CN109447165A - A kind of quality of agricultural product state identification method and device - Google Patents
A kind of quality of agricultural product state identification method and device Download PDFInfo
- Publication number
- CN109447165A CN109447165A CN201811299013.8A CN201811299013A CN109447165A CN 109447165 A CN109447165 A CN 109447165A CN 201811299013 A CN201811299013 A CN 201811299013A CN 109447165 A CN109447165 A CN 109447165A
- Authority
- CN
- China
- Prior art keywords
- agricultural product
- convolutional neural
- neural networks
- target
- targeted species
- 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
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the invention provides a kind of quality of agricultural product state identification method and devices, which comprises obtains the image of target agricultural product；Image is identified, determines the targeted species of target agricultural product；According to the targeted species of target agricultural product, the first convolutional neural networks to match with the targeted species are determined；By the first convolutional neural networks, identify that the quality state information of target agricultural product, quality state information include at least one of following information or a variety of: agricultural product maturity, agricultural product pest and disease damage, agricultural product kind.The quality state information of various target agricultural product can be accurately automatically determined out by the first convolutional neural networks, so that user is choosing each agricultural products, can be better understood by the quality state of chosen agricultural product, be substantially increased the consumption experience of user.
Description
Technical field
The present invention relates to technical field of image processing, more particularly to a kind of quality of agricultural product state identification method and dress
It sets.
Background technique
With the increasingly raising of people's living standard, the agricultural product of a large amount of various different cultivars enter the daily life of people
In work.For example, various new fruit and vegetables etc..Meanwhile as people's demands for quality of life are getting higher and higher, people couple
The quality requirements of the agricultural product such as the water fruits and vegetables usually bought are also higher and higher.
However, for most of ordinary consumer, and do not have the agricultural product knowledge of profession.So each choosing
When the agricultural product such as class water fruits and vegetables, tend not to identify or select the preferable or preferable agricultural product of quality that do well.Especially exist
When selecting individual novel agricultural product, often it is not easy to judge the quality state of the agricultural product, for example, selecting new varieties
When fruit, situations such as can not accurately judging the maturity of the fruit or whether there is pest and disease damage.
So how to allow the ordinary consumer without professional knowledge, chosen agricultural product can be accurately determined
Quality state, just become a urgent problem to be solved.
Summary of the invention
The embodiment of the present invention is designed to provide a kind of quality of agricultural product state identification method and device, can be realized pair
The automatic identification of the quality state of agricultural product.Specific technical solution is as follows:
The embodiment of the invention provides a kind of quality of agricultural product state identification methods, comprising:
Obtain the image of target agricultural product；
Described image is identified, determines the targeted species of the target agricultural product；
According to the targeted species of the target agricultural product, the first convolution nerve net to match with the targeted species is determined
Network, first convolutional neural networks are that the method training by deep learning obtains, for identification the targeted species
The convolutional neural networks of the quality state information of agricultural product, the quality state information include at least one of following information or
It is a variety of: agricultural product maturity, agricultural product pest and disease damage, agricultural product kind；
By first convolutional neural networks, the quality state information of the target agricultural product is identified.
Optionally, described that described image is identified, determine the targeted species of the target agricultural product, comprising:
Described image is inputted into the second convolutional neural networks, second convolutional neural networks are the method instruction by deep learning
It gets, for identification the convolutional neural networks of agricultural product type；
By second convolutional neural networks, the targeted species of the target agricultural product are determined.
Optionally, described that described image is identified, determine the targeted species of the target agricultural product, comprising:
The target image of the target agricultural product is extracted from described image；
The target image is identified, determines the targeted species of the target agricultural product.
Optionally, described by first convolutional neural networks, identify the quality shape of the target agricultural product
State information, comprising:
The target image is inputted into first convolutional neural networks；
The target image is identified by first convolutional neural networks, determines the described of the target agricultural product
Quality state information.
Optionally, the targeted species according to the target agricultural product are determined to match with the targeted species
First convolutional neural networks, comprising:
According to the targeted species of the target agricultural product, from preset multiple convolutional neural networks for variety classes agricultural product
In, determine the first convolutional neural networks to match with the targeted species.
Optionally, the targeted species according to the target agricultural product are determined to match with the targeted species
First convolutional neural networks, comprising:
The targeted species according to the target agricultural product determine multiple first convolution to match with the targeted species
Neural network, the multiple first convolutional neural networks are respectively the first convolutional Neural for identifying the different quality state information
Network.
Optionally, it is described identify the quality state information of the target agricultural product after, the method is also wrapped
It includes:
According to the quality state information, the related information to match with the quality state information, the related information are obtained
Including at least one of following information or a variety of: edible suggestion, storage method, kind information.
The embodiment of the invention also provides a kind of quality of agricultural product status identification means, comprising:
Module is obtained, for obtaining the image of target agricultural product；
First identification module determines the targeted species of the target agricultural product for identifying to described image；
Matching module determines to match with the targeted species for the targeted species according to the target agricultural product
One convolutional neural networks, first convolutional neural networks are that the method training by deep learning obtains, for identification institute
The convolutional neural networks of the quality state information of the agricultural product of targeted species are stated, the quality state information includes at least following letter
One of breath is a variety of: agricultural product maturity, agricultural product pest and disease damage, agricultural product kind；
Second identification module, for identifying the quality of the target agricultural product by first convolutional neural networks
Status information.
Optionally, first identification module, is specifically used for:
Described image is inputted into the second convolutional neural networks, second convolutional neural networks are the method instruction by deep learning
It gets, for identification the convolutional neural networks of agricultural product type；By second convolutional neural networks, the mesh is determined
Mark the targeted species of agricultural product.
Optionally, first identification module, is specifically used for:
The target image of the target agricultural product is extracted from described image；The target image is identified, is determined
The targeted species of the target agricultural product.
Optionally, second identification module, is specifically used for:
The target image is inputted into first convolutional neural networks；
The target image is identified by first convolutional neural networks, determines the described of the target agricultural product
Quality state information.
Optionally, the matching module, is specifically used for:
According to the targeted species of the target agricultural product, from preset multiple convolutional neural networks for variety classes agricultural product
In, determine the first convolutional neural networks to match with the targeted species.
Optionally, the matching module, is specifically used for:
The targeted species according to the target agricultural product determine multiple first convolution to match with the targeted species
Neural network, the multiple first convolutional neural networks are respectively the first convolutional Neural for identifying the different quality state information
Network.
Optionally, described device further include:
Relating module, for obtaining the related information to match with the quality state information according to the quality state information,
The related information includes at least one of following information or a variety of: edible suggestion, storage method, kind information.
A kind of quality of agricultural product state identification method and device provided in an embodiment of the present invention, available target agricultural product
Image, and the image is identified so that it is determined that targeted species of target agricultural product out.It is determined further according to targeted species
The first convolutional neural networks to match are identified by image of first convolutional neural networks to target agricultural product, thus
The quality state information of various target agricultural product can be accurately automatically determined out, so that user is choosing each agricultural products, especially
When it is novel kind or uncommon agricultural product, it can be better understood by the quality state of chosen agricultural product, greatly
The consumption experience of user is improved greatly.It avoids businessman from adulterating, cheats consumer, ensure that consumers' rights and interests.Certainly, implement
Any product of the invention or method must be not necessarily required to reach all the above advantage simultaneously.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is the flow chart of quality of agricultural product state identification method provided in an embodiment of the present invention；
Fig. 2 is the structure chart of quality of agricultural product status identification means provided in an embodiment of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Referring to Fig. 1, Fig. 1 is the flow chart of quality of agricultural product state identification method provided in an embodiment of the present invention, wherein wrapping
It includes:
Step 110, the image of target agricultural product is obtained.
The embodiment of the present invention can be applied to each class of electronic devices, for example, mobile phone, computer, server and server
Cloud platform constituted etc..
Electronic equipment can obtain the image of target agricultural product in several ways, for example, the camera shooting for passing through electronic equipment
First-class component shoots target agricultural product, to obtain the image of target agricultural product.Or can also by other means,
By the image input of target agricultural product or it is transmitted to electronic equipment.For example, user is clapped by mobile phone when electronic equipment is server
The image of target agricultural product is taken the photograph, it then can be by the image transmitting to server.Wherein, target agricultural product refer to that user needs
The agricultural product of its quality state are solved, for example, user is choosing agricultural product, and user's sense is emerging in an actual application scenarios
The agricultural product such as the interesting or unfamiliar novel water fruits and vegetables of user, can be used as target agricultural product.
Step 120, the image of target agricultural product is identified, determines the targeted species of target agricultural product.
Electronic equipment is after the image for obtaining target agricultural product, so that it may carry out image to the image of target agricultural product
Identification, so that it is determined that the specific targeted species of target agricultural product out, that is, determine that target agricultural product are any fruit or vegetables.
For example, determining that the targeted species of target agricultural product are apple, banana, salacca edulis salak, wax-apple, passion fruit, ball by image recognition
Stem wild cabbage, leek etc..
Specifically, the mode or algorithm of image recognition can there are many kinds of, in practical applications, can according to need progress
Selection can be applied in the embodiment of the present invention as long as can satisfy the effect in the embodiment of the present invention.
Preferably, in embodiments of the present invention, in order to fast and accurately determine the targeted species of target agricultural product,
Step 120, the image of target agricultural product is identified, determines the targeted species of target agricultural product, may include:
Step 121, the image of target agricultural product is inputted into the second convolutional neural networks, which is through too deep
Spend what the method training learnt obtained, for identification the convolutional neural networks of agricultural product type.
Step 122, by second convolutional neural networks, the targeted species of target agricultural product are determined.
Basic skills based on training convolutional neural networks can use a large amount of samples pictures, to the second convolutional Neural
Network is trained, and enables the second convolutional neural networks by convolution algorithm repeatedly, is identified and is saved various not of the same race
The agricultural product of class respectively possessed by feature.And it is based on obtained feature, image recognition is carried out to subsequent inputted picture, and
It is capable of determining that the type of agricultural product contained in the picture.The second convolutional neural networks are enabled to have identification not of the same race
The ability of the agricultural product of class.
So after the image of target agricultural product is inputted the second convolutional neural networks, the second convolutional neural networks energy
Enough images to target agricultural product carry out corresponding operation, obtain the feature of target agricultural product included in the image, and base
It is matched in saved various features with acquired feature, so that it is determined that going out the targeted species of the target agricultural product.
In an actual application scenarios, server or cloud platform obtain the figure of target agricultural product captured by user
Then the image of the target agricultural product can be inputted and have already passed through the second trained convolutional neural networks by picture, the second convolution mind
The features of agricultural product in the image can be obtained by calculating through network, by with the feature progress that is saved when preparatory training
Match, can determine that acquired feature matches with the type pre-saved for the feature of salacca edulis salak, so as to identify
The targeted species of the target agricultural product are salacca edulis salak.
Step 130, according to the targeted species of target agricultural product, the first convolutional Neural to match with targeted species is determined
Network, the first convolutional neural networks are that the method training by deep learning obtains, for identification the agricultural product of targeted species
Quality state information convolutional neural networks, quality state information includes at least one of following information or a variety of: agricultural production
Product maturity, agricultural product pest and disease damage, agricultural product kind.Wherein, agricultural product kind refers to, the kind of same agricultural product subdivision,
Such as, apple can be divided into multiple kinds such as Fuji apple, Red Star, state's light.
Electronic equipment is by the second convolutional neural networks, after the targeted species for determining target agricultural product, so that it may root
According to the targeted species, to select that the agricultural product of the targeted species can be carried out on the convolutional Neural net of quality state information identification
Network, as the first convolutional neural networks to match with targeted species.
Specifically, step 130 determines match with targeted species first according to the targeted species of target agricultural product
Convolutional neural networks may include:
According to the targeted species of target agricultural product, from preset multiple convolutional neural networks for variety classes agricultural product,
Determine the first convolutional neural networks to match with targeted species.
In practical application, being directed to each agricultural product, the agricultural product for being directed to the type can be preset, are carried out
The convolutional neural networks of quality state information identification.Each convolutional neural networks is just for a type of agricultural product.So
The targeted species according to target agricultural product are needed, from multiple convolutional neural networks, selection can be to the target of the targeted species
First convolutional neural networks of agricultural product progress quality state information identification.
For example, the water that can be familiar with for ordinary consumers such as salacca edulis salak, wax-apple, passion fruit, kohlrabi, leeks
Fruits and vegetables type is preset with the convolutional neural networks for each agricultural product respectively.When having confirmed target agricultural product
When targeted species are salacca edulis salak, then the convolutional neural networks of salacca edulis salak can will be directed to as the first convolutional neural networks.
Each preset convolutional neural networks all carries out a large amount of sample training for a kind of agricultural product, so that
The convolutional neural networks can have the ability that the agricultural product of the type are carried out with the identification of quality state information.
For example, being directed to the convolutional neural networks of salacca edulis salak, the picture of the salacca edulis salak of a large amount of different freshness can be used, or
Person largely has the picture of the salacca edulis salak of different pest and disease damages, or the largely picture of the salacca edulis salak of different cultivars, to initial volume
Product neural network is trained.So that the initial convolutional neural networks can have identify salacca edulis salak freshness,
The ability of pest and disease damage or kind.Wherein, freshness can be indicated using preset rank or percentage, for example, by fresh
Degree is divided into 5 ranks, and 1 grade of expression is most fresh, and 5 grades of expressions are addled.Then in training convolutional neural networks, for each
A rank is all trained using a large amount of picture sample, so that the convolutional neural networks can recognize that difference is fresh
Spend the agricultural product of rank.
In embodiments of the present invention, quality state information can there are many, such as may include freshness, pest and disease damage, product
Kind etc..
If in a convolutional neural networks, while realizing the identification to multiple quality state information, may make
The convolutional neural networks it is excessively complicated, the process of training process and identification can all become sufficiently complex, to reduce efficiency.
To solve the above-mentioned problems, optionally, in embodiments of the present invention, step 130, according to the mesh of target agricultural product
Type is marked, determines the first convolutional neural networks to match with targeted species, comprising:
Step 131, according to the targeted species of target agricultural product, multiple first convolutional Neurals to match with targeted species are determined
Network, multiple first convolutional neural networks are respectively the first convolutional neural networks for identifying different quality state information.
For each agricultural product, it can preset and be provided with multiple convolutional neural networks, wherein each convolutional Neural net
Network can detecte out the different quality state information of the agricultural product respectively.For example, being directed to a kind of this agricultural product of salacca edulis salak, it is provided with
3 convolutional neural networks, one for identification the freshness of salacca edulis salak, one for identification salacca edulis salak whether there is all kinds of disease pests
Evil, the kind of another salacca edulis salak for identification.
After electronic equipment determines the targeted species of target agricultural product, so that it may by the agriculture of the targeted species for identification
Multiple convolutional neural networks of the different quality state information of product are used as the first convolutional neural networks.For example, when electronics is set
It is standby determine that target agricultural product are salacca edulis salak after, then can by above-mentioned for 3 convolutional neural networks set by salacca edulis salak,
It is used as the first convolutional neural networks.
Step 140, by the first convolutional neural networks, the quality state information of target agricultural product is identified.
After electronic equipment determines the first convolutional neural networks, so that it may by the image of target agricultural product input this
In one convolutional neural networks, to carry out convolution algorithm by image of first convolutional neural networks to target agricultural product, and most
The quality state information of target agricultural product contained in the image is determined eventually.
When determined multiple first convolutional neural networks, then it is more the image of target agricultural product can be inputted this respectively
In a first convolutional neural networks, and different quality state information is determined respectively.
In an actual application scenarios, then user will by the image of the terminals photographic subjects agricultural product such as mobile phone
The image is sent to server or cloud platform, and server or cloud platform go out target agricultural product by the second convolution neural network recognization
Targeted species, e.g., determine the target agricultural product targeted species be salacca edulis salak.Then server or cloud platform determine with
The first convolutional neural networks of one or more that the targeted species match, and the image is inputted into the first convolutional neural networks,
Such as the image of the salacca edulis salak of user's shooting is inputted into multiple the first convolutional neural networks for salacca edulis salak respectively, then can known
Not Chu the corresponding salacca edulis salak of the image freshness, if there are pest and disease damage and the kinds of the salacca edulis salak.Server or Yun Ping
Above-mentioned quality state information can be sent to the terminal of user by platform, allow users to recognize that the quality state is believed in time
Breath.
It is easily understood that different freshness, or the target agricultural product with pest and disease damage or different cultivars, appearance
Feature necessarily will be different, but ordinary consumer is when being chosen, may not be able to by oneself knowledge experience into
Row accurately judgement.In embodiments of the present invention, the quality of agricultural product state identification method provided through the embodiment of the present invention, energy
Enough images to target agricultural product carry out image recognition using convolutional neural networks, various so as to accurately automatically determine out
The quality state information of target agricultural product, so that user is choosing each agricultural products, it is especially novel kind or uncommon
Agricultural product when, the quality state of chosen agricultural product can be better understood by, substantially increase the consumption experience of user.It keeps away
Exempt from businessman to adulterate, cheats consumer, ensure that consumers' rights and interests.
In conjunction with above-described embodiment, in practical applications, the image of target agricultural product is identified in order to further increase
Accuracy.Optionally, in quality of agricultural product state identification method provided in an embodiment of the present invention, step 120, to target agriculture
The image of product is identified, determines the targeted species of the target agricultural product, comprising:
Step 120a extracts the target image of target agricultural product from the image of target agricultural product.
In practical applications, the image of target agricultural product is generally captured by user, so, it is easy to wrap in the image
Contain the various background images for being easy to produce interference.For example, in addition to containing target agricultural product in the image of the target agricultural product, also
Other some environmental backgrounds can be contained.In some cases, excessively complex environment background can carrying out image recognition,
Interference is generated, causes to identify exception or mistake.
So preliminary knowledge can be carried out to the image first after electronic equipment obtains the images of target agricultural product
Not, the image range of target agricultural product included in the image is determined, for example, can be in such a way that frame selects, in the figure
As center selects the image range at the place of target agricultural product.Specifically, tentatively being identified to the image, can use existing
Identification methods, for example, can be the identification technology based on shape feature, the identification technology based on color character and base
In various ways such as the identification technologies of textural characteristics.
After determining the image range of target agricultural product included in image, so that it may from the image, by scratching
The modes such as figure or copy pixel point, extract the target image only containing target agricultural product, to eliminate more from the image
In background image.For example, the image range that institute's frame selects is carried out to scratch figure or duplication, using obtained small figure as target figure
Picture.
Step 120b, identifies target image, determines the targeted species of target agricultural product.
After having obtained target image, then directly the target image can be identified, such as the target image is defeated
Enter the second convolutional neural networks, so that logical can cross the targeted species that second convolutional neural networks determine target agricultural product.
In embodiments of the present invention, by extracting target image, the background image that may cause interference is eliminated, so that knowing
Other result can be more accurate, and improves recognition efficiency.
Equally, after extracting target image, when the quality using the first convolution neural network recognization to target agricultural product
When status information is identified, then the target image can also be directly used.So agricultural product product provided in an embodiment of the present invention
In matter state identification method, step 140, by the first convolutional neural networks, the quality state of target agricultural product is identified
Information may include:
Target image is inputted the first convolutional neural networks by step 140a.
Step 140b identifies target image by the first convolutional neural networks, determines the product of target agricultural product
Matter status information.
It, equally can be to avoid background image to identifying when the target image that is extracted inputs the first convolutional neural networks
Interference caused by journey.And the calculation amount that can reduce by the first convolutional neural networks, improves recognition efficiency.
Optionally, in order to further increase customer consumption experience, in embodiments of the present invention, in step 140, it identifies
After the quality state information of the target agricultural product, quality of agricultural product state identification method provided in an embodiment of the present invention
Can also include:
Step 150, according to quality state information, the related information to match with the quality state information is obtained, related information is extremely
It less include one of following information or a variety of: edible suggestion, storage method, kind information.
Electronic equipment passes through one or more first convolutional neural networks, determines the quality state information of target agricultural product
Later, so that it may according to the quality state information, be based on preset matching relationship, determine pass corresponding with quality state information
Join information.Specifically, preset matching relationship can be the form of table.Different quality state information, corresponds in the table
There is different related informations.For example, different freshness and whether there is pest and disease damage, different edible suggestions can be corresponded to and deposited
Method for storing, e.g., freshness be not high, and the salacca edulis salak of pest and disease damage is not present, then corresponding edible suggestion is eats as early as possible, storage side
Method is must be stored refrigerated.Freshness is not high, and there are the salacca edulis salaks of more serious pest and disease damage, then corresponding edible suggestion is unsuitable
Edible, storage method is that should not store.Alternatively, edible suggestion can also include the specific eating method of the target agricultural product, example
Such as, peeling skill, matching method etc..
Agricultural product kind in quality state information can be corresponding with corresponding kind information, including source area, cultivation side
Method, if the kind is original variety, or the kind obtained by engrafting and cultivating, or obtained by transgene method
The information such as kind.
In an actual application scenarios, the image of the target agricultural product of shooting is sent to server or Yun Ping by user
Platform, server or cloud platform obtain quality state information by the identification to image, and determine corresponding related information.So
Quality state information and corresponding related information can be sent to the terminals such as the mobile phone of user together afterwards.To which user can obtain
The more comprehensive information in relation to the target agricultural product, the understanding target agricultural product that can be more deep, largely
On improve the consumption experience of user.
Referring to fig. 2, Fig. 2 is the structure chart of quality of agricultural product status identification means provided in an embodiment of the present invention, wherein wrapping
It includes:
Module 201 is obtained, for obtaining the image of target agricultural product；
First identification module 202 determines the targeted species of the target agricultural product for identifying to described image；
Matching module 203 is determined to match with the targeted species for the targeted species according to the target agricultural product
First convolutional neural networks, first convolutional neural networks are that the method training by deep learning obtains, for identification
The convolutional neural networks of the quality state information of the agricultural product of the targeted species, the quality state information include at least following
One of information is a variety of: agricultural product maturity, agricultural product pest and disease damage, agricultural product kind；
Second identification module 204, for identifying the product of the target agricultural product by first convolutional neural networks
Matter status information.
In embodiments of the present invention, first obtain target agricultural product image, and the image is identified so that it is determined that
The targeted species of target agricultural product out.The first convolutional neural networks to match are determined further according to targeted species, pass through first
Convolutional neural networks identify the image of target agricultural product, so as to accurately automatically determine out various target agricultural product
Quality state information so that user is choosing each agricultural products, when especially novel kind or uncommon agricultural product,
It can be better understood by the quality state of chosen agricultural product, substantially increase the consumption experience of user.Avoid businessman with secondary
It substitutes the bad for the good, cheats consumer, ensure that consumers' rights and interests.
Optionally, in quality of agricultural product status identification means provided in an embodiment of the present invention, first identification module
202, it is specifically used for:
Described image is inputted into the second convolutional neural networks, second convolutional neural networks are the method instruction by deep learning
It gets, for identification the convolutional neural networks of agricultural product type；By second convolutional neural networks, the mesh is determined
Mark the targeted species of agricultural product.
Optionally, in quality of agricultural product status identification means provided in an embodiment of the present invention, first identification module
202, it is specifically used for:
The target image of the target agricultural product is extracted from described image；The target image is identified, is determined
The targeted species of the target agricultural product.
Optionally, in quality of agricultural product status identification means provided in an embodiment of the present invention, second identification module
204, it is specifically used for:
The target image is inputted into first convolutional neural networks；
The target image is identified by first convolutional neural networks, determines the described of the target agricultural product
Quality state information.
Optionally, in quality of agricultural product status identification means provided in an embodiment of the present invention, the matching module 203,
It is specifically used for:
According to the targeted species of the target agricultural product, from preset multiple convolutional neural networks for variety classes agricultural product
In, determine the first convolutional neural networks to match with the targeted species.
Optionally, in quality of agricultural product status identification means provided in an embodiment of the present invention, the matching module 203,
It is specifically used for:
The targeted species according to the target agricultural product determine multiple first convolution to match with the targeted species
Neural network, the multiple first convolutional neural networks are respectively the first convolutional Neural for identifying the different quality state information
Network.
Optionally, in quality of agricultural product status identification means provided in an embodiment of the present invention, described device further include:
Relating module, for obtaining the related information to match with the quality state information according to the quality state information,
The related information includes at least one of following information or a variety of: edible suggestion, storage method, kind information.
It should be noted that, in this document, relational terms such as first and second and the like are used merely to a reality
Body or operation are distinguished with another entity or operation, are deposited without necessarily requiring or implying between these entities or operation
In any actual relationship or order or sequence.Moreover, the terms "include", "comprise" or its any other variant are intended to
Non-exclusive inclusion, so that the process, method, article or equipment including a series of elements is not only wanted including those
Element, but also including other elements that are not explicitly listed, or further include for this process, method, article or equipment
Intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that
There is also other identical elements in process, method, article or equipment including the element.
Each embodiment in this specification is all made of relevant mode and describes, same and similar portion between each embodiment
Dividing may refer to each other, and each embodiment focuses on the differences from other embodiments.Especially for device reality
For applying example, since it is substantially similar to the method embodiment, so being described relatively simple, related place is referring to embodiment of the method
Part explanation.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the scope of the present invention.It is all
Any modification, equivalent replacement, improvement and so within the spirit and principles in the present invention, are all contained in protection scope of the present invention
It is interior.
Claims (10)
1. a kind of quality of agricultural product state identification method characterized by comprising
Obtain the image of target agricultural product；
Described image is identified, determines the targeted species of the target agricultural product；
According to the targeted species of the target agricultural product, the first convolution nerve net to match with the targeted species is determined
Network, first convolutional neural networks are that the method training by deep learning obtains, for identification the targeted species
The convolutional neural networks of the quality state information of agricultural product, the quality state information include at least one of following information or
It is a variety of: agricultural product maturity, agricultural product pest and disease damage, agricultural product kind；
By first convolutional neural networks, the quality state information of the target agricultural product is identified.
2. determining the target the method according to claim 1, wherein described identify described image
The targeted species of agricultural product, comprising:
Described image is inputted into the second convolutional neural networks, second convolutional neural networks are the method instruction by deep learning
It gets, for identification the convolutional neural networks of agricultural product type；
By second convolutional neural networks, the targeted species of the target agricultural product are determined.
3. method according to claim 1 or 2, which is characterized in that it is described that described image is identified, determine the mesh
Mark the targeted species of agricultural product, comprising:
The target image of the target agricultural product is extracted from described image；
The target image is identified, determines the targeted species of the target agricultural product.
4. according to the method described in claim 3, identifying it is characterized in that, described by first convolutional neural networks
The quality state information of the target agricultural product, comprising:
The target image is inputted into first convolutional neural networks；
The target image is identified by first convolutional neural networks, determines the described of the target agricultural product
Quality state information.
5. the method according to claim 1, wherein the targeted species according to the target agricultural product, really
Make the first convolutional neural networks to match with the targeted species, comprising:
According to the targeted species of the target agricultural product, from preset multiple convolutional neural networks for variety classes agricultural product
In, determine the first convolutional neural networks to match with the targeted species.
6. the method according to claim 1, wherein the targeted species according to the target agricultural product, really
Make the first convolutional neural networks to match with the targeted species, comprising:
The targeted species according to the target agricultural product determine multiple first convolution to match with the targeted species
Neural network, the multiple first convolutional neural networks are respectively the first convolutional Neural for identifying the different quality state information
Network.
7. method according to claim 1 or 6, which is characterized in that identified described in the target agricultural product described
After quality state information, the method also includes:
According to the quality state information, the related information to match with the quality state information, the related information are obtained
Including at least one of following information or a variety of: edible suggestion, storage method, kind information.
8. a kind of quality of agricultural product status identification means characterized by comprising
Module is obtained, for obtaining the image of target agricultural product；
First identification module determines the targeted species of the target agricultural product for identifying to described image；
Matching module determines to match with the targeted species for the targeted species according to the target agricultural product
One convolutional neural networks, first convolutional neural networks are that the method training by deep learning obtains, for identification institute
The convolutional neural networks of the quality state information of the agricultural product of targeted species are stated, the quality state information includes at least following letter
One of breath is a variety of: agricultural product maturity, agricultural product pest and disease damage, agricultural product kind；
Second identification module, for identifying the quality of the target agricultural product by first convolutional neural networks
Status information.
9. device according to claim 8, which is characterized in that first identification module is specifically used for:
Described image is inputted into the second convolutional neural networks, second convolutional neural networks are the method instruction by deep learning
It gets, for identification the convolutional neural networks of agricultural product type；
By second convolutional neural networks, the targeted species of the target agricultural product are determined.
10. device according to claim 8, which is characterized in that described device further include:
Relating module, for obtaining the related information to match with the quality state information according to the quality state information,
The related information includes at least one of following information or a variety of: edible suggestion, storage method, kind information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811299013.8A CN109447165A (en) | 2018-11-02 | 2018-11-02 | A kind of quality of agricultural product state identification method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811299013.8A CN109447165A (en) | 2018-11-02 | 2018-11-02 | A kind of quality of agricultural product state identification method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109447165A true CN109447165A (en) | 2019-03-08 |
Family
ID=65550351
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811299013.8A Pending CN109447165A (en) | 2018-11-02 | 2018-11-02 | A kind of quality of agricultural product state identification method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109447165A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110414532A (en) * | 2019-06-21 | 2019-11-05 | 广州利科科技有限公司 | A kind of commodity image recognition methods based on deep learning |
| CN110659676A (en) * | 2019-09-09 | 2020-01-07 | 卓尔智联(武汉)研究院有限公司 | Information processing method, device and storage medium |
| CN110796642A (en) * | 2019-10-09 | 2020-02-14 | 陈浩能 | Method for determining fruit quality degree and related product |
| CN111767802A (en) * | 2020-06-05 | 2020-10-13 | 京东数字科技控股有限公司 | Method and device for detecting abnormal state of object |
| CN112464762A (en) * | 2020-11-16 | 2021-03-09 | 中国科学院合肥物质科学研究院 | Agricultural product screening system and method based on image processing |
| CN112579882A (en) * | 2020-10-28 | 2021-03-30 | 深圳市农产品质量安全检验检测中心(深圳市动物疫病预防控制中心) | Agricultural product type identification method |
| CN116465888A (en) * | 2023-03-17 | 2023-07-21 | 南通锐越信息科技有限公司 | Agricultural product quality Internet of things monitoring method and device based on machine visual effect, electronic equipment and computer readable storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106991427A (en) * | 2017-02-10 | 2017-07-28 | 海尔优家智能科技（北京）有限公司 | The recognition methods of fruits and vegetables freshness and device |
| CN108224894A (en) * | 2018-01-08 | 2018-06-29 | 合肥美的智能科技有限公司 | Food material freshness identification method and device based on deep learning, refrigerator and medium |
| CN109190476A (en) * | 2018-08-02 | 2019-01-11 | 福建工程学院 | A kind of method and device of vegetables identification |
-
2018
- 2018-11-02 CN CN201811299013.8A patent/CN109447165A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106991427A (en) * | 2017-02-10 | 2017-07-28 | 海尔优家智能科技（北京）有限公司 | The recognition methods of fruits and vegetables freshness and device |
| CN108224894A (en) * | 2018-01-08 | 2018-06-29 | 合肥美的智能科技有限公司 | Food material freshness identification method and device based on deep learning, refrigerator and medium |
| CN109190476A (en) * | 2018-08-02 | 2019-01-11 | 福建工程学院 | A kind of method and device of vegetables identification |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110414532A (en) * | 2019-06-21 | 2019-11-05 | 广州利科科技有限公司 | A kind of commodity image recognition methods based on deep learning |
| CN110414532B (en) * | 2019-06-21 | 2021-08-06 | 广州利科科技有限公司 | Commodity image identification method based on deep learning |
| CN110659676A (en) * | 2019-09-09 | 2020-01-07 | 卓尔智联(武汉)研究院有限公司 | Information processing method, device and storage medium |
| CN110796642A (en) * | 2019-10-09 | 2020-02-14 | 陈浩能 | Method for determining fruit quality degree and related product |
| CN111767802A (en) * | 2020-06-05 | 2020-10-13 | 京东数字科技控股有限公司 | Method and device for detecting abnormal state of object |
| CN111767802B (en) * | 2020-06-05 | 2024-02-06 | 京东科技控股股份有限公司 | Method and device for detecting abnormal state of object |
| CN112579882A (en) * | 2020-10-28 | 2021-03-30 | 深圳市农产品质量安全检验检测中心(深圳市动物疫病预防控制中心) | Agricultural product type identification method |
| CN112464762A (en) * | 2020-11-16 | 2021-03-09 | 中国科学院合肥物质科学研究院 | Agricultural product screening system and method based on image processing |
| CN116465888A (en) * | 2023-03-17 | 2023-07-21 | 南通锐越信息科技有限公司 | Agricultural product quality Internet of things monitoring method and device based on machine visual effect, electronic equipment and computer readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109447165A (en) | A kind of quality of agricultural product state identification method and device | |
| US20210192247A1 (en) | Systems and methods for electronically identifying plant species | |
| CN110267119A (en) | The evaluation method and relevant device of video highlight degree | |
| CN110291538A (en) | Image recognition of filtered image classification output distribution | |
| CN110598700B (en) | Object display method and device, storage medium and electronic device | |
| CN106294483A (en) | For mating data processing method and the device of mouthpiece for product brand | |
| CN110674736A (en) | Method, device, server and storage medium for identifying freshness of food materials | |
| CN109992781A (en) | Processing, device, storage medium and the processor of text feature | |
| CN108198043A (en) | A kind of facial skin care product recommended based on user recommend method | |
| CN110119479A (en) | A kind of restaurant recommendation method, apparatus, equipment and readable storage medium storing program for executing | |
| CN106442329A (en) | Winter wheat leaf area index estimation method based on canopy image parameters | |
| Tanahashi et al. | Natsumushi: Image measuring software for entomological studies | |
| CN108830154A (en) | A kind of food nourishment composition detection method and system based on binocular camera | |
| CN110795959A (en) | Intelligent identification method and device for sliced food materials | |
| Correia et al. | Breaking the mould an evolutionary quest for innovation through style change | |
| Ignacio et al. | A YOLOv5-based deep learning model for in-situ detection and maturity grading of mango | |
| CN110427870A (en) | Eye image identification method, Model of Target Recognition training method and device | |
| CN110796148A (en) | Litchi insect pest monitoring and identifying system and litchi insect pest monitoring and identifying method | |
| Thorupunoori et al. | Camera Based Drunks Detection Mechanism Integrated with DL (Deep Learning) | |
| Banerjee et al. | A Deep Dive into Precision Horticulture: Unravelling Olive Peacock Spot Intensity with Hybrid Deep Learning | |
| Maheshwari et al. | Performance analysis of mango leaf disease using machine learning technique | |
| CN110399516A (en) | Image processing method and device, readable storage medium and electronic equipment | |
| WO2023088079A1 (en) | Display method and display system for plant disease diagnosis information, and readable storage medium | |
| Fan et al. | Computational analysis and eye movement experiments of white space in Chinese paintings | |
| CN109726648A (en) | A kind of facial image recognition method and device based on machine 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 |
Application publication date: 20190308 |
|
| RJ01 | Rejection of invention patent application after publication |