Grain mildew prediction method, device and equipment and storage medium

Abstract

The invention relates to the field of intelligent agriculture, and provides a grain mildew prediction method, device and equipment and a storage medium, a data collector is used for carrying out data collection on a simulated granary to obtain multiple groups of parameter data corresponding to the simulated granary, so that detailed parameters corresponding to the simulated granary are obtained; carrying out mildew detection on grains in the simulated granary through detection equipment to obtain a mildew grade corresponding to each group of parameter data so as to cooperate with each group of parameter data to generate training data; performing model training based on multiple groups of parameter data and mildew grades to obtain a multi-parameter mildew prediction model; the real-time data corresponding to the to-be-detected grain is collected through the data collector, the real-time data is input into the mildew prediction model, the prediction result corresponding to the to-be-detected grain is obtained, and therefore the accurate prediction result can be obtained, and the mildew analysis model obtained through the method is high in accuracy and high in practicability. And the grain condition change condition and the activity rule thereof can be accurately analyzed.

Description

Technical Field

[0001] This invention relates to the field of smart agriculture, and in particular to a method, apparatus, equipment, and storage medium for predicting grain mold growth. Background Technology

[0002] Grains are highly susceptible to microbial influences during storage, and the resulting molds and their metabolites are a major cause of grain loss. If it is necessary to analyze grain mold growth, it is necessary to simulate the grain mold growth process to obtain data parameters related to mold growth during grain storage. Based on these data parameters, a corresponding mold growth analysis model can be obtained to predict the grain condition.

[0003] CN109840854A This invention provides a method for predicting grain mold growth, comprising: acquiring storage information of a target grain variety; the storage information including storage temperature, moisture content of the target grain variety, and storage time; performing data preprocessing on the storage information to obtain multidimensional data corresponding to the storage information; inputting the multidimensional data corresponding to the storage information into a pre-established grain mold growth prediction model; and obtaining the mold growth prediction result of the target grain variety after processing by the pre-established grain mold growth prediction model. The grain mold growth prediction method provided by this invention, based on the moisture content, storage temperature, and storage time of the grain variety, can quickly predict the mold growth status of stored grain under the current environment, avoiding complex steps such as bacterial culture and fungal observation, and saving a significant amount of manpower and resources.

[0004] Existing methods for predicting grain mold are difficult to accurately simulate under various parameters such as different geographical locations and meteorological factors, and it is also difficult to obtain accurate data parameters of multiple sets during the mold process. As a result, the mold analysis models obtained through training have poor accuracy and are difficult to accurately analyze changes in grain conditions and their activity patterns. Summary of the Invention

[0005] This invention provides a method, apparatus, equipment, and storage medium for predicting grain mold growth, thereby at least solving the above-mentioned technical problems existing in the prior art.

[0006] According to a first aspect of the present invention, a method for predicting grain mold is provided. The method includes: collecting data from a simulated grain warehouse using a data acquisition device to obtain multiple sets of parameter data corresponding to the simulated grain warehouse; detecting mold in the grain within the simulated grain warehouse using a detection device to obtain a mold level corresponding to each set of parameter data; training a model based on the multiple sets of parameter data and the mold level to obtain a mold prediction model; collecting real-time data corresponding to the grain to be tested using the data acquisition device, inputting the real-time data into the mold prediction model, and obtaining a prediction result corresponding to the grain to be tested.

[0007] In one embodiment, data is collected from a simulated grain warehouse using a data acquisition device to obtain multiple sets of parameter data corresponding to the simulated grain warehouse. This includes: the data acquisition device collects data from the simulated grain warehouse at specified intervals to obtain multiple sets of parameter data collected at multiple specified intervals; wherein each set of parameter data includes at least first data related to the grain warehouse and second data related to weather.

[0008] In one embodiment, the step of detecting mold in grain in a grain silo using a detection device to obtain a mold level corresponding to each set of parameter data includes: within each specified period, performing mold detection on grain in a simulated grain silo using the detection device to obtain a mold level corresponding to each set of parameter data; determining that the grain is at a first mold level when the detection device detects first feature information, the first mold level being used to characterize that the grain is in a safe state, the first feature information including at least metabolic gas characteristics corresponding to rice; and determining that the grain is at a second mold level when the detection device detects second feature information, the second mold level being used to characterize that the grain is in a safe state. The detection device identifies the grain as being in a critical state. The second characteristic information includes at least the volatile gas characteristics corresponding to mold or the first spectral characteristics corresponding to the grain. When the detection device detects the third characteristic information, the grain is determined to be in a third mold level, which is used to characterize the grain as being in a spoiled state. The third characteristic information includes at least the second spectral characteristics corresponding to the grain or the first image characteristics corresponding to the grain. When the detection device detects the fourth characteristic information, the grain is determined to be in a fourth mold level, which is used to characterize the grain as being in a severely spoiled state. The fourth characteristic information includes the second image characteristics corresponding to the grain.

[0009] In one embodiment, the detection device includes at least an electronic nose, a hyperspectral camera, and an RGB camera; the electronic nose is used to detect and obtain the metabolic gas characteristics or volatile gas characteristics; the hyperspectral camera is used to detect and obtain the first spectral feature or the second spectral feature, the first spectral feature being used to characterize the texture feature distribution and reflectance frequency corresponding to Aspergillus glaucus, and the second spectral feature being used to characterize the texture feature distribution and reflectance frequency corresponding to Aspergillus glaucus and Aspergillus niger; the RGB camera is used to detect and obtain the first image feature or the second image feature; the first image feature is used to characterize the dark color of the grain surface or the brown outer shell, and the second image feature is used to characterize the appearance of black, clumping, or lumps on the grain surface.

[0010] In one embodiment, training a mold prediction model based on the multiple sets of parameter data and mold levels includes: generating training data and validation data according to the multiple sets of parameter data and mold levels; inputting the training data into a first model to obtain feature quantities corresponding to the training data, wherein the first model is a convolutional neural network; inputting the feature quantities into a second model to obtain output results corresponding to the training data, wherein the second model is a recurrent neural network, and the output results are used to characterize the mold level in the next specified period; and obtaining a mold prediction model when the output results correspond to the validation data.

[0011] In one embodiment, real-time data corresponding to the grain to be tested is collected by a data acquisition device, and the real-time data is input into the mold prediction model to obtain a prediction result corresponding to the grain to be tested. This includes: the data acquisition device collecting and detecting the grain to be tested to obtain real-time data of the grain to be tested within a specified period; inputting the real-time data into the mold prediction model to obtain an output result corresponding to the grain to be tested, the output result representing the mold level for the next specified period; determining a grain treatment method corresponding to the output result based on the output result; and generating a prediction result corresponding to the grain to be tested based on the output result and the grain treatment method.

[0012] In one embodiment, each set of parameter data includes at least: external temperature of the grain warehouse, external humidity of the grain warehouse, internal temperature of the grain warehouse, internal humidity of the grain warehouse, internal CO2 concentration of the grain warehouse, internal O2 concentration of the grain warehouse, internal temperature of the grain, internal moisture of the grain, temperature of the location of the grain warehouse, humidity of the location of the grain warehouse, wind speed of the location of the grain warehouse, and wind direction of the location of the grain warehouse.

[0013] According to a second aspect of the present invention, a grain mold prediction device is provided. The device includes: an acquisition module for acquiring data from a simulated grain warehouse using a data acquisition device to obtain multiple sets of parameter data corresponding to the simulated grain warehouse; a detection module for detecting mold in grain within the simulated grain warehouse using a detection device to obtain a mold level corresponding to each set of parameter data; a training module for training a model based on the multiple sets of parameter data and the mold level to obtain a mold prediction model; and a prediction module for acquiring real-time data corresponding to the grain to be tested using a data acquisition device, inputting the real-time data into the mold prediction model, and obtaining a prediction result corresponding to the grain to be tested.

[0014] In one embodiment, the acquisition module is further configured to have the data collector collect data from the simulated grain warehouse at a specified period to obtain multiple sets of parameter data collected at multiple specified periods; wherein each set of parameter data includes at least first data related to the grain warehouse and second data related to the weather.

[0015] In one embodiment, the detection module is further configured to perform mold detection on the grain in the simulated grain warehouse using the detection device within each specified period to obtain a mold level corresponding to each set of parameter data; when the detection device detects a first feature information, the grain is determined to be at a first mold level, the first mold level being used to characterize that the grain is in a safe state, and the first feature information includes at least a metabolic gas feature corresponding to rice; when the detection device detects a second feature information, the grain is determined to be at a second mold level, the second mold level being used to characterize that the grain is in a critical state, and the second feature information includes at least a volatile gas feature corresponding to mold, or a first spectral feature corresponding to the grain; when the detection device detects a third feature information, the grain is determined to be at a third mold level, the third mold level being used to characterize that the grain is in a deteriorated state, and the third feature information includes at least a second spectral feature corresponding to the grain, or a first image feature corresponding to the grain; when the detection device detects a fourth feature information, the grain is determined to be at a fourth mold level, the fourth mold level being used to characterize that the grain is in a severely deteriorated state, and the fourth feature information includes a second image feature corresponding to the grain.

[0016] In one embodiment, the detection device includes at least an electronic nose, a hyperspectral camera, and an RGB camera; the electronic nose is used to detect and obtain the metabolic gas characteristics or volatile gas characteristics; the hyperspectral camera is used to detect and obtain the first spectral feature or the second spectral feature, the first spectral feature being used to characterize the texture feature distribution and reflectance frequency corresponding to Aspergillus glaucus, and the second spectral feature being used to characterize the texture feature distribution and reflectance frequency corresponding to Aspergillus glaucus and Aspergillus niger; the RGB camera is used to detect and obtain the first image feature or the second image feature; the first image feature is used to characterize the dark color of the grain surface or the brown outer shell, and the second image feature is used to characterize the appearance of black, clumping, or lumps on the grain surface.

[0017] In one embodiment, the training module is further configured to generate training data and validation data based on the multiple sets of parameter data and mold levels; input the training data into a first model to obtain feature quantities corresponding to the training data, wherein the first model is a convolutional neural network; input the feature quantities into a second model to obtain output results corresponding to the training data, wherein the second model is a recurrent neural network, and the output results are used to characterize the mold level in the next specified period; and obtain a mold prediction model when the output results correspond to the validation data.

[0018] In one embodiment, the prediction module is further configured to: collect data from the grain to be tested using a data collector to obtain real-time data of the grain within a specified period; input the real-time data into the mold prediction model to obtain an output result corresponding to the grain to be tested, wherein the output result characterizes the mold level for the next specified period; determine the grain treatment method corresponding to the output result based on the output result; and generate a prediction result corresponding to the grain to be tested based on the output result and the grain treatment method.

[0019] In one embodiment, each set of parameter data includes at least: external temperature of the grain warehouse, external humidity of the grain warehouse, internal temperature of the grain warehouse, internal humidity of the grain warehouse, internal CO2 concentration of the grain warehouse, internal O2 concentration of the grain warehouse, internal temperature of the grain, internal moisture of the grain, temperature of the location of the grain warehouse, humidity of the location of the grain warehouse, wind speed of the location of the grain warehouse, and wind direction of the location of the grain warehouse.

[0020] According to a third aspect of the present invention, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method of the present invention.

[0021] According to a fourth aspect of the present invention, a non-transitory computer-readable storage medium is provided storing computer instructions for causing the computer to perform the method described in the present invention.

[0022] The present invention discloses a method, apparatus, equipment, and storage medium for predicting grain mold. It collects data from a simulated grain silo using a data acquisition device to obtain multiple sets of parameter data corresponding to the simulated silo, thus obtaining detailed parameters corresponding to the simulated silo. Then, it uses a detection device to detect mold in the grain within the simulated silo, obtaining the mold level corresponding to each set of parameter data, which is used to generate training data for each set of parameter data. Next, it trains a model based on multiple sets of parameter data and mold levels to obtain a multi-parameter mold prediction model. Finally, it collects real-time data corresponding to the grain to be tested using a data acquisition device and inputs this real-time data into the mold prediction model to obtain prediction results corresponding to the grain to be tested, thereby achieving accurate prediction results. The mold analysis model obtained through this application has high accuracy and can accurately analyze changes in grain conditions and their activity patterns.

[0023] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0024] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description taken in conjunction with the accompanying drawings. Several embodiments of the invention are illustrated in the drawings by way of example and not limitation, wherein:

[0025] In the accompanying drawings, the same or corresponding reference numerals indicate the same or corresponding parts.

[0026] Figure 1 A schematic diagram illustrating the implementation process of the grain mold prediction method according to an embodiment of the present invention is shown;

[0027] Figure 2 A schematic diagram of a grain mold prediction device according to an embodiment of the present invention is shown;

[0028] Figure 3 A schematic diagram of the composition structure of an electronic device according to an embodiment of the present invention is shown. Detailed Implementation

[0029] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0030] Figure 1 This diagram illustrates the implementation flow of the grain mold prediction method according to an embodiment of the present invention; please refer to... Figure 1 ;

[0031] According to a first aspect of the present invention, a method for predicting grain mold is provided, the method comprising: step 101, acquiring data from a simulated grain warehouse using a data acquisition device to obtain multiple sets of parameter data corresponding to the simulated grain warehouse; step 102, detecting mold in the grain within the simulated grain warehouse using a detection device to obtain a mold level corresponding to each set of parameter data; step 103, training a model based on the multiple sets of parameter data and the mold level to obtain a mold prediction model; and step 104, acquiring real-time data corresponding to the grain to be tested using a data acquisition device, inputting the real-time data into the mold prediction model, and obtaining a prediction result corresponding to the grain to be tested.

[0032] The present invention discloses a method, apparatus, equipment, and storage medium for predicting grain mold. It collects data from a simulated grain silo using a data acquisition device to obtain multiple sets of parameter data corresponding to the simulated silo, thus obtaining detailed parameters corresponding to the simulated silo. Then, it uses a detection device to detect mold in the grain within the simulated silo, obtaining the mold level corresponding to each set of parameter data, which is used to generate training data for each set of parameter data. Next, it trains a model based on multiple sets of parameter data and mold levels to obtain a multi-parameter mold prediction model. Finally, it collects real-time data corresponding to the grain to be tested using a data acquisition device and inputs this real-time data into the mold prediction model to obtain prediction results corresponding to the grain to be tested, thereby achieving accurate prediction results. The mold analysis model obtained through this application has high accuracy and can accurately analyze changes in grain conditions and their activity patterns.

[0033] In step 101 of the present invention, data is collected from the simulated grain warehouse using a data acquisition device to obtain multiple sets of parameter data corresponding to the simulated grain warehouse.

[0034] Specifically, the data acquisition unit includes various sensors, each used to collect different parameter data. Specifically, there are 12 sensors, including sensors for external grain silo temperature, external grain silo humidity, internal grain silo temperature, internal grain silo humidity, internal grain silo CO2 concentration, internal grain silo O2 concentration, internal grain temperature, internal grain moisture, grain silo location temperature, grain silo location humidity, grain silo location wind speed, and grain silo location wind direction. By collecting multi-dimensional data, a mold prediction model capable of multi-dimensional grain condition assessment can be established, allowing for deeper feature mining through parameter data to calculate more accurate prediction results. The data acquisition unit collects data from the simulated grain silo at specified intervals. Specifically, this can refer to collecting data at specified intervals during the grain simulation process, depending on the different grades of grain; or it can refer to collecting data at specified intervals throughout the entire grain simulation process. The latter approach is preferred. During data processing, a sample within a specific time unit is extracted as a data unit, and the datasets from the 12 sensors within this time unit are used as a set of parameter data. Data from different time units are acquired to obtain multiple sets of parameter data, thereby fusing one-dimensional parameters with time to obtain two-dimensional data, facilitating the mining of deeper features. The time unit can be determined based on the accuracy of the training results; specifically, the time unit can include the time length corresponding to 50-60 data points. Each set of parameter data includes at least first data related to the grain warehouse and second data related to meteorology. The first data corresponds to the aforementioned external temperature, external humidity, internal temperature, internal humidity, internal CO2 concentration, internal O2 concentration, internal grain temperature, and internal grain moisture. The second data corresponds to the aforementioned location temperature, location humidity, location wind speed, and location wind direction of the grain warehouse. By collecting the first data related to the grain warehouse, the influence of grain warehouse-related parameters on grain mold can be analyzed; by collecting the second data related to meteorology, the influence of environmental changes on grain mold can be analyzed. By processing the first and second data together, it is possible to collect most of the influencing factors related to grain mold, thereby enabling the trained model to have higher accuracy.

[0035] In step 102 of the present invention, the grain in the grain warehouse is tested for mold based on the detection equipment to obtain the mold level corresponding to each set of parameter data;

[0036] Specifically, within each specified period, while the data acquisition device collects data, the detection device performs mold detection on the grain in the simulated grain warehouse to obtain the mold level corresponding to each set of parameter data collected by the data acquisition device, and the label information corresponding to the training data.

[0037] Specifically, the testing for mold levels includes:

[0038] When the detection equipment detects the first characteristic information, the grain is determined to be at the first mold level. The first mold level indicates that the grain is in a safe state. The first characteristic information includes at least the metabolic gas characteristics corresponding to rice. Specifically, the first characteristic information indicates that there is essentially no harm to fungal growth, meaning the grain is at the first mold level and in a safe state, meaning it is safe to eat. The metabolic gas characteristics corresponding to rice specifically refer to the fact that only intermediate or final products of the rice's own metabolic process are collected inside the grain silo, such as some alcohols, aldehydes, and ketones. Specifically, these can include alkanes, benzene ring alcohols, aldehydes, and ketones.

[0039] When the detection equipment detects the second characteristic information, the grain is determined to be at the second mold level. The second mold level indicates that the grain is in a critical state. The second characteristic information includes at least the volatile gas characteristics corresponding to the mold, or the first spectral characteristics corresponding to the grain. Specifically, the second characteristic information indicates the presence of Aspergillus gray-green and a small amount of Aspergillus white in the grain, meaning the grain is at the second mold level, or in a critical state. The critical state refers to the early, slight mold stage, at which professional treatment can prevent most grains from further mold growth. The volatile gas characteristics corresponding to the mold refer to volatile components produced during mold metabolism. The first spectral characteristics corresponding to the grain refer to the texture feature distribution and reflectance frequency of the hyperspectral image of the grain corresponding to the critical state.

[0040] When the detection equipment detects the third characteristic information, the grain is determined to be in the third mold level. The third mold level is used to characterize that the grain is in a deteriorated state. The third characteristic information includes at least the second spectral feature corresponding to the grain, or the first image feature corresponding to the grain. The third characteristic information is used to characterize that there is obvious mold growth in the grain, mainly Aspergillus glaucus and Aspergillus oryzae, with a small amount of other fungi growing. At this time, the grain is in the third mold level, that is, the grain has obviously deteriorated. Obviously deteriorated means that the mold has endangered the safety of the grain. Specifically, the growth dominance of Aspergillus glaucus in the grain is gradually replaced by Aspergillus oryzae, and a small amount of other fungi grow. The second spectral feature corresponding to the grain refers to the texture feature distribution and reflectance frequency of the hyperspectral image of the grain to indicate the mold growth. The first image feature corresponding to the grain refers to the mold surface image obtained by an RGB camera. Because at this mold level, the mold-related features can be directly obtained by the camera. Specifically, the darkening of the grain color or the browning of the outer shell indicates that the grain is in the third mold level.

[0041] When the detection equipment detects the fourth feature information, it determines that the grain is in the fourth mold level. The fourth mold level is used to characterize that the grain is in a severely deteriorated state. The fourth feature information includes the second image feature corresponding to the grain.

[0042] The fourth feature is used to characterize the presence of a large amount of mold in the grain, with Aspergillus white as the main mold, and later, Aspergillus flavus, Penicillium, and other harmful fungi may appear. At this point, the grain is in the fourth mold level, which means that the grain has obviously deteriorated. Obviously deteriorated means that the mold has seriously endangered the safety of the grain. The second image feature corresponding to the grain refers to the mold surface image obtained by an RGB camera. Because at this mold level, the mold-related features can be directly obtained by the camera. Specifically, the grain is black and has clumping and lumps, which indicates that the grain is in the third mold level.

[0043] In summary, the detection equipment includes at least an electronic nose, a hyperspectral camera, and an RGB camera. The electronic nose is used to detect and obtain metabolic gas characteristics or volatile gas characteristics. The hyperspectral camera is used to detect and obtain a first spectral feature or a second spectral feature. The first spectral feature is used to characterize the distribution of texture features and reflectance frequency corresponding to Aspergillus glaucus, and the second spectral feature is used to characterize the distribution of texture features and reflectance frequency corresponding to Aspergillus glaucus and Aspergillus niger. The RGB camera is used to detect and obtain a first image feature or a second image feature. The first image feature is used to characterize the dark color of the grain surface or the brown color of the outer shell, and the second image feature is used to characterize the appearance of black spots, clumps, or lumps on the grain surface.

[0044] Specifically, as the simulated mold growth time changes, the rice begins to show slight discoloration, off-taste, and dampness, with a small amount of heat. Then, mold spots or clumps begin to appear on the germ or damaged parts of the grain. Subsequently, the heat and moisture accumulate further, the degree of mold growth deepens, until the grains significantly heat up and a large area of ​​mold colonies cover the surface, the grains deform, clump together, etc., accompanied by a strong musty smell, sour smell and odor.

[0045] During rice storage, the odor components mainly come from two sources: firstly, intermediate or final products of the rice's own metabolic processes, such as alcohols, aldehydes, and ketones; and secondly, volatile components produced by microorganisms in the rice, primarily molds, during their metabolism. An electronic nose was used to collect and determine the odor components of rice from its normal state to the point of mold growth.

[0046] During rice storage, mold flora, primarily of the genera *Aspergillus* and *Penicillium*, are randomly distributed on the grain surface when mold develops, exhibiting differential distributions in their near-infrared spectral reflectance values. By acquiring near-infrared spectral imaging data of different mold states and combining the differences in images resulting from varying degrees of mold, the texture features of the NIR images of moldy rice can be calculated. This allows for the analysis of NIR image characteristics of moldy rice, such as various textures and frequency domain properties, providing data indicators for detecting the degree of mold during rice storage and transportation.

[0047] For stages with obvious mold growth, sensory quality assessment of moldy grains can be used: Detection of moldy grains is a sensory testing method and an internationally accepted method for evaluating grain quality. This method is fast, simple to operate, and widely used. The definition of moldy grains is clearly stated in the national standard GB 2715-2016: "Grains with obvious mold growth on the surface, damaging the embryo, endosperm, or cotyledons, and having no edible value are moldy grains." The definition of suitable storage conditions is based on color and odor, as defined in GB / T 20569-2006, the rules for judging the storage quality of rice. Color is described as normal, basically normal, or distinctly yellow, dark gray, brown, or other abnormal colors unacceptable to humans. Therefore, RGB cameras are used to photograph moldy grains to obtain the severity of the mold. Early-stage micro-moldiness in rice can be detected using electronic nose technology to collect chemical gases generated during the mold growth process. These gases mainly include: natural mold odor substances in rice such as alkanes, benzene ring alcohols, aldehydes, ketones, and other volatile gases. It is possible to quantify the stage from safety to criticality. Mid-stage mold growth in rice can be quantified from the critical to the harmful stage by analyzing the distribution of texture features and the frequency of reflectance values ​​in near-infrared spectral images.

[0048] Late-stage mold growth in rice can be detected promptly through model comparison or by expert experience based on subtle changes in grain color, temperature, and moisture. The severity and extent of mold damage are quantified by combining recorded images of the mold growth process with expert experience. Accurate classification of mold levels is achieved based on gas, near-infrared spectral, and RGB image data collected by an electronic nose during the mold growth process, combined with human sensory analysis methods.

[0049] In step 103 of the present invention, a model is trained based on multiple sets of parameter data and mold level to obtain a mold prediction model.

[0050] The specific steps may include: generating training data and validation data based on multiple sets of parameter data and mold levels; inputting the training data into a first model to obtain feature quantities corresponding to the training data, the first model being a convolutional neural network; inputting the feature quantities into a second model to obtain output results corresponding to the training data, the second model being a recurrent neural network, the output results being used to characterize the mold level in the next specified period; obtaining a mold prediction model when the output results correspond to the validation data; wherein, by first passing the training data through the first model, its deep-seated feature quantities can be mined, and then input into the second model, based on its feature quantities and temporal sequence, a prediction result for the grain is obtained, and the output prediction result is validated through validation data, thereby training the model.

[0051] Specifically, the training method is as follows:

[0052] 1. Preprocess the raw data first, and normalize the data using the standard deviation method.

[0053] 2. Input the data into the CNN model, perform convolution operations using multiple convolution kernels, and then perform pooling. The convolutional layers extract features from the original data and deeply explore the inherent relationships within the data, while the pooling layers reduce network complexity and the number of training parameters.

[0054] 3. A Dropout layer was added after the CNN layer, which randomly blocks a number of neurons each time to prevent the model from relying too much on certain features and improve the robustness of the model.

[0055] 4. The newly extracted features are passed as input to the GRU recurrent neural network, and the temporal patterns in the training data and the Adam optimization algorithm are used to update and iterate the parameters.

[0056] 5. The GRU layer takes the hidden state at the last moment of the time series as its output.

[0057] 6. The final output value is obtained by weighted summation through fully connected layers.

[0058] 7. The output value is denormalized to obtain the predicted mold level for the next time step.

[0059] The CNN combined with GRU model proposed in this invention has better prediction performance than traditional models, improves the training accuracy of the model, and has a high degree of fitting in the training results, thus achieving efficient, accurate and reliable prediction of grain mold grading.

[0060] In step 104 of this invention, real-time data corresponding to the grain to be tested is collected by a data acquisition device, and the real-time data is input into the mold prediction model to obtain the prediction result corresponding to the grain to be tested.

[0061] Specifically, the data acquisition device collects and tests the grain to be tested, obtaining real-time data on the grain within a specified period; the real-time data is input into the mold prediction model to obtain the output result corresponding to the grain, which represents the mold level for the next specified period; based on the output result, the corresponding grain treatment method is determined; based on the output result and the grain treatment method, the prediction result corresponding to the grain is generated, thus providing a warning for grain storage based on the output result.

[0062] Figure 2 A schematic diagram of a grain mold prediction device according to an embodiment of the present invention is shown; please refer to it. Figure 2 ;

[0063] According to a second aspect of the present invention, a grain mold prediction device is provided, comprising: an acquisition module 201, used to acquire data from a simulated grain warehouse based on a data acquisition device to acquire multiple sets of parameter data corresponding to the simulated grain warehouse; a detection module 202, used to detect mold in the grain in the simulated grain warehouse based on a detection device to acquire a mold level corresponding to each set of parameter data; a training module 203, used to train a model based on multiple sets of parameter data and mold levels to acquire a mold prediction model; and a prediction module 204, used to acquire real-time data corresponding to the grain to be tested through a data acquisition device, input the real-time data into the mold prediction model, and obtain a prediction result corresponding to the grain to be tested.

[0064] In one embodiment, the acquisition module is further configured to have the data collector collect data from the simulated grain warehouse at specified intervals to obtain multiple sets of parameter data collected at multiple specified intervals; wherein each set of parameter data includes at least first data related to the grain warehouse and second data related to the weather.

[0065] In one embodiment, the detection module is further configured to perform mold detection on the grain in the simulated grain warehouse using a detection device within each specified period to obtain a mold level corresponding to each set of parameter data; when the detection device detects a first feature information, the grain is determined to be at a first mold level, the first mold level being used to characterize that the grain is in a safe state, and the first feature information includes at least the metabolic gas feature corresponding to rice; when the detection device detects a second feature information, the grain is determined to be at a second mold level, the second mold level being used to characterize that the grain is in a critical state, and the second feature information includes at least the volatile gas feature corresponding to mold, or a first spectral feature corresponding to the grain; when the detection device detects a third feature information, the grain is determined to be at a third mold level, the third mold level being used to characterize that the grain is in a deteriorated state, and the third feature information includes at least the second spectral feature corresponding to the grain, or a first image feature corresponding to the grain; when the detection device detects a fourth feature information, the grain is determined to be at a fourth mold level, the fourth mold level being used to characterize that the grain is in a severely deteriorated state, and the fourth feature information includes the second image feature corresponding to the grain.

[0066] In one embodiment, the detection device includes at least an electronic nose, a hyperspectral camera, and an RGB camera; the electronic nose is used to detect and obtain metabolic gas characteristics or volatile gas characteristics; the hyperspectral camera is used to detect and obtain a first spectral feature or a second spectral feature, the first spectral feature being used to characterize the texture feature distribution and reflectance frequency corresponding to Aspergillus glaucus, and the second spectral feature being used to characterize the texture feature distribution and reflectance frequency corresponding to Aspergillus glaucus and Aspergillus niger; the RGB camera is used to detect and obtain a first image feature or a second image feature; the first image feature is used to characterize the dark color of the grain surface or the brown outer shell, and the second image feature is used to characterize the appearance of black, clumping, or lumps on the grain surface.

[0067] In one embodiment, the training module is further configured to generate training data and validation data based on multiple sets of parameter data and mold levels; input the training data into a first model to obtain feature quantities corresponding to the training data, wherein the first model is a convolutional neural network; input the feature quantities into a second model to obtain output results corresponding to the training data, wherein the second model is a recurrent neural network, and the output results are used to characterize the mold level in the next specified period; when the output results correspond to the validation data, a mold prediction model is obtained.

[0068] In one embodiment, the prediction module is further configured to: collect data from the data collector to detect the grain to be tested and obtain real-time data of the grain to be tested within a specified period; input the real-time data into the mold prediction model to obtain the output result corresponding to the grain to be tested, wherein the output result represents the mold level for the next specified period; determine the grain treatment method corresponding to the output result based on the output result; and generate a prediction result corresponding to the grain to be tested based on the output result and the grain treatment method.

[0069] In one embodiment, each set of parameter data includes at least: external temperature of the grain silo, external humidity of the grain silo, internal temperature of the grain silo, internal humidity of the grain silo, internal CO2 concentration of the grain silo, internal O2 concentration of the grain silo, internal temperature of the grain, internal moisture of the grain, temperature of the location of the grain silo, humidity of the location of the grain silo, wind speed of the location of the grain silo, and wind direction of the location of the grain silo.

[0070] According to embodiments of the present invention, the present invention also provides an electronic device and a readable storage medium.

[0071] Figure 3 A schematic block diagram of an example electronic device 300 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0072] like Figure 3 As shown, device 300 includes a computing unit 301, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 302 or a computer program loaded from storage unit 303 into random access memory (RAM) 303. The RAM 303 may also store various programs and data required for the operation of device 300. The computing unit 301, ROM 302, and RAM 303 are interconnected via bus 304. Input / output (I / O) interface 305 is also connected to bus 304.

[0073] Multiple components in device 300 are connected to I / O interface 305, including: input unit 306, such as keyboard, mouse, etc.; output unit 307, such as various types of monitors, speakers, etc.; storage unit 303, such as disk, optical disk, etc.; and communication unit 309, such as network card, modem, wireless transceiver, etc. Communication unit 309 allows device 300 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0074] The computing unit 301 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 301 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 301 performs the various methods and processes described above, such as the grain mold prediction method. For example, in some embodiments, the grain mold prediction method may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 303. In some embodiments, part or all of the computer program may be loaded and / or installed on device 300 via ROM 302 and / or communication unit 309. When the computer program is loaded into RAM 303 and executed by the computing unit 301, one or more steps of the grain mold prediction method described above may be performed. Alternatively, in other embodiments, the computing unit 301 may be configured to perform the grain mold prediction method by any other suitable means (e.g., by means of firmware).

[0075] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0076] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0077] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0078] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0079] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0080] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0081] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0082] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0083] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for predicting grain mold growth, characterized in that, The method includes: Data is collected from the simulated grain warehouse using a data acquisition device to obtain multiple sets of parameter data corresponding to the simulated grain warehouse. Based on the detection equipment, mold detection was performed on the grain in the simulated grain warehouse to obtain the mold level corresponding to each set of parameter data; Based on the aforementioned multiple sets of parameter data and mold levels, a mold prediction model is trained to obtain the mold prediction model. Real-time data corresponding to the grain to be tested is collected by a data acquisition device, and the real-time data is input into the mold prediction model to obtain the prediction result corresponding to the grain to be tested.

2. The method according to claim 1, characterized in that, Data is collected from the simulated grain silo using a data acquisition device to obtain multiple sets of parameter data corresponding to the simulated grain silo, including: The data acquisition device collects data from the simulated grain warehouse at a specified period to obtain multiple sets of parameter data collected at multiple specified periods. Each set of parameter data includes at least the first set of data related to the grain depot and the second set of data related to the weather.

3. The method according to claim 2, characterized in that, The method involves using detection equipment to detect mold in grain in a grain warehouse and obtaining the mold level corresponding to each set of parameter data, including: Within each specified period, the detection equipment is used to detect mold in the simulated grain warehouse to obtain the mold level corresponding to each set of parameter data; When the detection device detects the first feature information, it determines that the grain is in the first mold level. The first mold level is used to characterize that the grain is in a safe state. The first feature information includes at least the metabolic gas characteristics corresponding to rice. When the detection device detects the second feature information, it determines that the grain is in the second mold level. The second mold level is used to characterize that the grain is in a critical state. The second feature information includes at least the volatile gas characteristics corresponding to mold, or the first spectral characteristics corresponding to grain. When the detection device detects the third feature information, it determines that the grain is in the third mold level. The third mold level is used to characterize that the grain is in a deteriorated state. The third feature information includes at least a second spectral feature corresponding to the grain or a first image feature corresponding to the grain. When the detection device detects the fourth feature information, it determines that the grain is in the fourth mold level. The fourth mold level is used to characterize that the grain is in a severely deteriorated state. The fourth feature information includes the second image feature corresponding to the grain.

4. The method according to claim 3, characterized in that, The detection equipment includes at least an electronic nose, a hyperspectral camera, and an RGB camera; The electronic nose is used to detect and obtain the metabolic gas characteristics or volatile gas characteristics; The hyperspectral camera is used to detect and obtain the first spectral feature or the second spectral feature. The first spectral feature is used to characterize the texture feature distribution and reflectance frequency corresponding to Aspergillus glaucus. The second spectral feature is used to characterize the texture feature distribution and reflectance frequency corresponding to Aspergillus glaucus and Aspergillus niger. The RGB camera is used to detect and obtain the first image feature or the second image feature; the first image feature is used to characterize the dark color of the grain surface or the brown outer shell, and the second image feature is used to characterize the appearance of black, clumping or lumps on the grain surface.

5. The method according to claim 2, characterized in that, Based on the aforementioned multiple sets of parameter data and mold severity levels, a mold prediction model is obtained through training, including: Based on the multiple sets of parameter data and mold levels, training data and validation data are generated; The training data is input into the first model to obtain the feature quantity corresponding to the training data. The first model is a convolutional neural network. The feature quantity is input into the second model to obtain the output result corresponding to the training data. The second model is a recurrent neural network, and the output result is used to characterize the mold level in the next specified cycle. When the output corresponds to the verification data, a mold prediction model is obtained.

6. The method according to claim 1, characterized in that, Real-time data corresponding to the grain to be tested is collected by a data acquisition device, and the real-time data is input into the mold prediction model to obtain the prediction result corresponding to the grain to be tested, including: The data acquisition device collects and detects the grain to be tested, and obtains real-time data of the grain to be tested within a specified period. The real-time data is input into the mold prediction model to obtain the output result corresponding to the grain to be tested. The output result represents the mold level for the next specified period. Based on the output results, determine the grain processing method corresponding to the output results; Based on the output results and the grain processing method, a prediction result corresponding to the grain to be tested is generated.

7. The method according to claims 1-6, characterized in that, Each set of parameter data includes at least: external temperature of the grain warehouse, external humidity of the grain warehouse, internal temperature of the grain warehouse, internal humidity of the grain warehouse, internal CO2 concentration of the grain warehouse, internal O2 concentration of the grain warehouse, internal temperature of the grain, internal moisture of the grain, temperature of the location of the grain warehouse, humidity of the location of the grain warehouse, wind speed of the location of the grain warehouse, and wind direction of the location of the grain warehouse.

8. A grain mold prediction device, characterized in that, The device includes: The acquisition module is used to collect data from the simulated grain warehouse based on the data acquisition device, and obtain multiple sets of parameter data corresponding to the simulated grain warehouse; The detection module is used to detect mold in the grain in the simulated grain warehouse based on the detection equipment, and obtain the mold level corresponding to each set of parameter data; The training module is used to train the model based on the multiple sets of parameter data and mold level to obtain a mold prediction model; The prediction module is used to collect real-time data corresponding to the grain to be tested through a data collector, input the real-time data into the mold prediction model, and obtain the prediction result corresponding to the grain to be tested.

9. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-8.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-8.

Images