CN116449713A - Tobacco leaf baking simulation method and system based on high-temperature baking room - Google Patents
Tobacco leaf baking simulation method and system based on high-temperature baking room Download PDFInfo
- Publication number
- CN116449713A CN116449713A CN202310429012.5A CN202310429012A CN116449713A CN 116449713 A CN116449713 A CN 116449713A CN 202310429012 A CN202310429012 A CN 202310429012A CN 116449713 A CN116449713 A CN 116449713A
- Authority
- CN
- China
- Prior art keywords
- tobacco
- baking
- sample
- tobacco leaf
- target
- 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
- 235000002637 Nicotiana tabacum Nutrition 0.000 title claims abstract description 540
- 238000000034 method Methods 0.000 title claims abstract description 45
- 238000004088 simulation Methods 0.000 title claims abstract description 22
- 244000061176 Nicotiana tabacum Species 0.000 title description 2
- 241000208125 Nicotiana Species 0.000 claims abstract description 538
- 238000004458 analytical method Methods 0.000 claims abstract description 265
- 210000002435 tendon Anatomy 0.000 claims abstract description 83
- 238000005070 sampling Methods 0.000 claims abstract description 9
- 238000010276 construction Methods 0.000 claims description 49
- 238000012360 testing method Methods 0.000 claims description 27
- 238000012549 training Methods 0.000 claims description 24
- 238000012795 verification Methods 0.000 claims description 17
- 239000003086 colorant Substances 0.000 claims description 10
- 238000011176 pooling Methods 0.000 claims description 10
- 238000007781 pre-processing Methods 0.000 claims description 10
- 238000004383 yellowing Methods 0.000 claims description 7
- 238000013527 convolutional neural network Methods 0.000 claims description 5
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 238000010801 machine learning Methods 0.000 claims description 4
- 230000000694 effects Effects 0.000 abstract description 7
- 238000005516 engineering process Methods 0.000 abstract description 5
- 230000008569 process Effects 0.000 description 8
- 239000011159 matrix material Substances 0.000 description 6
- 230000002787 reinforcement Effects 0.000 description 6
- 230000006870 function Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 230000001276 controlling effect Effects 0.000 description 3
- 238000002372 labelling Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 238000003908 quality control method Methods 0.000 description 2
- 125000003118 aryl group Chemical group 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- QDOXWKRWXJOMAK-UHFFFAOYSA-N dichromium trioxide Chemical compound O=[Cr]O[Cr]=O QDOXWKRWXJOMAK-UHFFFAOYSA-N 0.000 description 1
- 238000001035 drying Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000003306 harvesting Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 235000016709 nutrition Nutrition 0.000 description 1
- 230000035764 nutrition Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011946 reduction process Methods 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 210000003462 vein Anatomy 0.000 description 1
Classifications
-
- A—HUMAN NECESSITIES
- A24—TOBACCO; CIGARS; CIGARETTES; SIMULATED SMOKING DEVICES; SMOKERS' REQUISITES
- A24B—MANUFACTURE OR PREPARATION OF TOBACCO FOR SMOKING OR CHEWING; TOBACCO; SNUFF
- A24B3/00—Preparing tobacco in the factory
- A24B3/10—Roasting or cooling tobacco
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
-
- 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
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Automation & Control Theory (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Medical Informatics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Manufacture Of Tobacco Products (AREA)
Abstract
The present disclosure provides a tobacco leaf baking simulation method and system based on a high temperature baking room, and relates to the technical field of tobacco leaf baking, wherein the method comprises: obtaining a target baking type; acquiring a temperature information set, a humidity information set and baking time; constructing a corresponding target baking control model to obtain a temperature control parameter and a humidity control parameter; sampling and selecting tobacco leaves, and collecting image information; obtaining a tobacco leaf analysis result and a tobacco rib analysis result; when the tobacco leaf analysis result and the tobacco tendon analysis result reach the preset tobacco leaf color threshold value and the preset tobacco tendon color threshold value, the next baking stage is performed or baking is completed, the technical problem that the tobacco leaf baking quality is poor due to insufficient control accuracy of baking temperature and humidity in the existing tobacco leaf baking technology is solved, the control accuracy of temperature and humidity is improved, and the technical effect of improving the tobacco leaf baking quality is achieved.
Description
Technical Field
The disclosure relates to the technical field of tobacco leaf baking, in particular to a tobacco leaf baking simulation method and system based on a high-temperature baking room.
Background
The automatic control of the tobacco leaf baking process is a key technology for guaranteeing the tobacco leaf quality, and the traditional mode is to collect data by using a dry-wet ball glass thermometer, so that the manual long-time monitoring is realized, the labor intensity is high, the operation procedure is complex, the temperature and humidity distribution in a baking room is uneven, and the tobacco leaf baking quality level is low.
At present, the existing tobacco leaf baking technology has the technical problem of poor tobacco leaf baking quality due to insufficient control accuracy of baking temperature and humidity.
Disclosure of Invention
The present disclosure provides a tobacco leaf baking simulation method and system based on a high temperature baking room, which are used for solving the technical problem of poor tobacco leaf baking quality caused by insufficient control accuracy of baking temperature and humidity in the existing tobacco leaf baking technology.
According to a first aspect of the present disclosure, there is provided a tobacco leaf curing simulation method based on a high temperature curing barn, comprising: acquiring the tobacco leaf type of the current tobacco leaf baking test and a target baking stage for baking the target tobacco leaf type, and taking the tobacco leaf type as the target baking type; acquiring temperature and humidity information in the last preset time period and the current baking time to obtain a temperature information set, a humidity information set and the baking time; constructing and acquiring a corresponding target baking control model according to the target baking type, inputting the baking time, the temperature information set and the humidity information set into the target baking control model, acquiring temperature control parameters and humidity control parameters, and controlling the temperature and the humidity; sampling and selecting the tobacco leaves which are baked currently according to the preset time period to obtain a plurality of sample tobacco leaves, and collecting image information of the plurality of sample tobacco leaves to obtain an image information set; preprocessing the image information set, and inputting the image information set into a tobacco leaf analysis module and a tobacco tendon analysis module in a baking stage analysis model corresponding to the target baking type to obtain a tobacco leaf analysis result and a tobacco tendon analysis result; and when the tobacco leaf analysis result and the tobacco tendon analysis result reach the preset tobacco leaf color threshold value and the preset tobacco tendon color threshold value of the target baking stage, entering the baking in the next baking stage or finishing the baking.
According to a second aspect of the present disclosure, there is provided a tobacco curing simulation system based on a high temperature curing barn, comprising: the target baking type determining module is used for acquiring the tobacco type of the current tobacco baking test and the target baking stage for baking the target tobacco type as the target baking type; the baking information acquisition module is used for acquiring temperature and humidity information in the last preset time period and the current baking time to acquire a temperature information set, a humidity information set and the baking time; the baking control module is used for constructing and acquiring a corresponding target baking control model according to the target baking type, inputting the baking time, the temperature information set and the humidity information set into the target baking control model, acquiring temperature control parameters and humidity control parameters, and controlling the temperature and the humidity; the sample tobacco leaf image acquisition module is used for sampling and selecting the tobacco leaves which are baked currently according to the preset time period to obtain a plurality of sample tobacco leaves, and acquiring image information of the plurality of sample tobacco leaves to obtain an image information set; the tobacco leaf tobacco tendon analysis result acquisition module is used for preprocessing the image information set and inputting the image information set into a tobacco leaf analysis module and a tobacco tendon analysis module in a baking stage analysis model corresponding to the target baking type to obtain a tobacco leaf analysis result and a tobacco tendon analysis result; and the baking result judging module is used for entering the baking in the next baking stage or finishing the baking when the tobacco leaf analysis result and the tobacco tendon analysis result reach the preset tobacco leaf color threshold value and the preset tobacco tendon color threshold value of the target baking stage.
According to the tobacco leaf baking simulation method based on the high-temperature baking room, the tobacco leaf type and the baking stage are used as target baking types, the corresponding target baking control model is built according to the target baking types, the temperature information set, the humidity information set and the baking time in the last preset time period are further acquired, and are input into the target baking control model to acquire the temperature control parameters and the humidity control parameters, the temperature and the humidity are controlled, and the accuracy of the temperature control parameters and the humidity control parameters can be effectively improved. Further sampling and selecting the tobacco leaves which are currently baked to obtain an image information set, preprocessing the image information set to obtain a tobacco leaf image information set and a tobacco rib image information set, respectively inputting the tobacco leaf image information set and the tobacco rib image information set into a tobacco leaf analysis module and a tobacco rib analysis module in a baking stage analysis model corresponding to the target baking type to obtain a tobacco leaf analysis result and a tobacco rib analysis result, analyzing the tobacco leaf color and the tobacco rib color in the tobacco leaf analysis result and the tobacco rib analysis result, and ending baking or baking in the next stage under the condition that the tobacco leaf color and the tobacco rib color meet the expected requirements to monitor the baking condition of the tobacco leaves and improve the technical effect of the tobacco leaf baking quality.
It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification.
Drawings
For a clearer description of the present disclosure or of the prior art, the drawings that are required to be used in the description of the embodiments or of the prior art will be briefly described, it being apparent that the drawings in the description below are merely illustrative and that other drawings may be obtained, without inventive effort, by a person skilled in the art from the drawings provided.
Fig. 1 is a schematic flow chart of a tobacco leaf baking simulation method based on a high-temperature baking room according to an embodiment of the disclosure;
FIG. 2 is a flow chart of obtaining a target toast type in an embodiment of the present disclosure;
FIG. 3 is a schematic flow chart of obtaining a tobacco leaf analysis result and a tobacco tendon analysis result in an embodiment of the disclosure;
fig. 4 is a schematic structural diagram of a tobacco leaf baking simulation system based on a high-temperature baking room according to an embodiment of the disclosure.
Reference numerals illustrate: the system comprises a target baking type determining module 11, a baking information acquiring module 12, a baking control module 13, a sample tobacco leaf image acquiring module 14, a tobacco leaf tendon analysis result acquiring module 15 and a baking result judging module 16.
Detailed Description
Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and should be considered as merely exemplary. Accordingly, one of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
In order to solve the technical problem that the tobacco leaf baking quality is poor due to insufficient control accuracy of baking temperature and humidity in the existing tobacco leaf baking technology in the prior art, the inventor of the present disclosure obtains the tobacco leaf baking simulation method and system based on the high-temperature baking room through creative labor.
Example 1
Fig. 1 is a diagram of a tobacco leaf baking simulation method based on a high-temperature baking room according to an embodiment of the disclosure, as shown in fig. 1, the method includes:
step S100: acquiring the tobacco leaf type of the current tobacco leaf baking test and a target baking stage for baking the target tobacco leaf type, and taking the tobacco leaf type as the target baking type;
As shown in fig. 2, step S100 of the embodiment of the disclosure further includes:
step S110: obtaining the variety, the position and the maturity of the tobacco leaves which are currently baked as the tobacco leaf types;
step S120: the method comprises the steps of obtaining a baking stage for baking tobacco leaves at present as a target baking stage, wherein the baking stage comprises a yellowing stage, a color fixing stage and a baking stage;
step S130: and combining the tobacco leaf type and the target baking stage to obtain the target baking type.
Specifically, the varieties of tobacco leaves comprise flue-cured tobacco, burley tobacco, aromatic tobacco and other types, the positions of the tobacco leaves refer to the positions of the tobacco leaves on tobacco plants, the tobacco leaves can be divided into five positions of foot leaves, lower two shed leaves, waist leaves, upper two shed leaves and top leaves from bottom to top, and the maturity of the tobacco leaves comprises two types: firstly, under the condition of sufficient nutrition, the tobacco leaves grow to reach the maturity degree, namely, the fresh tobacco leaves in the field are She Chengshou degrees; secondly, harvesting mature tobacco leaves, baking to reach the degree of maturity, and dividing the tobacco leaves into immature, pseudo-mature, primary mature, proper mature, complete mature and overmature, wherein the degree of maturity of the tobacco leaves subjected to baking at present needs to be determined according to actual conditions. Based on the above, the variety, the position and the maturity of the tobacco leaves currently being cured are determined as the tobacco types, and the curing stage currently being cured is further obtained as a target curing stage, wherein the curing stage comprises a yellowing stage, a fixing stage and a curing stage, curing parameters set in different curing stages and curing effects to be achieved are different, for example, the requirements of the tobacco leaves in the yellowing stage for achieving the change are as follows: yellow leaf green ribs of tobacco leaves and green leaf base parts; the tobacco leaf change requirement in the fixed color period: the tobacco leaves reach small curls of yellow slices and yellow ribs; the tobacco leaf change requirement in the drying period: all tobacco leaves in the curing barn are dried in main veins. The tobacco type and the target baking stage which are currently baked are used as target baking types, so that the subsequent temperature and humidity control during baking according to the tobacco type and the baking stage is facilitated.
Step S200: acquiring temperature and humidity information in the last preset time period and the current baking time to obtain a temperature information set, a humidity information set and the baking time;
specifically, the last preset time period refers to a time period when the tobacco leaves to be baked are baked at the last stage, for example, the tobacco leaves to be baked at the current stage need to be baked at a fixed color period, then the last stage is a yellowing period, temperature and humidity information during baking in the yellowing period is acquired, the temperature information refers to the ambient temperature in a baking room, the humidity information refers to the air humidity in the baking room, and particularly, the temperature and humidity can be acquired by arranging a temperature and humidity sensor in the baking room. The baking time refers to the baking stage required to be performed on the tobacco leaves currently baked, namely the time period corresponding to the target baking stage.
Step S300: constructing and acquiring a corresponding target baking control model according to the target baking type, inputting the baking time, the temperature information set and the humidity information set into the target baking control model, acquiring temperature control parameters and humidity control parameters, and controlling the temperature and the humidity;
Wherein, step S300 of the embodiment of the present disclosure further includes:
step S310: according to the baking data of the target baking type in the historical time, a plurality of sample baking times, a plurality of sample temperature information sets and a plurality of sample humidity information sets are obtained;
step S320: respectively formulating and acquiring a plurality of sample temperature control parameters and a plurality of sample humidity control parameters according to the plurality of sample baking times, the plurality of sample temperature information sets and the plurality of sample humidity information sets;
step S330: and constructing the target baking control model by adopting the plurality of sample baking times, the plurality of sample temperature information sets, the plurality of sample humidity information sets, the plurality of sample temperature control parameters and the plurality of sample humidity control parameters as construction data.
Wherein, step S330 of the embodiment of the present disclosure further includes:
step S331: based on BP neural network in machine learning, constructing a network structure of the target baking control model, wherein input data of the target baking control model are baking time, a temperature information set and a sample humidity information set, and output data are temperature control parameters and humidity control parameters;
step S332: performing data annotation on the plurality of sample baking times, the plurality of sample temperature information sets, the plurality of sample humidity information sets, the plurality of sample temperature control parameters and the plurality of sample humidity control parameters to obtain a first training set, a first verification set and a first test set;
Step S333: and performing supervision training on the target baking control model by adopting the first training set, the first verification set and the first test set, performing gradient descent updating on network parameters through errors of actual output and expected output, and performing verification and test to obtain the target baking control model with accuracy meeting preset conditions.
Specifically, the requirements of tobacco leaves in different types and different baking stages on temperature and humidity during baking are different, a corresponding target baking control model is constructed and obtained according to the target baking type, baking time, a temperature information set and a humidity information set are input into the target baking control model, temperature control parameters and humidity control parameters are obtained, the temperature control parameters and the humidity control parameters are the temperature and humidity of the tobacco leaves required to be baked currently, the temperature control parameters refer to the temperature requirements of a baking room, the humidity control parameters refer to the humidity requirements of air in the baking room, and the temperature control parameters and the humidity control parameters include the temperature and the humidity requirements under different times, for example, the temperature of the baking room is firstly increased to 35 ℃, and the temperature of the baking room is increased to 41-42 ℃ at a speed of 1 ℃ every 2 hours. Therefore, the temperature control parameters and the humidity control parameters output by the target baking control model are changed in stages, so that more accurate baking control is facilitated, the baking quality of tobacco leaves is guaranteed, and the temperature and the humidity in a baking room are controlled through the temperature control parameters and the humidity control parameters, so that the tobacco leaves are baked.
Specifically, the historical time is that the baking data of the tobacco leaves with the same target baking type in the historical time is collected for a period of time, for example, one month, the baking data are analyzed and integrated, a plurality of sample baking times, a plurality of sample temperature information sets and a plurality of sample humidity information sets are extracted from the baking data, the sample baking times, the sample temperature information sets and the sample humidity information sets are consistent with the data types indicated by the temperature information sets, the humidity information sets and the baking times, the data in the sample baking times, the sample temperature information sets and the sample humidity information sets are analyzed respectively, corresponding sample temperature control parameters and sample humidity control parameters are configured for the data, and the sample temperature control parameters and the sample humidity control parameters are consistent with the data types indicated by the temperature control parameters and the humidity control parameters and also respectively comprise temperature and humidity requirements under different time. And taking the plurality of sample baking times, the plurality of sample temperature information sets, the plurality of sample humidity information sets, the plurality of sample temperature control parameters and the plurality of sample humidity control parameters as construction data to construct a target baking control model.
Specifically, the process of constructing the target baking control model is as follows: based on BP neural network in machine learning, constructing a network structure of a target baking control model, wherein the target baking control model comprises a plurality of simple units simulating human brain neurons, the target baking control model can form parameters such as weights, thresholds and the like connected among the simple units in the supervision training process, the trained target baking control model can carry out complex nonlinear logic operation according to input data, output data obtained through prediction is output, the input data of the target baking control model is baking time, temperature information set and humidity information set, and the output data is temperature control parameters and humidity control parameters. The method comprises the steps of carrying out data marking on a plurality of sample baking times, a plurality of sample temperature information sets, a plurality of sample humidity information sets, a plurality of sample temperature control parameters and a plurality of sample humidity control parameters, dividing the data marking according to a certain proportion to obtain a first training set, a first verification set and a first test set, wherein the first training set comprises the plurality of sample baking times, the plurality of sample temperature information sets and the plurality of sample humidity information sets, the first verification set comprises a plurality of sample temperature control parameters and a plurality of sample humidity control parameters corresponding to the plurality of sample baking times, the plurality of sample temperature information sets and the plurality of sample humidity information sets in the first training set, and the first test set comprises a plurality of sample baking times, the plurality of sample temperature information sets, the plurality of sample temperature control parameters and the plurality of sample humidity control parameters corresponding to one.
And inputting each group of sample baking time, sample temperature information set and sample humidity information set in the first training set into a target baking control model, and performing supervision adjustment on the output of the target baking control model by utilizing the corresponding sample temperature control parameters and sample humidity control parameters in the first verification set so that the output data of the target baking control model are consistent with the sample temperature control parameters and the sample humidity control parameters. After all data in the first training set are trained, the accuracy test is carried out on the target baking control model by using the first testing set, a plurality of sample baking times, a plurality of sample temperature information sets and a plurality of sample humidity information sets in the first testing set are respectively input into the target baking control model, a plurality of output data are obtained to be used as actual output, a plurality of sample temperature control parameters and a plurality of sample humidity control parameters corresponding to the input data in the first testing set are used as expected output, the error between the actual output and the expected output is calculated, gradient descent update is carried out on the network parameters, in short, the smaller the loss function is, the smaller the error is, the gradient descent update process on the network parameters is the process of minimizing the loss function, namely the error reduction process, the iterative solution can be carried out step by using a gradient descent method, the minimized loss function and the corresponding network parameters are obtained, the updated target baking control model is verified and tested, whether the loss function accords with preset conditions or not is judged, therefore, the accuracy accords with the preset conditions, the target baking control model is obtained, the accuracy accords with the preset conditions, the output quality control model is improved, and the tobacco leaf baking quality control model is improved, and the technical effect is achieved.
Step S400: sampling and selecting the tobacco leaves which are baked currently according to the preset time period to obtain a plurality of sample tobacco leaves, and collecting image information of the plurality of sample tobacco leaves to obtain an image information set;
specifically, the preset time period refers to a baking time period of the tobacco leaves which are currently baked, the tobacco leaves which are currently baked are randomly sampled and selected in the baking time of the tobacco leaves which are currently baked to obtain a plurality of sample tobacco leaves, the image acquisition equipment such as an intelligent camera and a video camera is used for acquiring images of the plurality of sample tobacco leaves, and the image information of the plurality of sample tobacco leaves is acquired to form an image information set.
Step S500: preprocessing the image information set, and inputting the image information set into a tobacco leaf analysis module and a tobacco tendon analysis module in a baking stage analysis model corresponding to the target baking type to obtain a tobacco leaf analysis result and a tobacco tendon analysis result;
as shown in fig. 3, step S500 of the embodiment of the disclosure further includes:
step S510: dividing images in the image information set according to tobacco leaves and tobacco rib areas to obtain a tobacco leaf image information set and a tobacco rib image information set;
Step S520: acquiring a sample tobacco leaf image information set and a sample tobacco rib image information set of the target baking type;
step S530: identifying the image colors in the sample tobacco leaf image information set to obtain a sample tobacco leaf analysis result set, and identifying the image colors in the sample tobacco leaf rib image information set to obtain a sample tobacco leaf rib analysis result set;
step S540: the sample tobacco leaf image information set and the sample tobacco leaf analysis result set are used as construction data to construct the tobacco leaf analysis module;
step S550: the sample tobacco tendon image information set and the sample tobacco tendon analysis result set are used as construction data to construct the tobacco tendon analysis module, and the baking stage analysis model is obtained;
step S560: respectively inputting the tobacco leaf image information set and the tobacco tendon image information set into the tobacco leaf analysis module and the tobacco tendon analysis module to obtain a tobacco leaf analysis result set and a tobacco tendon analysis result set;
step S570: and calculating expected values in the tobacco leaf analysis result set and the tobacco tendon analysis result set to obtain the tobacco leaf analysis result and the tobacco tendon analysis result.
Wherein, step S540 of the embodiment of the present disclosure further includes:
step S541: randomly selecting J groups of data in the sample tobacco leaf image information set and the sample tobacco leaf analysis result set respectively in a put-back way to obtain a first constructed data set, wherein J is a positive integer and is smaller than the number of data in the sample tobacco leaf image information set;
step S542: constructing a first tobacco leaf analysis unit in the tobacco leaf analysis module by adopting the first construction data set;
step S543: continuously randomly selecting J groups of data in the sample tobacco leaf image information set and the sample tobacco leaf analysis result set respectively in a put-back way to obtain a second construction data set, and constructing a second tobacco leaf analysis unit;
step S544: and continuously constructing and obtaining K tobacco leaf analysis units, wherein K is an integer greater than 1, and obtaining the tobacco leaf analysis module.
Wherein, step S542 of the embodiment of the present disclosure further includes:
step S5421: based on a convolutional neural network, constructing a network structure of the first tobacco leaf analysis unit, wherein the first tobacco leaf analysis unit comprises an input layer, a convolutional layer, a pooling layer, a full-connection layer and an output layer;
step S5422: and performing supervision training on the first tobacco leaf analysis unit by adopting the first construction data set, updating network parameters of the first tobacco leaf analysis unit through errors of actual output and expected output until convergence conditions are reached, and performing verification and test, or obtaining the first tobacco leaf analysis unit if accuracy accords with preset conditions.
Specifically, the baking stage analysis model comprises a tobacco leaf analysis module and a tobacco reinforcement analysis module, wherein the tobacco leaf analysis module is used for analyzing a tobacco leaf image to obtain a tobacco leaf analysis result, and the tobacco leaf analysis result comprises the color characteristics of tobacco leaves; the tobacco stem analysis module is used for analyzing the tobacco stem image to obtain a tobacco stem analysis result, wherein the tobacco stem analysis result contains the image characteristics of the tobacco stem. The preprocessing is to analyze an image information set, divide and divide images in the image information set according to tobacco leaves and tobacco tendon areas to obtain the tobacco leaf image information set and the tobacco tendon image information set, and input the tobacco leaf image information set and the tobacco tendon image information set into a tobacco leaf analysis module and a tobacco tendon analysis module respectively to obtain a tobacco leaf analysis result and a tobacco tendon analysis result.
Specifically, firstly, dividing images in an image information set according to distribution areas of tobacco leaves and tobacco ribs to obtain the tobacco leaf image information set and the tobacco rib image information set, if the tobacco leaves and the tobacco rib parts in the images are difficult to divide, respectively selecting and labeling the tobacco leaf parts and the tobacco rib parts, and dividing the frame selected parts to construct the tobacco leaf image information set and the tobacco rib image information set. And further acquiring a sample tobacco leaf image information set and a sample tobacco rib image information set which are the same as the target baking type based on the big data. And identifying the image colors in the sample tobacco leaf image information set, wherein all colors can be specifically represented by RGB (red, green and blue) color values, and when the colors are identified, if common color adjectives such as yellow, pale yellow and the like are directly used as identification results, the analysis results are inaccurate due to the fact that color boundaries are fuzzy, and therefore the RGB color values of the tobacco leaf image are used as sample tobacco leaf analysis results corresponding to the sample tobacco leaf image, so that the sample tobacco leaf analysis result set is obtained. And adopting the same method to perform RGB color value recognition on the image colors in the sample tobacco rod image information set, and forming a sample tobacco rod analysis result set by using the color value recognition result.
Further adopting a sample tobacco image information set and a sample tobacco analysis result set as construction data to construct a tobacco analysis module, wherein the input of the tobacco analysis module is the tobacco image information set, and the output is the tobacco analysis result set; and the sample tobacco rib image information set and the sample tobacco rib analysis result set are used as construction data, the input of the construction tobacco rib analysis module is the tobacco rib image information set, the output is the tobacco rib analysis result, and the tobacco leaf analysis module and the tobacco rib analysis module form a baking stage analysis model together.
The process of constructing the tobacco leaf analysis module is as follows: the method comprises the steps that a sample tobacco image information set and a sample tobacco analysis result set contain multiple groups of sample tobacco image information, J groups of data are randomly selected in the sample tobacco image information set and the sample tobacco analysis result set in a put-back mode respectively, J is a positive integer and smaller than the number of data in the sample tobacco image information set, the randomly selected J groups of data form a first construction data set, a first tobacco analysis unit in a tobacco analysis module is constructed, J groups of data are randomly selected in the sample tobacco image information set and the sample tobacco analysis result set continuously in a put-back mode respectively to obtain a second construction data set, a second tobacco analysis unit is constructed through the second construction data set, K tobacco analysis units form a tobacco analysis module, K is an integer larger than 1, J and K can be set according to practical conditions, for example, the sample tobacco image information set and the sample tobacco analysis result set respectively contain 100 sample tobacco image information and 100 sample tobacco analysis results, and the first sample tobacco analysis result set is 30, and the sample tobacco analysis result 30 is selected from the 100 sample tobacco image information sets for the first time, and the sample tobacco analysis result set has a corresponding relation to the sample tobacco analysis result data set; then, 30 groups of data are selected from 100 sample tobacco leaf image information and 100 sample tobacco leaf analysis results at random for the second time to serve as a second construction data set, and the like, 30 groups of data are selected from 100 sample tobacco leaf image information and 100 sample tobacco leaf analysis results at random for multiple times to serve as a third construction data set and a fourth construction data set, K can be set by oneself until the K construction data set, K construction data sets can be obtained, the data in the K construction data sets are different, K tobacco leaf analysis units are constructed by using the K construction data sets, and the K tobacco leaf analysis units form a tobacco leaf analysis module.
The method is characterized in that the same method as that of constructing a tobacco leaf analysis module is adopted, J groups of data are randomly selected in a sample tobacco tendon image information set and a sample tobacco tendon analysis result set for multiple times in a put-back mode respectively, K tobacco tendon construction data sets are obtained, K tobacco tendon construction data sets are adopted, K tobacco tendon analysis units are constructed, the tobacco tendon analysis module is composed of the K tobacco tendon analysis units, and K is an integer larger than 1.
Inputting the images in the tobacco image information set into K tobacco analysis units in a tobacco analysis module to obtain K tobacco analysis results to form a tobacco analysis result set; inputting the images in the tobacco tendon image information set into K tobacco tendon analysis units in the tobacco tendon analysis module to obtain K tobacco tendon analysis results to form a tobacco tendon analysis result set, calculating expected values, namely average values, in the tobacco leaf analysis result set and the tobacco tendon analysis result set, and taking the average values of the data in the tobacco leaf analysis result set and the tobacco tendon analysis result set as the tobacco leaf analysis result and the tobacco tendon analysis result respectively.
By randomly selecting a small amount of data from the sample data to construct a plurality of tobacco leaf analysis units and a plurality of tobacco reinforcement analysis units, the convergence rate of the tobacco leaf analysis module during training can be effectively reduced, the construction rate is improved, images in a tobacco leaf image information set are input into the plurality of tobacco leaf analysis units and the plurality of tobacco reinforcement analysis units, the output results of the plurality of tobacco leaf analysis units and the plurality of tobacco reinforcement analysis units possibly have differences, average value calculation is performed on the output data of the plurality of tobacco leaf analysis units and the plurality of tobacco reinforcement analysis units respectively, and the average value calculation result is used as a tobacco leaf analysis result and a tobacco reinforcement analysis result, so that the effect of improving the accuracy of the output data of the tobacco leaf analysis module is achieved.
The process of constructing the first tobacco leaf analysis unit is as follows: based on a convolutional neural network, constructing a network structure of a first tobacco leaf analysis unit, wherein the first tobacco leaf analysis unit comprises an input layer, a convolutional layer, a pooling layer, a full-connection layer and an output layer, the first tobacco leaf analysis unit is used for analyzing and identifying a tobacco leaf image, the input layer generally represents a pixel matrix of the tobacco leaf image, a three-dimensional matrix can be used for representing a picture, the length and the width of the three-dimensional matrix represent the size of the image, and the depth of the three-dimensional matrix represents a color channel of the image, namely RGB color values; the core of the convolutional neural network is a convolutional layer, the core part of the convolutional layer is a convolutional operation, and the operation of inner product (element-by-element multiplication and summation) of the tobacco leaf image and the filter matrix (a group of fixed weights: because a plurality of weights of each neuron are fixed and can be regarded as a constant filter) is the convolutional operation; the pooling layer takes the overall statistical characteristics of a certain position adjacent region of a pixel matrix of an input tobacco leaf image as the output of the position, mainly comprises average pooling, maximum pooling and the like, and simply speaking, the pooling is to designate an RGB color value on the region to represent the whole region; after the processing of the multi-pass convolution layer and the pooling layer, the information in the tobacco leaf image can be considered to have been abstracted to specific RGB color values. The convolution layer and the pooling layer can be regarded as the process of extracting the tobacco image characteristics, after the characteristic extraction is finished, the full connection layer is still needed to finish the integration of RGB color values, and finally, the tobacco analysis result, namely the color characteristics (RGB color values) of the tobacco corresponding to the input tobacco image, is output through the output layer.
The method for constructing the target baking control model is similar to the method for constructing the target baking control model, the data in the first construction data set can be divided to obtain a second training set, a second verification set and a second test set, the sample tobacco leaf image information in the second training set is input into the first tobacco leaf analysis unit, the output of the first tobacco leaf analysis unit is supervised and regulated through the sample tobacco leaf analysis result corresponding to the second verification set, the output of the first tobacco leaf analysis unit is consistent with the sample tobacco leaf analysis result corresponding to the second verification set, all the sample tobacco leaf image information in the second training set is trained, the accuracy of the first tobacco leaf analysis unit is tested by utilizing the second test set, the network parameters of the first tobacco leaf analysis unit are updated according to errors of actual output and expected output until the errors of the actual output and the expected output reach the minimum, that is, the accuracy accords with the preset condition, the constructed first tobacco leaf analysis unit is obtained, and the accuracy of the first tobacco leaf analysis unit is ensured. And continuously constructing other K tobacco leaf analysis units and K tobacco tendon analysis units by using the same method as that for constructing the first tobacco leaf analysis unit, so that the constructed K tobacco leaf analysis units form a tobacco leaf analysis module, the K tobacco tendon analysis units form a tobacco tendon analysis module, and the tobacco leaf analysis module and the tobacco tendon analysis module form a baking stage analysis model.
Step S600: and when the tobacco leaf analysis result and the tobacco tendon analysis result reach the preset tobacco leaf color threshold value and the preset tobacco tendon color threshold value of the target baking stage, entering the baking in the next baking stage or finishing the baking.
Specifically, the preset tobacco color threshold and the preset tobacco tendon color threshold are the expected tobacco color and the expected eye color after tobacco baking, and are required to be set according to actual conditions, the preset tobacco color threshold and the preset tobacco tendon color threshold corresponding to different types of tobacco in different baking stages are represented by RGB color values, and the tobacco analysis result and the tobacco tendon analysis result are respectively compared with the preset tobacco color threshold and the preset tobacco tendon color threshold, if the tobacco analysis result and the tobacco tendon analysis result reach the preset tobacco color threshold and the preset tobacco tendon color threshold in the target baking stage, the target baking stage is illustrated to be completed, the next stage baking can be started or ended, and the tobacco baking in the next stage can be continuously performed by using the tobacco baking simulation method based on the high-temperature baking room provided by the embodiment of the disclosure, so that the temperature and humidity control accuracy of tobacco baking is improved, and the technical effect of improving the tobacco baking quality is achieved.
Based on the above analysis, the present disclosure provides a tobacco leaf baking simulation method based on a high-temperature baking room, in this embodiment, a tobacco leaf type and a baking stage are used as target baking types, a corresponding target baking control model is constructed according to the target baking types, a temperature information set, a humidity information set and baking time in a last preset time period are further acquired, and are input into the target baking control model to obtain a temperature control parameter and a humidity control parameter, and temperature and humidity control is performed, so that accuracy of the temperature control parameter and the humidity control parameter can be effectively improved. Further sampling and selecting the tobacco leaves which are currently baked to obtain an image information set, preprocessing the image information set to obtain a tobacco leaf image information set and a tobacco rib image information set, respectively inputting the tobacco leaf image information set and the tobacco rib image information set into a tobacco leaf analysis module and a tobacco rib analysis module in a baking stage analysis model corresponding to the target baking type to obtain a tobacco leaf analysis result and a tobacco rib analysis result, analyzing the tobacco leaf color and the tobacco rib color in the tobacco leaf analysis result and the tobacco rib analysis result, and ending baking or baking in the next stage under the condition that the tobacco leaf color and the tobacco rib color meet the expected requirements to monitor the baking condition of the tobacco leaves and improve the technical effect of the tobacco leaf baking quality.
Example two
Based on the same inventive concept as the tobacco curing simulation method based on the high-temperature curing barn in the previous embodiment, as shown in fig. 4, the present disclosure further provides a tobacco curing simulation system based on the high-temperature curing barn, the system comprising:
the target baking type determining module 11, wherein the target baking type determining module 11 is used for obtaining the tobacco type of the current tobacco baking test and the target baking stage for baking the target tobacco type as the target baking type;
the baking information acquisition module 12 is used for acquiring temperature and humidity information in the last preset time period and the current baking time, and acquiring a temperature information set, a humidity information set and the baking time;
the baking control module 13 is configured to construct and acquire a corresponding target baking control model according to the target baking type, input the baking time, the temperature information set and the humidity information set into the target baking control model, acquire a temperature control parameter and a humidity control parameter, and perform temperature and humidity control;
the sample tobacco leaf image acquisition module 14 is used for sampling and selecting the tobacco leaves which are baked currently according to the preset time period to obtain a plurality of sample tobacco leaves, and acquiring image information of the plurality of sample tobacco leaves to obtain an image information set;
The tobacco leaf tobacco tendon analysis result acquisition module 15 is used for preprocessing the image information set, inputting the image information set into a tobacco leaf analysis module and a tobacco tendon analysis module in a baking stage analysis model corresponding to the target baking type, and acquiring a tobacco leaf analysis result and a tobacco tendon analysis result;
and the baking result judging module 16, wherein the baking result judging module 16 is used for entering the baking in the next baking stage or finishing the baking when the tobacco leaf analysis result and the tobacco tendon analysis result reach the preset tobacco leaf color threshold value and the preset tobacco tendon color threshold value of the target baking stage.
Further, the system further comprises:
the tobacco type acquisition module is used for acquiring the variety, the position and the maturity of the tobacco which is currently baked as the tobacco type;
the tobacco leaf baking device comprises a baking stage acquisition module, a target baking stage acquisition module and a control module, wherein the baking stage acquisition module is used for acquiring a current baking stage for baking tobacco leaves and taking the current baking stage as a target baking stage, and the baking stage comprises a yellowing stage, a color fixing stage and a baking stage;
and the baking type acquisition module is used for combining the tobacco type and the target baking stage to acquire the target baking type.
Further, the system further comprises:
the first sample information acquisition module is used for acquiring a plurality of sample baking times, a plurality of sample temperature information sets and a plurality of sample humidity information sets according to the baking data of the target baking type in the historical time;
the second sample information acquisition module is used for respectively formulating and acquiring a plurality of sample temperature control parameters and a plurality of sample humidity control parameters according to the plurality of sample baking times, the plurality of sample temperature information sets and the plurality of sample humidity information sets;
the target baking control model construction module is used for constructing the target baking control model by adopting the sample baking time, the sample temperature information set, the sample humidity information set, the sample temperature control parameters and the sample humidity control parameters as construction data.
Further, the system further comprises:
the network structure construction module is used for constructing a network structure of the target baking control model based on a BP neural network in machine learning, wherein input data of the target baking control model are baking time, a temperature information set and a sample humidity information set, and output data are temperature control parameters and humidity control parameters;
The data labeling module is used for carrying out data labeling on the plurality of sample baking times, the plurality of sample temperature information sets, the plurality of sample humidity information sets, the plurality of sample temperature control parameters and the plurality of sample humidity control parameters to obtain a first training set, a first verification set and a first test set;
and the supervision and training module is used for performing supervision and training on the target baking control model by adopting the first training set, the first verification set and the first test set, performing gradient descent update on network parameters through errors of actual output and expected output, and performing verification and test to obtain the target baking control model with accuracy meeting preset conditions.
Further, the system further comprises:
the image dividing module is used for dividing the images in the image information set according to tobacco leaves and tobacco rib areas to obtain a tobacco leaf image information set and a tobacco rib image information set;
the sample image acquisition module is used for acquiring a sample tobacco leaf image information set and a sample tobacco rib image information set of the target baking type;
The color recognition module is used for recognizing the image colors in the sample tobacco leaf image information set to obtain a sample tobacco leaf analysis result set, and recognizing the image colors in the sample tobacco leaf image information set to obtain a sample tobacco leaf analysis result set;
the first construction module is used for constructing the tobacco leaf analysis module by adopting the sample tobacco leaf image information set and the sample tobacco leaf analysis result set as construction data;
the second construction module is used for constructing the tobacco rib analysis module by adopting the sample tobacco rib image information set and the sample tobacco rib analysis result set as construction data to obtain the baking stage analysis model;
the analysis result set acquisition module is used for respectively inputting the tobacco leaf image information set and the tobacco tendon image information set into the tobacco leaf analysis module and the tobacco tendon analysis module to obtain a tobacco leaf analysis result set and a tobacco tendon analysis result set;
the expected calculation module is used for calculating expected values in the tobacco leaf analysis result set and the tobacco tendon analysis result set to obtain the tobacco leaf analysis result and the tobacco tendon analysis result.
Further, the system further comprises:
the first construction data set acquisition module is used for randomly selecting J groups of data in the sample tobacco leaf image information set and the sample tobacco leaf analysis result set in a replacement mode respectively to obtain a first construction data set, wherein J is a positive integer and is smaller than the number of data in the sample tobacco leaf image information set;
the first tobacco leaf analysis unit construction module is used for constructing a first tobacco leaf analysis unit in the tobacco leaf analysis module by adopting the first construction data set;
the second tobacco leaf analysis unit construction module is used for continuing to randomly select J groups of data in the sample tobacco leaf image information set and the sample tobacco leaf analysis result set respectively in a replacement mode to obtain a second construction data set and constructing a second tobacco leaf analysis unit;
k tobacco leaf analysis unit acquisition modules are used for continuously constructing and obtaining K tobacco leaf analysis units, the tobacco leaf analysis modules are obtained, and K is an integer greater than 1.
Further, the system further comprises:
The network structure construction module is used for constructing a network structure of the first tobacco leaf analysis unit based on a convolutional neural network, wherein the first tobacco leaf analysis unit comprises an input layer, a convolutional layer, a pooling layer, a full-connection layer and an output layer;
and the second supervision and training module is used for performing supervision and training on the first tobacco leaf analysis unit by adopting the first construction data set, updating the network parameters of the first tobacco leaf analysis unit through errors of actual output and expected output until convergence conditions are reached, verifying and testing, or obtaining the first tobacco leaf analysis unit if the accuracy rate meets preset conditions.
A specific example of a tobacco curing simulation method based on a high-temperature curing barn in the foregoing embodiment is also applicable to a tobacco curing simulation system based on a high-temperature curing barn in the present embodiment, and by the foregoing detailed description of a tobacco curing simulation method based on a high-temperature curing barn, those skilled in the art can clearly know a tobacco curing simulation system based on a high-temperature curing barn in the present embodiment, so that the details will not be described here for brevity of the description. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps recited in the present disclosure may be performed in parallel, sequentially, or in a different order, provided that the desired results of the disclosed aspects are achieved, and are not limited herein.
The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure are intended to be included within the scope of the present disclosure.
Claims (8)
1. A tobacco leaf baking simulation method based on a high-temperature flue-curing barn, which is characterized by comprising the following steps:
acquiring the tobacco leaf type of the current tobacco leaf baking test and a target baking stage for baking the target tobacco leaf type, and taking the tobacco leaf type as the target baking type;
acquiring temperature and humidity information in the last preset time period and the current baking time to obtain a temperature information set, a humidity information set and the baking time;
Constructing and acquiring a corresponding target baking control model according to the target baking type, inputting the baking time, the temperature information set and the humidity information set into the target baking control model, acquiring temperature control parameters and humidity control parameters, and controlling the temperature and the humidity;
sampling and selecting the tobacco leaves which are baked currently according to the preset time period to obtain a plurality of sample tobacco leaves, and collecting image information of the plurality of sample tobacco leaves to obtain an image information set;
preprocessing the image information set, and inputting the image information set into a tobacco leaf analysis module and a tobacco tendon analysis module in a baking stage analysis model corresponding to the target baking type to obtain a tobacco leaf analysis result and a tobacco tendon analysis result; and
and when the tobacco leaf analysis result and the tobacco tendon analysis result reach the preset tobacco leaf color threshold value and the preset tobacco tendon color threshold value of the target baking stage, entering the baking in the next baking stage or finishing the baking.
2. The method according to claim 1, wherein the step of obtaining the type of tobacco currently subjected to the tobacco curing test and the target curing stage of curing the target type of tobacco as the target curing type includes:
Obtaining the variety, the position and the maturity of the tobacco leaves which are currently baked as the tobacco leaf types;
the method comprises the steps of obtaining a baking stage for baking tobacco leaves at present as a target baking stage, wherein the baking stage comprises a yellowing stage, a color fixing stage and a baking stage;
and combining the tobacco leaf type and the target baking stage to obtain the target baking type.
3. The method of claim 1, wherein constructing and acquiring a corresponding target toast control model based on the target toast type comprises:
according to the baking data of the target baking type in the historical time, a plurality of sample baking times, a plurality of sample temperature information sets and a plurality of sample humidity information sets are obtained;
respectively formulating and acquiring a plurality of sample temperature control parameters and a plurality of sample humidity control parameters according to the plurality of sample baking times, the plurality of sample temperature information sets and the plurality of sample humidity information sets;
and constructing the target baking control model by adopting the plurality of sample baking times, the plurality of sample temperature information sets, the plurality of sample humidity information sets, the plurality of sample temperature control parameters and the plurality of sample humidity control parameters as construction data.
4. The method of claim 3, wherein constructing the target bake control model using the plurality of sample bake times, the plurality of sample temperature information sets, the plurality of sample humidity information sets, the plurality of sample temperature control parameters, and the plurality of sample humidity control parameters as construction data comprises:
based on BP neural network in machine learning, constructing a network structure of the target baking control model, wherein input data of the target baking control model are baking time, a temperature information set and a sample humidity information set, and output data are temperature control parameters and humidity control parameters;
performing data annotation on the plurality of sample baking times, the plurality of sample temperature information sets, the plurality of sample humidity information sets, the plurality of sample temperature control parameters and the plurality of sample humidity control parameters to obtain a first training set, a first verification set and a first test set;
and performing supervision training on the target baking control model by adopting the first training set, the first verification set and the first test set, performing gradient descent updating on network parameters through errors of actual output and expected output, and performing verification and test to obtain the target baking control model with accuracy meeting preset conditions.
5. The method of claim 1, wherein preprocessing the image information set and inputting a tobacco leaf analysis module and a tobacco tendon analysis module in a baking stage analysis model corresponding to the target baking type, comprises:
dividing images in the image information set according to tobacco leaves and tobacco rib areas to obtain a tobacco leaf image information set and a tobacco rib image information set;
acquiring a sample tobacco leaf image information set and a sample tobacco rib image information set of the target baking type;
identifying the image colors in the sample tobacco leaf image information set to obtain a sample tobacco leaf analysis result set, and identifying the image colors in the sample tobacco leaf rib image information set to obtain a sample tobacco leaf rib analysis result set;
the sample tobacco leaf image information set and the sample tobacco leaf analysis result set are used as construction data to construct the tobacco leaf analysis module;
the sample tobacco tendon image information set and the sample tobacco tendon analysis result set are used as construction data to construct the tobacco tendon analysis module, and the baking stage analysis model is obtained;
respectively inputting the tobacco leaf image information set and the tobacco tendon image information set into the tobacco leaf analysis module and the tobacco tendon analysis module to obtain a tobacco leaf analysis result set and a tobacco tendon analysis result set;
And calculating expected values in the tobacco leaf analysis result set and the tobacco tendon analysis result set to obtain the tobacco leaf analysis result and the tobacco tendon analysis result.
6. The method of claim 5, wherein constructing the tobacco analysis module using the set of sample tobacco image information and the set of sample tobacco analysis results as construction data comprises:
randomly selecting J groups of data in the sample tobacco leaf image information set and the sample tobacco leaf analysis result set respectively in a put-back way to obtain a first constructed data set, wherein J is a positive integer and is smaller than the number of data in the sample tobacco leaf image information set;
constructing a first tobacco leaf analysis unit in the tobacco leaf analysis module by adopting the first construction data set;
continuously randomly selecting J groups of data in the sample tobacco leaf image information set and the sample tobacco leaf analysis result set respectively in a put-back way to obtain a second construction data set, and constructing a second tobacco leaf analysis unit;
and continuously constructing and obtaining K tobacco leaf analysis units, wherein K is an integer greater than 1, and obtaining the tobacco leaf analysis module.
7. The method of claim 6, wherein constructing a first tobacco leaf analysis unit within the tobacco leaf analysis module using the first construction dataset comprises:
Based on a convolutional neural network, constructing a network structure of the first tobacco leaf analysis unit, wherein the first tobacco leaf analysis unit comprises an input layer, a convolutional layer, a pooling layer, a full-connection layer and an output layer;
and performing supervision training on the first tobacco leaf analysis unit by adopting the first construction data set, updating network parameters of the first tobacco leaf analysis unit through errors of actual output and expected output until convergence conditions are reached, and performing verification and test, or obtaining the first tobacco leaf analysis unit if accuracy accords with preset conditions.
8. A tobacco curing simulation system based on a high temperature curing barn, the system comprising:
the target baking type determining module is used for acquiring the tobacco type of the current tobacco baking test and the target baking stage for baking the target tobacco type as the target baking type;
the baking information acquisition module is used for acquiring temperature and humidity information in the last preset time period and the current baking time to acquire a temperature information set, a humidity information set and the baking time;
The baking control module is used for constructing and acquiring a corresponding target baking control model according to the target baking type, inputting the baking time, the temperature information set and the humidity information set into the target baking control model, acquiring temperature control parameters and humidity control parameters, and controlling the temperature and the humidity;
the sample tobacco leaf image acquisition module is used for sampling and selecting the tobacco leaves which are baked currently according to the preset time period to obtain a plurality of sample tobacco leaves, and acquiring image information of the plurality of sample tobacco leaves to obtain an image information set;
the tobacco leaf tobacco tendon analysis result acquisition module is used for preprocessing the image information set and inputting the image information set into a tobacco leaf analysis module and a tobacco tendon analysis module in a baking stage analysis model corresponding to the target baking type to obtain a tobacco leaf analysis result and a tobacco tendon analysis result; and
and the baking result judging module is used for entering the baking in the next baking stage or finishing the baking when the tobacco leaf analysis result and the tobacco tendon analysis result reach the preset tobacco leaf color threshold value and the preset tobacco tendon color threshold value of the target baking stage.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310429012.5A CN116449713A (en) | 2023-04-20 | 2023-04-20 | Tobacco leaf baking simulation method and system based on high-temperature baking room |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310429012.5A CN116449713A (en) | 2023-04-20 | 2023-04-20 | Tobacco leaf baking simulation method and system based on high-temperature baking room |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116449713A true CN116449713A (en) | 2023-07-18 |
Family
ID=87121713
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310429012.5A Pending CN116449713A (en) | 2023-04-20 | 2023-04-20 | Tobacco leaf baking simulation method and system based on high-temperature baking room |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116449713A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117876380A (en) * | 2024-03-13 | 2024-04-12 | 昆明昊拜农业科技有限公司 | Tobacco leaf environment temperature and humidity and microlayer difference prediction method and system |
-
2023
- 2023-04-20 CN CN202310429012.5A patent/CN116449713A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117876380A (en) * | 2024-03-13 | 2024-04-12 | 昆明昊拜农业科技有限公司 | Tobacco leaf environment temperature and humidity and microlayer difference prediction method and system |
| CN117876380B (en) * | 2024-03-13 | 2024-05-14 | 昆明昊拜农业科技有限公司 | Tobacco leaf environment temperature and humidity and microlayer difference prediction method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104914720B (en) | Electronic nose intelligence baking control system and control method with autolearn feature | |
| CN111274860A (en) | Machine vision-based online automatic tobacco leaf grade sorting identification method | |
| Zhou et al. | Strawberry maturity classification from UAV and near-ground imaging using deep learning | |
| Burgos-Artizzu et al. | Improving weed pressure assessment using digital images from an experience-based reasoning approach | |
| CN116449713A (en) | Tobacco leaf baking simulation method and system based on high-temperature baking room | |
| CN113002816B (en) | Intelligent control method for flue-cured tobacco curing process | |
| CN103919258A (en) | Densification tobacco flue-cure dry-wet bulb temperature automatic control technique based on tobacco image processing | |
| CN114862858A (en) | Cigar harvesting maturity identification method and system based on ensemble learning | |
| CN115620151B (en) | Method and device for identifying phenological period, electronic equipment and storage medium | |
| CN110827273A (en) | Tea disease detection method based on regional convolution neural network | |
| CN109325410B (en) | Electroencephalogram EEG (electroencephalogram) feature extraction method based on convolutional neural network | |
| CN112836623A (en) | Facility tomato farming decision auxiliary method and device | |
| CN114304699A (en) | Method for extracting green and yellow ratio of primary flue-cured tobacco leaves | |
| CN110874835B (en) | Crop leaf disease resistance identification method and system, electronic equipment and storage medium | |
| CN114858984A (en) | Muskmelon germplasm resource commodity data acquisition device and acquisition and analysis method | |
| CN118053154A (en) | Oyster mushroom growth monitoring method and device, electronic equipment and storage medium | |
| CN110766665A (en) | Tongue picture data analysis method based on strong supervision algorithm and deep learning network | |
| KR20230061034A (en) | Method of training machine learning model for estimating plant chlorophyll contents, method of estimating plant growth quantity and plant growth system | |
| CN114511777A (en) | Peach tree flowering phase monitoring and evaluating method based on image recognition method | |
| CN114241221B (en) | Control system based on neural network prediction algorithm | |
| CN116524279A (en) | Artificial intelligent image recognition crop growth condition analysis method for digital agriculture | |
| CN115169751B (en) | Method and system for improving maintenance quality of traditional Chinese medicine decoction pieces | |
| CN113076309B (en) | System and method for predicting water addition amount of raw tobacco | |
| CN113670917B (en) | Analysis method and system for green tea quality | |
| CN117788960B (en) | Tobacco leaf turning judging method, medium and system based on double models |
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 |