest picture to be identified is the pest in solid background photographs
Picture.
S202 is parsed by Canny edge detection algorithm and is obtained the edge feature data in pest picture.
The edge feature data are the character shape datas that can show pest ontology in pest picture.It is described wait know
Other pest picture is the pest picture in solid background photographs.
For S203 according to the edge feature data from default sample database, screening obtains the first matching result.
The default sample database prestores sample pest picture, and side corresponding with the sample pest picture
Edge characteristic and color value characteristic.Edge feature data and the sample pest picture by pest picture to be identified
Edge feature data matched one by one.
S204 obtains the background colour color value range of pest photo to be identified.
The color value section of statistics remainder and each color value proportion after S205 exclusion background color value range.
This step is to obtain the background colour color value range of pest photo to be identified；Count surplus after excluding background color value range
The color value section of remaining part point and each color value proportion.So-called color value is to represent the data of color.
Color value section by counting remainder after statistics exclusion background color value range can obtain the color of pest
Feature, color value section allow for the color nuance between pest individual itself.Each color value proportion is counted as matching
One of according to, it is in order to distinguish the pest species with Similar color, accordingly even when there are the similar different pests of color composition
Type can also be distinguished according to the difference of each color proportion, such as the differentiation between various ladybugs.
S206 is according to the color value section and each color value proportion for counting remainder after the exclusion background color value range
From the first matching result, screening obtains the second matching result.
Color value section and each color value proportion in this step by statistics remainder after exclusion background color value range
Two item datas are matched one by one with the color value of the sample pest picture in the first matching result and each color value proportion,
Filter out immediate second matching result.
Technical solution of the present invention passes through edge feature data first and filters out the first matching result from sample database,
First matching result is a series of samples morphologically close with pest picture to be identified.Then schemed again by pest
The color value section of piece and color value proportion, two parameters filter out the second matching result from the first matching result.Due to knot
Color has been closed as identification feature, can the close pest of cog region division aspect, improve recognition efficiency and accuracy.
Embodiment three:
As shown in figure 3, the embodiment of the present invention provides a kind of pest and disease damage identification device, described device includes:
First parsing module 301, for parsing the edge feature data for obtaining pest picture to be identified.
The pest picture to be identified is the pest picture in solid background photographs.It is calculated by Canny edge detection
Method parses pest picture to be identified, obtains edge feature data therein, the edge feature data are can to show evil
The character shape data of pest ontology in worm picture.
First matching module 302, for from default sample database, screening to obtain the according to the edge feature data
One matching result.
The default sample database prestores sample pest picture, and side corresponding with the sample pest picture
Edge characteristic and color value characteristic.Edge feature data and the sample pest picture by pest picture to be identified
Edge feature data matched one by one.Filter out a certain number of set of metadata of similar data from high to low according to matching degree, i.e.,
One matching result.First matching result as it is tentatively identifying as a result, be one group of pest in shape with evil to be identified
The approximate sample pest picture of worm picture.
Second parsing module 303, for parsing the color value characteristic for obtaining the pest picture.
So-called color value is to represent the data of color.Color value characteristic is to be represented in pest picture using color value data
The color characteristic of pest.
Second matching module 304, for from the first matching result, screening to obtain second according to the color value characteristic
Matching result.
By counting the color value characteristic of remainder after exclusion background color value range, with the sample in the first matching result
The color value characteristic of this pest picture is matched one by one, filters out immediate second matching result.It in this way can be logical
It crosses on the basis of the first matching result that shape edges Feature Selection goes out, postsearch screening is carried out by color value characteristic.To
It can effectively identify the pest species distinguished and there is similar profile shape.Due to combining color as identification feature, Neng Goushi
It Qu Fen not the close pest of form, raising recognition efficiency and accuracy.
Example IV:
As shown in figure 4, the embodiment of the present invention is to provide another pest and disease damage to the further supplement of embodiment three and optimization
Identification device, the present embodiment and the difference of embodiment three are:
Second parsing module 303 includes:
Acquiring unit 401, for obtaining the background colour color value range of pest photo to be identified.
Statistic unit 402, for being counted shared by color value section and each color value of remainder after excluding background color value range
Ratio.
Described device further includes sample database 403, and the sample database prestores sample pest picture, Yi Jiyu
The corresponding edge feature data of the sample pest picture and color value characteristic.
Color value section by counting remainder after statistics exclusion background color value range can obtain the color of pest
Feature, color value section allow for the color nuance between pest individual itself.Each color value proportion is counted as matching
One of according to, it is in order to distinguish the pest species with Similar color, accordingly even when there are the similar different pests of color composition
Type can also be distinguished according to the difference of each color proportion, such as the differentiation between various ladybugs.
Claims (10)
1. a kind of pest and disease damage recognition methods, characterized in that the described method comprises the following steps:
Parsing obtains the edge feature data of pest picture to be identified；
According to the edge feature data from default sample database, screening obtains the first matching result；
Parsing obtains the color value characteristic of the pest picture；
According to the color value characteristic from the first matching result, screening obtains the second matching result.
2. a kind of pest and disease damage recognition methods according to claim 1, characterized in that the sample database prestores sample
This pest picture, and edge feature data corresponding with the sample pest picture and color value characteristic.
3. a kind of pest and disease damage recognition methods according to claim 1, characterized in that the parsing obtains the pest picture
Edge feature data the step of, further comprise:
It is parsed by Canny edge detection algorithm and obtains the edge feature data in pest picture.
4. a kind of pest and disease damage recognition methods according to claim 1, characterized in that the color value characteristic includes pest
It include color value data and each color value proportion in picture, the parsing obtains the color value characteristic of the pest picture
Step further comprises:
Obtain the background colour color value range of pest photo to be identified；
The color value section of statistics remainder and each color value proportion after exclusion background color value range.
5. a kind of pest and disease damage recognition methods according to claim 1, characterized in that the pest picture to be identified be
The pest picture of solid background photographs.
6. a kind of pest and disease damage identification device, characterized in that described device includes:
First parsing module, for parsing the edge feature data for obtaining pest picture to be identified；
First matching module, for from default sample database, screening to obtain the first matching according to the edge feature data
As a result；
Second parsing module, for parsing the color value characteristic for obtaining the pest picture；
Second matching module, for from the first matching result, screening to obtain the second matching knot according to the color value characteristic
Fruit.
7. a kind of pest and disease damage identification device according to claim 6, characterized in that it further include sample database, it is described
Sample database prestores sample pest picture, and edge feature data corresponding with the sample pest picture and color value spy
Levy data.
8. a kind of pest and disease damage identification device according to claim 6, characterized in that first parsing module passes through
Canny edge detection algorithm parses and obtains the edge feature data in pest picture.
9. a kind of pest and disease damage identification device according to claim 6, characterized in that second parsing module includes:
Acquiring unit, for obtaining the background colour color value range of pest photo to be identified；
Statistic unit, color value section and each color value proportion for excluding to count remainder after background color value range.
10. a kind of pest and disease damage identification device according to claim 6, characterized in that the pest picture to be identified is
In the pest picture of solid background photographs.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810869379.8A CN109325499A (en) | 2018-08-02 | 2018-08-02 | Pest and disease damage recognition methods and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810869379.8A CN109325499A (en) | 2018-08-02 | 2018-08-02 | Pest and disease damage recognition methods and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109325499A true CN109325499A (en) | 2019-02-12 |
Family
ID=65263559
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810869379.8A Pending CN109325499A (en) | 2018-08-02 | 2018-08-02 | Pest and disease damage recognition methods and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109325499A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110796148A (en) * | 2019-10-12 | 2020-02-14 | 广西大学 | Litchi insect pest monitoring and identifying system and litchi insect pest monitoring and identifying method |
| CN111143601A (en) * | 2019-12-31 | 2020-05-12 | 深圳市芭田生态工程股份有限公司 | Image processing method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102393902A (en) * | 2011-12-01 | 2012-03-28 | 昆山市工业技术研究院有限责任公司 | Vehicle color detection method based on H_S two-dimensional histogram and regional color matching |
| CN103093208A (en) * | 2013-01-23 | 2013-05-08 | 中国科学技术大学 | Method and system for fruit and vegetable recognition |
| CN104392240A (en) * | 2014-10-28 | 2015-03-04 | 中国疾病预防控制中心寄生虫病预防控制所 | Parasite egg identification method based on multi-feature fusion |
| US20160140411A1 (en) * | 2011-09-28 | 2016-05-19 | Casio Computer Co., Ltd. | Image processing apparatus for converting image in characteristic region of original image into image of brushstroke patterns |
| CN106886758A (en) * | 2017-01-20 | 2017-06-23 | 北京农业信息技术研究中心 | Based on insect identifying device and method that 3 d pose is estimated |
| CN107886077A (en) * | 2017-11-17 | 2018-04-06 | 苏州健雄职业技术学院 | A kind of crop pests recognition methods and its system based on wechat public number |
-
2018
- 2018-08-02 CN CN201810869379.8A patent/CN109325499A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160140411A1 (en) * | 2011-09-28 | 2016-05-19 | Casio Computer Co., Ltd. | Image processing apparatus for converting image in characteristic region of original image into image of brushstroke patterns |
| CN102393902A (en) * | 2011-12-01 | 2012-03-28 | 昆山市工业技术研究院有限责任公司 | Vehicle color detection method based on H_S two-dimensional histogram and regional color matching |
| CN103093208A (en) * | 2013-01-23 | 2013-05-08 | 中国科学技术大学 | Method and system for fruit and vegetable recognition |
| CN104392240A (en) * | 2014-10-28 | 2015-03-04 | 中国疾病预防控制中心寄生虫病预防控制所 | Parasite egg identification method based on multi-feature fusion |
| CN106886758A (en) * | 2017-01-20 | 2017-06-23 | 北京农业信息技术研究中心 | Based on insect identifying device and method that 3 d pose is estimated |
| CN107886077A (en) * | 2017-11-17 | 2018-04-06 | 苏州健雄职业技术学院 | A kind of crop pests recognition methods and its system based on wechat public number |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110796148A (en) * | 2019-10-12 | 2020-02-14 | 广西大学 | Litchi insect pest monitoring and identifying system and litchi insect pest monitoring and identifying method |
| CN110796148B (en) * | 2019-10-12 | 2020-07-07 | 广西大学 | Litchi insect pest monitoring and identifying system and litchi insect pest monitoring and identifying method |
| CN111143601A (en) * | 2019-12-31 | 2020-05-12 | 深圳市芭田生态工程股份有限公司 | Image processing method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Rastogi et al. | Leaf disease detection and grading using computer vision technology & fuzzy logic | |
| TWI435234B (en) | Plant disease identification method, system and record media | |
| CN112464971A (en) | Method for constructing pest detection model | |
| CN110569786B (en) | Fruit tree identification and quantity monitoring method and system based on unmanned aerial vehicle data acquisition | |
| Wang et al. | Mobile smart device-based vegetable disease and insect pest recognition method | |
| Malik et al. | Detection and counting of on-tree citrus fruit for crop yield estimation | |
| CN103955938A (en) | Wheat growing status diagnosing method based on mobile internet mode and leaf color analysis | |
| CN109325499A (en) | Pest and disease damage recognition methods and device | |
| CN109101891A (en) | A kind of rice pest detection system and its detection method merging artificial intelligence | |
| Shaikh et al. | Citrus leaf unhealthy region detection by using image processing technique | |
| CN106650822A (en) | Identification method and device for diseases and insect pests | |
| CN112686862A (en) | Pest identification and counting method, system and device and readable storage medium | |
| Williams et al. | Leaf only SAM: a segment anything pipeline for zero-shot automated leaf segmentation | |
| CN109934297A (en) | A kind of rice species test method based on deep learning convolutional neural networks | |
| Liu et al. | Detection of Ginseng leaf cicatrices base on K-means clustering algorithm | |
| Dhanesha et al. | Segmentation of arecanut bunches using YCgCr color model | |
| Alchanatis et al. | Apple yield mapping using hyperspectral machine vision | |
| Wang et al. | Automatic identification of asian rice plant-hopper based on image processing | |
| Tripathi et al. | A new technique for CBIR with contrast enhancement using multi-feature and multi class SVM classification | |
| Deng et al. | Disease Feature Recognition of Hydroponic Lettuce Images Based on Support Vector Machine. | |
| Chodey et al. | Crop pest detection and classification by K-means and EM clustering | |
| Lim et al. | Development of a classification algorithm for butterflies and ladybugs | |
| CN116994064B (en) | Seed lesion particle identification method and seed intelligent screening system | |
| Rao et al. | Monitoring and Detection of Plant Diseases Using Neural Networks | |
| Watcharabutsarakham et al. | Mobile-device based image processing for rice brown planthopper classification and outbreak monitoring |
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: 20190212 |