CN116882546A - Agricultural product traceability system and method - Google Patents

Agricultural product traceability system and method Download PDF

Info

Publication number
CN116882546A
CN116882546A CN202310649548.8A CN202310649548A CN116882546A CN 116882546 A CN116882546 A CN 116882546A CN 202310649548 A CN202310649548 A CN 202310649548A CN 116882546 A CN116882546 A CN 116882546A
Authority
CN
China
Prior art keywords
apple
transportation
output
node
container
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
Application number
CN202310649548.8A
Other languages
Chinese (zh)
Inventor
付菊芳
林兴相
林志钊
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Fujian Zhongke Zhongxin Intelligent Technology Co ltd
Original Assignee
Fujian Zhongke Zhongxin Intelligent Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Fujian Zhongke Zhongxin Intelligent Technology Co ltd filed Critical Fujian Zhongke Zhongxin Intelligent Technology Co ltd
Priority to CN202310649548.8A priority Critical patent/CN116882546A/en
Publication of CN116882546A publication Critical patent/CN116882546A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • G06F18/253—Fusion techniques of extracted features
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
    • G06Q10/083—Shipping
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Business, Economics & Management (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Economics (AREA)
  • Biophysics (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Biomedical Technology (AREA)
  • Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • General Business, Economics & Management (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Development Economics (AREA)
  • Tourism & Hospitality (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Marketing (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Game Theory and Decision Science (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention relates to the technical field of AI (advanced technology attachment), and discloses a tracing system and a tracing method for agricultural products, wherein the tracing system for agricultural products comprises: the infrared image acquisition module is used for acquiring infrared images of five angles of the apple container; the transportation environment information acquisition module is used for acquiring transportation environment information of each transportation node through which the apple container is transported; a transportation environment feature generation module for generating transportation environment features from the transportation environment information; a prediction module that predicts a damage proportion of apples within the apple tote based on the infrared image and the transport environment characteristics; according to the method, the transportation environment factors and time continuity factors of the apple fruits are comprehensively considered, and the damage proportion of the apple container is predicted through the neural network model, so that the loss of damage proportion information of agricultural products in the transportation stage is compensated.

Description

Agricultural product traceability system and method
Technical Field
The invention relates to the technical field of AI, in particular to a tracing system and a tracing method for agricultural products.
Background
The existing agricultural product tracing system relates to a plurality of links, the agricultural product is easy to damage in the transportation process, the problem that damage proportion information is lost in agricultural product tracing information occurs, and the agricultural product tracing system is unfavorable for merchants and consumers to trace agricultural product information.
Disclosure of Invention
The invention provides a tracing system and a tracing method for agricultural products, which solve the technical problems that agricultural products are easy to damage in the transportation process, damage proportion information is absent in agricultural product tracing information in the related art, and agricultural product information tracing by merchants and consumers is not facilitated.
The invention provides a tracing system for agricultural products, which comprises:
the infrared image acquisition module is used for acquiring infrared images of five angles of the apple container;
the transportation environment information acquisition module is used for acquiring transportation environment information of each transportation node through which the apple container is transported;
a transportation environment feature generation module for generating transportation environment features from the transportation environment information;
a prediction module that predicts a damage proportion of apples within the apple tote based on the infrared image and the transport environment characteristics;
the prediction module comprises: the first hiding layer, the second hiding layer and the classifier;
the first hidden layer comprises a convolution layer, a pooling layer, a full connection layer and a combiner;
the outputs of the five channels of the convolution layer serve as inputs of the five channels of the pooling layer;
the outputs of the five channels of the pooling layer are used as the inputs of the full-connection layer;
the full connection layer outputs a first feature vector;
a combiner of the t first hidden layer combines the first feature vector with the transport environment feature of the t transport node to obtain a second feature vector;
the second hidden layer comprises N LSTM units, and the input of the t LSTM unit isWherein->A second feature vector representing the output of the t first hidden layer;
the output of each LSTM unit of the second hidden layer is connected with a classifier, and the set of classification labels of the classifier is thatThe value range of the damage proportion of the apple container is 0-100 percent to carry out mean discretization,corresponding to a discretized damage ratio value;
further, the apple container is a carriage, and infrared images are respectively acquired by an infrared camera right above, left side, right front and right rear of the carriage of each transportation node;
further, the transportation environment information of the transportation node includes: the number of the apple boxes, the number of apples in the apple boxes, the weight of apples in the apple boxes of the current transportation node, the number of the current transportation node, the transportation time from the previous transportation node to the current transportation node, the temperature in the apple boxes of the current transportation node and the humidity in the apple boxes of the current transportation node;
further, the transportation environment characteristics comprise 7 dimensions, wherein the value of each dimension is respectively a serial number value of an apple container, a numerical value of an apple in the apple container, a weight value of an apple in the apple container of a current transportation node, a serial number value of the current transportation node, a transportation time length value from a previous transportation node to the current transportation node, a temperature value in the apple container of the current transportation node and a humidity value in the apple container of the current transportation node; the starting value of the serial number value of the apple container is 1, and the starting value of the serial number value of the current transportation node is 1;
further, the operation procedure of the t-th LSTM unit of the second hidden layer is as follows:
forgetting doorThe calculation formula of (2) is as follows:
wherein the method comprises the steps ofRepresenting the input of the t-th LSTM cell, i.e. the second eigenvector of the t-th first hidden layer output,represents the output of the t-1 th LSTM cell,>representation->Transfer to->Corresponding weight matrix, < >>Representation->Transfer to->Corresponding weight matrix, < >>Representing a bias term; by activating the function->(sigmoid function) will->The calculation result of (1) is defined between (0, 1);
further, an input doorThe calculation formula of (2) is as follows:
wherein the method comprises the steps ofRepresents the input of the t-th LSTM cell, < >>Represents the output of the t-1 th LSTM cell,>representation input +.>Transfer to->Corresponding weight matrix, < >>Representation->Transfer to->Corresponding weight matrix, < >>Representing a bias term; by activating the function->Will->The calculation result of (1) is defined between (0, 1); further, intermediate state->Can be expressed as follows: />
Wherein the method comprises the steps ofRepresents the input of the t-th LSTM cell, < >>Represents the output of the t-1 th LSTM cell,>representation input +.>Transfer to->Corresponding weight matrix, < >>Representation->Transfer to->Corresponding weight matrix, < >>Representing a bias term; by activating the function->Will->Is defined between (-1, 1); further, the methodOutput state->Expressed by the following formula:
wherein the method comprises the steps ofIs the output state transferred by the t-1 th LSTM,>、/>、/>is the calculation result of the forget gate, the input gate and the intermediate state;
indicating forgetfulness door->And output state of t-1 th LSTM +.>Point-by-point multiplication is performed to make->；
Representing I/O gate->And intermediate state->Performing point-by-point multiplication to obtain new candidate memory, namely new features needing to be memorized; />；
The output gate is expressed as:
wherein the method comprises the steps ofRepresents the input of the t-th LSTM cell, < >>Represents the output of the t-1 th LSTM cell,>representation input +.>Transfer to->Corresponding weight matrix, < >>Representation->Transfer to->Corresponding weight matrix, < >>Representing a bias term; by activating the function->Defining the calculation result between (0, 1); output->Can be expressed as follows:output door->And->Multiplying point by point to obtain the output +.>The method comprises the steps of carrying out a first treatment on the surface of the Definitions->；
Further, the first hidden layer is independently trained, when the first hidden layer is trained, the output of the first hidden layer is connected with a training classifier, and the set of classification labels of the training classifier is as followsThe value range of the damage proportion of the apple container is 0-100 percent and is subjected to mean discretization, and the apple container is->Corresponding to a discretized damage ratio value; the weight parameters of the first hidden layer are not updated in the process of training the prediction module;
the tracing system for the agricultural products executes the following steps:
step one, the method is used for collecting infrared images of five angles of an apple container;
step two, the method is used for collecting the transportation environment information of each transportation node through which the apple container is transported;
step three, it is used for producing the transportation environment characteristic according to the transportation environment information;
and step four, predicting the damage proportion of apples in the apple container based on the infrared image and the transportation environment characteristics.
The invention has the beneficial effects that: according to the method, the transportation environment factors and time continuity factors of the apple fruits are comprehensively considered, and the damage proportion of the apple container is predicted through the neural network model, so that the loss of damage proportion information of agricultural products in the transportation stage is compensated.
Drawings
FIG. 1 is a block diagram of a traceability system for agricultural products of the present invention;
fig. 2 is a flow chart of a method of tracing agricultural products of the present invention.
Detailed Description
The subject matter described herein will now be discussed with reference to example embodiments. It is to be understood that these embodiments are merely discussed so that those skilled in the art may better understand and implement the subject matter described herein and that changes may be made in the function and arrangement of the elements discussed without departing from the scope of the disclosure herein. Various examples may omit, replace, or add various procedures or components as desired. In addition, features described with respect to some examples may be combined in other examples as well.
As shown in fig. 1, a tracing system for agricultural products includes:
an infrared image acquisition module 101 for acquiring infrared images of five angles of the apple container;
in one embodiment of the invention, the apple container is a carriage, and infrared images are respectively acquired by an infrared camera right above, left side, right front and right rear of the carriage of each transportation node;
in one embodiment of the invention, the apple container is a packaging box, and infrared images are respectively acquired by an infrared camera right above, left side, right front and right rear of the packaging box at each transportation node.
A transportation environment information collection module 102 for collecting transportation environment information of each transportation node through which the apple totes are transported;
the transportation environment information of the transportation node includes: the number of apple boxes, the number of apples in the apple boxes, the weight of apples in the apple boxes of the current transportation node, the number of the current transportation node, the transportation time from the previous transportation node to the current transportation node (the transportation time of the first transportation node is 0), the temperature in the apple boxes of the current transportation node and the humidity in the apple boxes of the current transportation node;
a transportation environment feature generation module 103 for generating transportation environment features from the transportation environment information;
in one embodiment of the invention, the transportation environment characteristics comprise 7 dimensions, wherein the value of each dimension is respectively a serial number value of an apple container, a numerical value of an apple in the apple container, a weight value of an apple in the apple container of a current transportation node, a serial number value of the current transportation node, a transportation time length value from a previous transportation node to the current transportation node, a temperature value in the apple container of the current transportation node, and a humidity value in the apple container of the current transportation node; the starting value of the serial number value of the apple container is 1, and the starting value of the serial number value of the current transportation node is 1;
a prediction module 104 that predicts a damage proportion of apples within the apple tote based on the infrared images and the transport environment characteristics;
the prediction module comprises: the first hiding layer, the second hiding layer and the classifier;
the first hidden layer comprises a convolution layer, a pooling layer, a full connection layer and a combiner;
five channels of the convolution layer of the t first hidden layer are respectively input into five-angle infrared images of the apple container of the t transport node;
the outputs of the five channels of the convolution layer serve as inputs of the five channels of the pooling layer;
the outputs of the five channels of the pooling layer are used as the inputs of the full-connection layer;
the full connection layer outputs a first feature vector;
a combiner of the t first hidden layer combines the first feature vector with the transport environment feature of the t transport node to obtain a second feature vector; the second hidden layer comprises N LSTM units, and the input of the t LSTM unit isWherein->A second feature vector representing the output of the t first hidden layer; the operation process of the t LSTM unit is as follows: amnesia door->The calculation formula of (2) is as follows: />Wherein->Representing the input of the t-th LSTM cell, i.e. the second eigenvector of the t-th first hidden layer output,>represents the output of the t-1 th LSTM cell,>representation->Transfer to->Corresponding weight matrix, < >>Representation->Transfer to->Corresponding weight matrix, < >>Representing the bias term. By activating the function->(sigmoid function) will->Is defined between (0, 1). Input door->The calculation formula of (2) is as follows:wherein->Represents the input of the t-th LSTM cell, < >>Represents the output of the t-1 th LSTM cell,>representation input +.>Transfer to->Corresponding weight matrix, < >>Representation->Transfer to->Corresponding weight matrix, < >>Representing a bias term; by activating the function->Will->Is defined between (0, 1). Intermediate state->Can be expressed as follows: />Wherein->Represents the input of the t-th LSTM cell, < >>Represents the output of the t-1 th LSTM cell,>representation input +.>Transfer to->The corresponding weight matrix is used to determine the weight matrix,representation->Transfer to->Corresponding weight matrix, < >>Representing a bias term; by activating the function->Will->Is defined between (-1, 1). Output state->Expressed by the following formula: />
Wherein the method comprises the steps ofIs the output state transferred by the t-1 th LSTM,>、/>、/>is the calculation result of forget gate, input gate and intermediate state. />Indicating forgetfulness door->And output state of t-1 th LSTM +.>The multiplication is performed point by point,。
representing I/O gate->And intermediate state->And (5) carrying out point-by-point multiplication to obtain new candidate memory, namely new features needing to be memorized. />. The output gate is expressed as: />Wherein->Represents the input of the t-th LSTM cell, < >>Represents the output of the t-1 th LSTM cell,>representation input +.>Transfer to->Corresponding weight matrix, < >>Representation->Transfer to->Corresponding weight matrix, < >>Representing a bias term; by activating the function->The calculation result is defined between (0, 1).
Output ofCan be expressed as follows: />Output door->And->Multiplying point by point to obtain the output +.>. Definitions->。
The output of each LSTM unit of the second hidden layer is connected with a classifier, and the set of classification labels of the classifier is thatThe value range of the damage proportion of the apple container is 0-100 percent to carry out mean discretization,corresponding to a discretized damage ratio value;
in one embodiment of the invention, the first hidden layer is independently trained, and when the first hidden layer is trained, the output of the first hidden layer is connected with a training classifier, and the set of classification labels of the training classifier is thatThe value range of the damage proportion of the apple container is 0-100 percent to carry out mean discretization,corresponding to a discretized damage ratio value;
the weight parameters of the first hidden layer are not updated during the training of the prediction module.
Training of the neural network model is a conventional technical means, and is not described in detail herein;
as shown in fig. 2, a tracing method for agricultural products, through the tracing system for agricultural products, performs the following steps:
step one, the method is used for collecting infrared images of five angles of an apple container;
step two, the method is used for collecting the transportation environment information of each transportation node through which the apple container is transported;
step three, it is used for producing the transportation environment characteristic according to the transportation environment information;
predicting the damage proportion of apples in the apple container based on the infrared image and the transportation environment characteristics;
the invention has the beneficial effects that: according to the method, the transportation environment factors and time continuity factors of the apple fruits are comprehensively considered, and the damage proportion of the apple container is predicted through the neural network model, so that the loss of damage proportion information of agricultural products in the transportation stage is compensated.
The embodiment has been described above with reference to the embodiment, but the embodiment is not limited to the above-described specific implementation, which is only illustrative and not restrictive, and many forms can be made by those of ordinary skill in the art, given the benefit of this disclosure, are within the scope of this embodiment.

Claims (10)

1. A traceability system for agricultural products, comprising:
the infrared image acquisition module is used for acquiring infrared images of five angles of the apple container;
the transportation environment information acquisition module is used for acquiring transportation environment information of each transportation node through which the apple container is transported;
a transportation environment feature generation module for generating transportation environment features from the transportation environment information;
a prediction module that predicts a damage proportion of apples within the apple tote based on the infrared image and the transport environment characteristics;
the prediction module comprises: the first hiding layer, the second hiding layer and the classifier;
the first hidden layer comprises a convolution layer, a pooling layer, a full connection layer and a combiner;
the outputs of the five channels of the convolution layer serve as inputs of the five channels of the pooling layer;
the outputs of the five channels of the pooling layer are used as the inputs of the full-connection layer;
the full connection layer outputs a first feature vector;
a combiner of the t first hidden layer combines the first feature vector with the transport environment feature of the t transport node to obtain a second feature vector;
the second hidden layer comprises N LSTM units, and the input of the t LSTM unit isWherein->A second feature vector representing the output of the t first hidden layer;
the output of each LSTM unit of the second hidden layer is connected with a classifier, and the set of classification labels of the classifier is thatPerforming mean discretization on the value range of the damage proportion of the apple container from 0% to 100%, and performing->Corresponds to a discretized one of the damage ratio values.
2. The agricultural product tracing system according to claim 1, wherein the apple container is a carriage, and infrared images are respectively collected by infrared cameras right above, left side, right front and right rear of the carriage of each transportation node.
3. The agricultural product tracing system of claim 1, wherein the transportation environment information of the transportation node comprises: the number of the apple boxes, the number of apples in the apple boxes, the weight of apples in the apple boxes of the current transportation node, the number of the current transportation node, the transportation time from the previous transportation node to the current transportation node, the temperature in the apple boxes of the current transportation node and the humidity in the apple boxes of the current transportation node.
4. The agricultural product tracing system according to claim 1, wherein the transportation environment characteristics comprise 7 dimensions, and the value of each dimension is respectively a serial number value of an apple container, a numerical value of an apple in the apple container, a weight value of an apple in the apple container of a current transportation node, a serial number value of the current transportation node, a transportation time length value from a previous transportation node to the current transportation node, a temperature value in the apple container of the current transportation node, and a humidity value in the apple container of the current transportation node; the starting value of the serial number value of the apple container is 1, and the starting value of the serial number value of the current transportation node is 1.
5. The agricultural product tracing system of claim 1, wherein the operation procedure of the t-th LSTM unit of the second hidden layer is as follows:
forgetting doorThe calculation formula of (2) is as follows: />Wherein->Representing the input of the t-th LSTM cell, i.e. the second eigenvector of the t-th first hidden layer output,>represents the output of the t-1 th LSTM cell,>representation input +.>Transfer to->Corresponding weight matrix, < >>Representation->Transfer to->Corresponding weight matrix, < >>Representing a bias term; by activating the function->Will->Is defined between (0, 1).
6. The agricultural product tracing system according to claim 5, wherein the operation process of the t-th LSTM unit of the second hidden layer is as follows:
input doorThe calculation formula of (2) is as follows: />Wherein->Represents the input of the t-th LSTM cell, < >>Represents the output of the t-1 th LSTM cell,>representation input +.>Transfer to->Corresponding weight matrix, < >>Representation->Transfer to->Corresponding weight matrix, < >>Representing a bias term; by activating the function->Will->Is defined between (0, 1).
7. The agricultural product tracing system according to claim 6, wherein the operation process of the t-th LSTM unit of the second hidden layer is as follows: intermediate stateCan be expressed as follows: />Wherein->Represents the input of the t-th LSTM cell, < >>Represents the output of the t-1 th LSTM cell,>representation input +.>Transfer to->Corresponding weight matrix, < >>Representation->Transfer to->Corresponding weight matrix, < >>Representing a bias term; by activating the function->Will->Is defined between (-1, 1).
8. The agricultural product tracing system of claim 7, wherein the operation procedure of the t-th LSTM unit of the second hidden layer is as follows:
output stateExpressed by the following formula: />Wherein->Is the output state transferred by the t-1 th LSTM,>、/>、/>is the calculation result of the forget gate, the input gate and the intermediate state; />Indicating forgetfulness door->And output state of t-1 th LSTM +.>Point-by-point multiplication is performed to make->；/>Representing I/O gate->And intermediate state->Performing point-by-point multiplication to obtain new candidate memory, namely new features needing to be memorized; />；
The output gate is expressed as:wherein->Represents the input of the t-th LSTM cell, < >>Represents the output of the t-1 th LSTM cell,>representation input +.>Transfer to->The corresponding weight matrix is used to determine the weight matrix,representation->Transfer to->Corresponding weight matrix, < >>Representing a bias term; by activating the function->Defining the calculation result between (0, 1); output->Can be expressed as follows: />Output door->And->Multiplying point by point to obtain the output +.>The method comprises the steps of carrying out a first treatment on the surface of the Definitions->。
9. The agricultural product tracing system of claim 1, wherein the first hidden layer is independently trained, the output of the first hidden layer is connected to a training classifier when the first hidden layer is trained, and the set of classification labels of the training classifier isThe value range of the damage proportion of the apple container is 0-100 percent and is subjected to mean discretization, and the apple container is->Corresponding to a discretized damage ratio value; the weight parameters of the first hidden layer are not updated during the training of the prediction module.
10. A method for tracing agricultural products, characterized in that it comprises the following steps performed by a tracing system for agricultural products according to any one of claims 1 to 9:
step one, the method is used for collecting infrared images of five angles of an apple container;
step two, the method is used for collecting the transportation environment information of each transportation node through which the apple container is transported;
step three, it is used for producing the transportation environment characteristic according to the transportation environment information;
and step four, predicting the damage proportion of apples in the apple container based on the infrared image and the transportation environment characteristics.
CN202310649548.8A 2023-06-02 2023-06-02 Agricultural product traceability system and method Pending CN116882546A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310649548.8A CN116882546A (en) 2023-06-02 2023-06-02 Agricultural product traceability system and method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310649548.8A CN116882546A (en) 2023-06-02 2023-06-02 Agricultural product traceability system and method

Publications (1)

Publication Number Publication Date
CN116882546A true CN116882546A (en) 2023-10-13

Family

ID=88262750

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310649548.8A Pending CN116882546A (en) 2023-06-02 2023-06-02 Agricultural product traceability system and method

Country Status (1)

Country Link
CN (1) CN116882546A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117252346A (en) * 2023-11-15 2023-12-19 江西珉轩智能科技有限公司 Material traceability system and method

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117252346A (en) * 2023-11-15 2023-12-19 江西珉轩智能科技有限公司 Material traceability system and method
CN117252346B (en) * 2023-11-15 2024-02-13 江西珉轩智能科技有限公司 Material traceability system and method

Similar Documents

Publication Publication Date Title
Xu et al. Learning low-rank label correlations for multi-label classification with missing labels
CN116882546A (en) Agricultural product traceability system and method
CN111291809A (en) Processing device, method and storage medium
CN113160200B (en) Industrial image defect detection method and system based on multi-task twin network
Zhang et al. Self-taught semisupervised dictionary learning with nonnegative constraint
WO2020188794A1 (en) Video system, imaging device, and video processing device
CN115861246A (en) Product quality abnormity detection method and system applied to industrial Internet
CN113449815B (en) Abnormal packet detection method and system based on deep packet analysis
Lei et al. Hybrid Low-Order and Higher-Order Graph Convolutional Networks.
Dong et al. Research on image classification based on capsnet
Seo et al. Graph neural networks and implicit neural representation for near-optimal topology prediction over irregular design domains
Liu et al. Optimization-derived learning with essential convergence analysis of training and hyper-training
CN112380919A (en) Vehicle category statistical method
CN115578325A (en) Image anomaly detection method based on channel attention registration network
Das et al. Recurrent neural network based classification of fetal heart rate using cardiotocograph
Dai et al. Clothing recognition based on improved resnet18 model
Zhang et al. Network art image classification and print propagation extraction based on depth algorithm
CN112185543A (en) Construction method of medical induction data flow classification model
Panchal A novel approach of hyperspectral imaging classification using hybrid ConvNet
CN117404853B (en) External circulating water cooling system and method for tunnel boring machine
Galchonkov et al. Using a neural network in the second stage of the ensemble classifier to improve the quality of classification of objects in images
CN116720635B (en) Actual measurement data-based Guangxi oil tea estimation method
Herok et al. Cotton leaf disease identification using transfer learning
Zang et al. Joint Transfer Extreme Learning Machine with Cross-Domain Mean Approximation and Output Weight Alignment
EP4369251A1 (en) Systems and methods for tensorizing convolutional neural networks

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