Multi-zone adaptive gas circulation ai physical preservation technology control system

Abstract

The present application relates to the technical field of preservation control, in particular to an AI physical preservation technology control system with multi-zone adaptive gas circulation, comprising a first data acquisition module, which acquires the storage item parameters of the i-th independent airtight zone of the overall storage area, the storage item parameters including the item category and the target preservation duration; a first strategy generation module, which inputs the acquired storage item parameters into a pre-constructed first item preservation strategy model and outputs a first item preservation strategy; a second data acquisition module, which acquires the ripeness characteristic parameters of the storage items; and a second strategy generation module, which outputs a second item preservation strategy. The control system can not only maintain the stability of the storage environment, but also accurately adjust the gas component ratio according to the real-time state of the items, ultimately realizing the accurate matching of the environmental parameters and the physiological needs of the items and significantly improving the preservation adaptive ability.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of preservation control, in particular to an AI physical preservation technology control system with multi-region adaptive gas circulation. BACKGROUND

[0002] In the existing article preservation technology, the storage mode mainly adopts constant atmosphere control (such as setting the proportion of oxygen, carbon dioxide and nitrogen) or single refrigeration preservation mode for environmental regulation, so as to delay the metabolism speed of the article, inhibit the growth of microorganisms, and thus realize the basic preservation purpose. However, the traditional system generally adopts static control means, and the operating parameters in the whole storage period are mostly preset values, which cannot be adjusted in real time according to the dynamic changes of the actual stored articles. More seriously, the existing technology generally lacks the sensing ability of the state evolution of the stored articles in the preservation process, cannot obtain key state information such as whether the article starts to deteriorate and the deterioration speed, and thus cannot dynamically optimize the atmosphere matching, temperature and humidity and other control parameters according to the feedback signals. SUMMARY

[0003] In order to solve the above problems, the application provides an AI physical preservation technology control system with multi-region adaptive gas circulation.

[0004] The application adopts the following technical scheme, an AI physical preservation technology control system with multi-region adaptive gas circulation, comprising:

[0005] A first data acquisition module acquires the storage article parameters of the i-th independent airtight area of the overall storage area, i≥1, wherein the storage article parameters include the article category and the target preservation time length;

[0006] A first strategy generation module inputs the acquired storage article parameters into a pre-constructed first article preservation strategy model, and outputs a first article preservation strategy;

[0007] A second data acquisition module acquires the ripeness characteristic parameters of the stored article, and acquires the maturity coefficient of the article based on the ripeness characteristic parameters of the stored article;

[0008] A second strategy generation module inputs the acquired maturity coefficient, the first article preservation strategy and the storage article parameters into a pre-constructed second article preservation strategy model, and outputs a second article preservation strategy;

[0009] A preservation control module controls the parameter matching, temperature and humidity of the circulating gas of the i-th independent airtight area based on the generated first article preservation strategy and second article preservation strategy;

[0010] The third data acquisition module acquires comprehensive environmental parameters of the i-th independent airtight zone in real time using a set of sensors based on a preset time interval. These comprehensive environmental parameters include ethylene concentration and TVOC gas content. Feature extraction is performed on the acquired comprehensive environmental parameters to obtain the ethylene concentration change rate. and the rate of change of TVOC gas content ;

[0011] The strategy optimization module is based on the rate of change in ethylene concentration. and the rate of change of TVOC gas content Obtain the spoilage level of stored items, obtain the current second item preservation strategy, and trigger different atmosphere optimization strategies for different spoilage levels of stored items.

[0012] As a further description of the above technical solution: the first item preservation strategy includes storage temperature and storage humidity;

[0013] The second item preservation strategy is a gas parameter ratio, which is the ratio of nitrogen, oxygen and carbon dioxide content.

[0014] As a further description of the above technical solution: the training method for the first item preservation strategy model includes:

[0015] G sets of training data are collected in advance, where G is a positive integer greater than 0. The training data includes the item category and the target shelf life, as well as the first item preservation strategy corresponding to the item category and the target shelf life.

[0016] A multi-output gradient boosting regression model was selected.

[0017] The collected data is divided into training set, validation set and test set according to a preset ratio;

[0018] The model is trained using the training set, the optimizer and learning rate are set, the model parameters are updated using the backpropagation algorithm, the loss function is minimized, and the model is validated using the validation set to achieve the best performance on the validation set.

[0019] The trained model is evaluated using a test set, and the mean squared error index of the model is calculated to assess its performance. Once the model's performance evaluation meets the standards, the trained first-item preservation strategy model is deployed and applied.

[0020] As a further description of the above technical solution: the method for obtaining the maturity characteristic parameters of stored items includes:

[0021] By installing industrial cameras on the inbound conveyor belt, image data of items being transported to the i-th independent airtight zone can be obtained;

[0022] The collected images are subjected to resolution unification, color correction and exposure correction to obtain pre-processed images, and then an image segmentation model is used to extract the outline of the article to generate a feature image;

[0023] The feature image is converted from an RGB color space to an HSV color space;

[0024] The HSV image is subjected to kernel density estimation to obtain a target color peak value;

[0025] The hue center value and the standard deviation are calculated.

[0026] As a further description of the above technical solution: the method for obtaining the maturity coefficient of the article based on the maturity feature parameter of the stored article comprises:

[0027] Defining a hue value to a piecewise linear function of maturity score;

[0028] Setting a hue value weight for each hue value in turn;

[0029] The maturity coefficient is calculated based on the weighted sum of the maturity score and the set hue value weight.

[0030] As a further description of the above technical solution: the second article preservation strategy model training method comprises:

[0031] Pre-collecting F sets of training data, F being a positive integer greater than 0, the training data including maturity coefficient, storage temperature and storage humidity and storage article parameters, and the corresponding second article preservation strategy of the maturity coefficient, storage temperature and storage humidity and storage article parameters;

[0032] The F sets of training data are divided into a training set and a validation set, wherein the training set is used to train the second article preservation strategy model, and the validation set is used to evaluate the generalization performance of the second article preservation strategy model;

[0033] During the training of the second article preservation strategy model, the cross-entropy loss function is minimized as the optimization objective, the early stopping strategy is used to monitor the performance of the validation set, the network parameters are continuously adjusted to optimize the model performance; when the prediction accuracy on the validation set reaches the expected accuracy, it is considered that the second article preservation strategy model has converged, and the training is stopped; the second article preservation strategy model is trained using a deep neural network based on a multilayer perceptron;

[0034] The maturity coefficient, storage temperature, storage humidity, and storage article parameters are converted into a feature vector; the input layer of the second article preservation strategy model receives the feature vector, extracts the nonlinear relationship in the data through multiple hidden layers, and finally the output layer of the second article preservation strategy model calculates the probability distribution of the gas parameter ratio through a softmax activation function, and outputs the gas parameter ratio corresponding to the maximum probability as the final prediction result.

[0035] As a further description of the above technical solution: the deterioration grade includes a first deterioration grade, a second deterioration grade, and a third deterioration grade, and the deterioration degrees of the first deterioration grade, the second deterioration grade, and the third deterioration grade increase in turn;

[0036] The method for obtaining the deterioration grade of the storage article includes:

[0037] The preset ethylene concentration change rate threshold Ymax and the TVOC gas content change rate threshold Tmax are obtained;

[0038] When > Ymax, and > Tmax, it is determined to generate a third deterioration grade;

[0039] When < Ymax, and < Tmax, it is determined to generate a first deterioration grade;

[0040] The remaining cases are determined to generate a second deterioration grade.

[0041] As a further description of the above technical solution: the current second article preservation strategy is obtained, and different atmosphere optimization strategies are triggered for different deterioration grades of the storage article, including:

[0042] When the deterioration grade of the storage article is a first deterioration grade, no atmosphere optimization strategy is generated;

[0043] When the deterioration grade of the storage article is a second deterioration grade or a third deterioration grade, the second article preservation strategy and the deterioration grade of the storage article are input into a pre-constructed gas parameter optimization model, and a corrected gas parameter ratio is output as an atmosphere optimization strategy.

[0044] As a further description of the above technical solution: the training method of the gas parameter optimization model includes:

[0045] Q sets of training data are collected in advance, Q is a positive integer greater than 0, the training data includes the second article preservation strategy and the deterioration grade of the storage article, and the gas parameter ratio corresponding to the second article preservation strategy and the deterioration grade of the storage article;

[0046] The deep learning model is selected as the gas parameter optimization model, the training data is used to train the gas parameter optimization model, the second item preservation strategy and the deterioration grade of the stored item are taken as the input of the gas parameter optimization model, the gas parameter ratio is taken as the output of the gas parameter optimization model, the random gradient descent method is adopted, the weight and bias of the gas parameter optimization model are adjusted through the back propagation algorithm, so that the error between the prediction result and the actual result of the gas parameter optimization model is minimized, the loss function is set, the loss function is mean square error, when the loss function value reaches convergence, the training of the gas parameter optimization model is stopped, and the gas parameter optimization model corresponding to the loss function value reaching convergence is taken as the trained gas parameter optimization model.

[0047] As a further description of the above technical solution: the sensor group includes an optical ethylene sensor and a PID type TVOC sensor, the optical ethylene sensor is used to collect the ethylene concentration of the i-th independent airtight area, and the PID type TVOC sensor is used to collect the TVOC gas content of the i-th independent airtight area.

[0048] Beneficial effects:

[0049] In the above technical solution, the control system realizes the synergistic optimization of preservation efficiency through double-strategy deep coupling, the first strategy generation module generates basic temperature and humidity parameters based on item parameters and target preservation period to construct a static storage framework; the second strategy generation module, based on the constructed static storage framework, dynamically analyzes the physiological state change of the item relying on real-time visual maturity analysis, obtains a maturity coefficient, and inputs the maturity coefficient, the first item preservation strategy and the storage item parameters into a neural network model, and the two strategies form a closed-loop linkage - the basic temperature and humidity setting provides a stable environmental benchmark for maturity monitoring, and the maturity feedback drives dynamic correction of the gas ratio, so that the control system can maintain a steady state of the storage environment and accurately adjust the proportion of gas components according to the real-time state of the item, and finally realize the accurate matching of environmental parameters and physiological needs of the item, significantly improving the preservation adaptive ability.

[0050] Further, by introducing a third data acquisition module and a strategy optimization module, the ethylene concentration and TVOC content in each independent airtight area can be collected in real time based on a set time interval, and the change rate thereof is calculated, so as to accurately reflect the deterioration trend of the stored goods. By constructing a deterioration grade determination, the dynamic change characteristics of ethylene and TVOC are mapped into three types of deterioration grades, so as to realize scientific division and identification of the fresh-keeping state of the goods. In combination with the second goods fresh-keeping strategy and the gas parameter optimization model constructed by deep learning, a corrected gas ratio is generated under the second or third deterioration grade, and the storage atmosphere is dynamically adjusted. This method not only improves the early perception ability and response speed of the system to the deterioration signs, avoids lagging or excessive intervention, but also realizes differentiation and precision of the atmosphere adjustment, effectively prolongs the fresh-keeping period of the goods, reduces energy consumption, and significantly enhances the adaptability, intelligence and stability of the overall fresh-keeping system. BRIEF DESCRIPTION OF DRAWINGS

[0051] The application will be further explained in conjunction with the drawings and examples:

[0052] Figure 1 A module connection diagram of the AI physical fresh-keeping technology control system of the multi-zone adaptive gas circulation provided for Embodiment 1 of the application;

[0053] Figure 2 A module connection diagram of the AI physical fresh-keeping technology control system of the multi-zone adaptive gas circulation provided for Embodiment 2 of the application;

[0054] Figure 3 A flowchart of the AI physical fresh-keeping technology control method of the multi-zone adaptive gas circulation provided for Embodiment 3 of the application;

[0055] Figure 4 A flowchart of the method for obtaining the ripeness characteristic parameters of the stored goods provided for Embodiment 1 of the application. DETAILED DESCRIPTION

[0056] In order to make the technical means, creative features, purposes and effects achieved by the application easy to understand, the application will be further described in conjunction with specific drawings. It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0057] Embodiment 1

[0058] Please refer to Figure 1 and Figure 4 The embodiment of the application provides a technical solution: an AI physical fresh-keeping technology control system of multi-zone adaptive gas circulation, which comprises a first data acquisition module, a first strategy generation module, a second data acquisition module, a second strategy generation module and a fresh-keeping control module, and each module is connected through wired and / or wireless connection:

[0059] a first data acquisition module, configured to acquire a storage item parameter of an i-th independent airtight area of the overall storage area, i≥1, the storage item parameter comprising an item category and a target preservation time length;

[0060] It should be noted that the overall storage area has M independent airtight areas, M being a positive integer greater than 1, wherein the M independent airtight areas are isolated by a sealing door or an air curtain, and each independent airtight area is provided with an air inlet and an air outlet.

[0061] It should be noted that the item category includes green leaf vegetables, apples, strawberries, shiitake mushrooms, etc., and the item category and the target preservation time length are manually input by a user through an operation interface.

[0062] a first strategy generation module, configured to input the acquired storage item parameter into a pre-constructed first item preservation strategy model, and output a first item preservation strategy, the first item preservation strategy comprising a storage temperature and a storage humidity;

[0063] The first item preservation strategy model training method comprises:

[0064] pre-collecting G groups of training data, G being a positive integer greater than 0, the training data comprising an item category and a target preservation time length, and a first item preservation strategy corresponding to the item category and the target preservation time length;

[0065] a multi-output gradient boosting regression model is selected;

[0066] The collected data is divided into a training set, a validation set and a test set according to a preset ratio;

[0067] The training set is used to train the model, an optimizer and a learning rate are set, the parameters of the model are updated through a back propagation algorithm, a loss function is minimized, and the validation set is used for validation to make the model achieve the best performance on the validation set;

[0068] The test set is used to evaluate the trained model, the mean square error index of the model is calculated, the performance of the model is evaluated, and after the performance evaluation of the model meets the standard, the trained first item preservation strategy model is deployed.

[0069] a second data acquisition module, configured to acquire a ripeness characteristic parameter of the storage item, and acquire a maturity coefficient of the item based on the ripeness characteristic parameter of the storage item;

[0070] The method for acquiring the ripeness characteristic parameter of the storage item comprises:

[0071] An industrial camera is installed on an entry conveyor belt to acquire image data of the item conveyed to the i-th independent airtight area;

[0072] The collected images are subjected to resolution unification, color correction and exposure correction to obtain pre-processed images, and then an image segmentation model (such as Mask R-CNN) is used to extract the outline of the article to generate a feature image;

[0073] It should be noted that the pre-trained Mask R-CNN model is used to input the pre-processed image, and the model outputs a binary mask of the article. The mask identifies the article region (foreground is 1 and background is 0). The feature image is the pixel-by-pixel product of the mask and the original image, and only the article region is retained.

[0074] The feature image is converted from the RGB color space to the HSV color space;

[0075] Kernel density estimation is used on the HSV image to obtain the target color peak value;

[0076] The method for obtaining the target color peak value comprises:

[0077] Collecting the H values of all non-zero pixels of the HSV image , wherein is the total number of pixels in the article region, is the H value of the th non-zero pixel.

[0078] The kernel density estimation is used to detect the target color peak value, and the kernel density formula is:

[0079] ;

[0080] wherein, ; ;

[0081] In the formula, is the density estimation at the hue value , is a bandwidth parameter, is a ring distance function, is a Gaussian kernel function;

[0082] Peak detection, find all peaks in the range that satisfy the following conditions :

[0083] ;

[0084] wherein, is the search step, and optionally ; , is a constant, and optionally 1.5, is the mean value of , is the mean value of the standard deviation of hue;

[0085] calculating the hue center value and the standard deviation;

[0086] It should be noted that the hue center value can determine the maturity stage, and the standard deviation is used to determine the uniformity of maturity.

[0087] hue center value ;

[0088] the hue center value (unit: °) of the first main color, representing the average hue angle of the main color, is the serial number of the main color peak (for example, =1 represents the first main color, =2 represents the second main color, etc.); is the position (unit: °) of the first peak value found by kernel density estimation, ranging from [0°, 360°);

[0089] The formula for calculating the standard deviation is:

[0090] ;

[0091] In the formula, is the standard deviation (unit: °) of the first main color, used to quantify the dispersion degree of color distribution; a small value indicates that the color is pure and concentrated; a large value indicates that the color is dispersed and mixed; is the number of valid pixels in the neighborhood , is the annular neighborhood centered on , is the hue center value of the current main color, is the hue value of the first pixel in the neighborhood ;

[0092] The method for obtaining the maturity coefficient of the stored goods based on the maturity characteristic parameters of the goods includes:

[0093] defining a piecewise linear function of hue value to maturity score;

[0094] ; In the formula is the maturity score, is the unripe threshold value, is the ripe threshold value;

[0095] The hue value weight is set for each hue value in sequence, and a method for setting the hue value weight comprises: ; wherein is a hue value weight of the th primary color, ; is a sensitivity parameter, and the sensitivity parameter is optionally = 20°, is a standard deviation of the th primary color; , is a number of pixels in a neighborhood , is a total number of pixels.

[0096] A maturity coefficient is calculated based on a weighted summation of the maturity scores and the set hue value weights;

[0097] A calculation formula of the maturity coefficient is: ;

[0098] is the maturity coefficient, a number of detected primary color peak values, a hue center value of the th primary color, a maturity score of the primary color, a hue value weight of the th primary color;

[0099] A second strategy generation module inputs the obtained maturity coefficient, the first item preservation strategy, and the storage item parameter into a pre-constructed second item preservation strategy model, and outputs a second item preservation strategy; the second item preservation strategy is a gas parameter ratio, and the gas parameter ratio is a content ratio of nitrogen, oxygen, and carbon dioxide;

[0100] The second item preservation strategy model training method comprises:

[0101] F sets of training data are pre-collected, F is a positive integer greater than 0, the training data include the maturity coefficient, the storage temperature, the storage humidity, and the storage item parameter, and the maturity coefficient, the storage temperature, the storage humidity, and the storage item parameter correspond to the second item preservation strategy;

[0102] The F sets of training data are divided into a training set and a validation set, wherein the training set is used to train the second item preservation strategy model, and the validation set is used to evaluate the generalization performance of the second item preservation strategy model;

[0103] In the training process of the second item preservation strategy model, a cross-entropy loss function is minimized as an optimization objective, an early stopping strategy is used to monitor the performance of the validation set, and the model performance is optimized by continuously adjusting network parameters; when the prediction accuracy on the validation set reaches the expected accuracy, it is considered that the second item preservation strategy model has converged, and the training is stopped; the second item preservation strategy model is trained using a deep neural network based on a multilayer perceptron;

[0104] The maturity coefficient, storage temperature and humidity, and storage item parameters are converted into a feature vector; the input layer of the second item preservation strategy model receives the feature vector, extracts the nonlinear relationship in the data through multiple hidden layers, and finally the output layer of the second item preservation strategy model calculates the probability distribution of the gas parameter ratio through a softmax activation function, and outputs the gas parameter ratio corresponding to the maximum probability as the final prediction result.

[0105] The preservation control module controls the parameter ratio, temperature and humidity of the circulating gas in the ith independent airtight area based on the generated first item preservation strategy and second item preservation strategy.

[0106] In this embodiment, the control system realizes the synergistic optimization of preservation efficiency through double-strategy deep coupling, the first strategy generation module generates basic temperature and humidity parameters based on item parameters and target preservation period to construct a static storage framework; the second strategy generation module is based on the constructed static storage framework, and relies on real-time visual maturity analysis to dynamically analyze the physiological state change of the item, obtain the maturity coefficient, and input the maturity coefficient, the first item preservation strategy and the storage item parameters into the neural network model together, and the two strategies form a closed loop linkage - the basic temperature and humidity setting provides a stable environmental benchmark for maturity monitoring, and the maturity feedback drives the dynamic correction of the gas ratio, so that the control system can maintain the stability of the storage environment and accurately adjust the proportion of the gas component according to the real-time state of the item, and finally realize the accurate matching of the environmental parameters and the physiological needs of the item, and significantly improve the preservation adaptive ability.

[0107] Embodiment 2

[0108] Please refer to Figure 2 On the basis of the above-mentioned embodiments, the third data acquisition module and the strategy optimization module are added in this embodiment.

[0109] The third data acquisition module acquires the comprehensive environmental parameters of the i-th independent airtight area in real time based on a preset time interval through a set sensor group, the comprehensive environmental parameters including ethylene concentration and TVOC gas content, and the acquired comprehensive environmental parameters are subjected to feature extraction to acquire the ethylene concentration change rate and the TVOC gas content change rate; wherein the sensor group includes an optical ethylene sensor and a PID type TVOC sensor, the optical ethylene sensor being used to collect the ethylene concentration of the i-th independent airtight area, and the PID type TVOC sensor being used to collect the TVOC gas content of the i-th independent airtight area.

[0110] The method for acquiring the ethylene concentration change rate and the TVOC gas content change rate includes that the calculation formula of the ethylene concentration change rate is: ; in the formula, is the ethylene concentration change rate, is the detection time interval, is the difference between the current detection ethylene concentration and the last detection ethylene concentration;

[0111] The calculation formula of the TVOC gas content change rate is: ; in the formula, is the TVOC gas content change rate, is the detection time interval, is the difference between the current detection TVOC gas content and the last detection TVOC gas content;

[0112] The strategy optimization module acquires the deterioration grade of the stored goods based on the ethylene concentration change rate and the TVOC gas content change rate acquires the current second goods preservation strategy, triggers different atmosphere optimization strategies for different deterioration grades of the stored goods, and

[0113] The deterioration grade includes a first deterioration grade, a second deterioration grade and a third deterioration grade, and the deterioration degrees of the first deterioration grade, the second deterioration grade and the third deterioration grade increase in turn.

[0114] The method for acquiring the deterioration grade of the stored goods includes:

[0115] A preset ethylene concentration change rate threshold Ymax and a TVOC gas content change rate threshold Tmax are set;

[0116] When > Ymax, and > Tmax, it is determined that the third deterioration grade is generated.

[0117] When < Ymax, and < Tmax, it is determined that the first deterioration grade is generated.determining to generate the first deterioration grade when Tmax;

[0118] determining to generate the second deterioration grade for the remaining cases.

[0119] In the embodiment, the ethylene concentration and the TVOC content are introduced as gas index parameters of the deterioration of the stored goods, which helps to timely capture the key volatile organic compounds released by fruits and vegetables and other stored goods in the process of ripening- over-ripening- deterioration. Compared with the traditional single- direction decision limit relying only on the ripeness coefficient and the initial preservation strategy model, the change rates of ethylene and TVOC are judged by the preset threshold in a double- index combined manner, and three types of deterioration grades are scientifically divided, which further refines the granularity of the stored state recognition.

[0120] The method for obtaining the current second goods preservation strategy includes:

[0121] When the deterioration grade of the stored goods is the first deterioration grade, no atmosphere optimization strategy is generated.

[0122] When the deterioration grade of the stored goods is the second deterioration grade or the third deterioration grade, the second goods preservation strategy and the deterioration grade of the stored goods are input into a pre-constructed gas parameter optimization model, and a corrected gas parameter ratio is output as the atmosphere optimization strategy.

[0123] The training method of the gas parameter optimization model includes:

[0124] Q sets of training data are collected in advance, Q is a positive integer greater than 0, the training data includes the second goods preservation strategy and the deterioration grade of the stored goods, and the gas parameter ratio corresponding to the second goods preservation strategy and the deterioration grade of the stored goods;

[0125] A deep learning model is selected as the gas parameter optimization model, the training data is used to train the gas parameter optimization model, the second goods preservation strategy and the deterioration grade of the stored goods are used as the input of the gas parameter optimization model, and the gas parameter ratio is used as the output of the gas parameter optimization model. The random gradient descent method is adopted, the weight and bias of the gas parameter optimization model are adjusted through the back propagation algorithm, so that the error between the prediction result of the gas parameter optimization model and the actual result is minimized. A loss function is set, the loss function is mean square error, when the loss function value reaches convergence, the training of the gas parameter optimization model is stopped, and the gas parameter optimization model corresponding to the loss function value reaching convergence is used as the trained gas parameter optimization model.

[0126] In this embodiment, by introducing the third data acquisition module and the strategy optimization module, the ethylene concentration and the TVOC content in each independent airtight area can be collected in real time based on the set time interval, and the change rate thereof can be calculated, so that the deterioration trend of the stored goods can be accurately reflected. By constructing a deterioration grade determination, the dynamic change characteristics of ethylene and TVOC are mapped into three types of deterioration grades, so that the scientific division and identification of the fresh-keeping state of the goods can be realized. In combination with the second goods fresh-keeping strategy and the gas parameter optimization model constructed by deep learning, the corrected gas ratio is generated under the second or third deterioration grade, and the storage atmosphere is dynamically adjusted. This method not only improves the early perception ability and response speed of the system to the deterioration signs, avoids lagging or excessive intervention, but also realizes the differentiation and precision of the atmosphere adjustment, effectively prolongs the fresh-keeping period of the goods, reduces the energy consumption, and significantly enhances the adaptability, intelligence and stability of the overall fresh-keeping system.

[0127] Embodiment 3

[0128] Please refer to Figure 3 The embodiment of the present application provides a technical solution: an AI physical fresh-keeping technology control method for multi-region adaptive gas circulation, comprising:

[0129] Obtain the i-th independent airtight area storage goods parameters of the overall storage area, wherein the storage goods parameters include the goods category and the target fresh-keeping time length, the goods category includes green leaf vegetables, apples, strawberries, shiitake mushrooms, etc., and the goods category and the target fresh-keeping time length are manually input by the user through the operation interface;

[0130] Input the obtained storage goods parameters into the pre-constructed first goods fresh-keeping strategy model, and output the first goods fresh-keeping strategy; the first goods fresh-keeping strategy includes the storage temperature and the storage humidity;

[0131] Obtain the ripeness characteristic parameters of the stored goods, and obtain the maturity coefficient of the goods based on the ripeness characteristic parameters of the stored goods;

[0132] Input the obtained maturity coefficient, the first goods fresh-keeping strategy and the storage goods parameters into the pre-constructed second goods fresh-keeping strategy model, and output the second goods fresh-keeping strategy; the second goods fresh-keeping strategy is a gas parameter ratio, i.e., the content ratio of nitrogen, oxygen and carbon dioxide;

[0133] Control the parameter ratio, temperature and humidity of the delivery circulating gas of the i-th independent airtight area based on the generated first goods fresh-keeping strategy and second goods fresh-keeping strategy;

[0134] Based on a preset time interval, a comprehensive environment parameter of the i-th independent airtight area is acquired in real time by a set sensor group, the comprehensive environment parameter including ethylene concentration and TVOC gas content, feature extraction is performed on the acquired comprehensive environment parameter, and an ethylene concentration change rate and a TVOC gas content change rate are acquired;

[0135] Based on the ethylene concentration change rate and the TVOC gas content change rate, a deterioration grade of the stored article is acquired, a current second article preservation strategy is acquired, and different atmosphere optimization strategies are triggered for different deterioration grades of the stored article.

[0136] The basic principles, main features and advantages of the present application are shown and described above. It should be understood by those skilled in the art that the present application is not limited by the above examples, and the above examples and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A multi-zone adaptive gas circulation AI physical preservation technology control system, characterized in that, include: The first data acquisition module acquires the storage item parameters of the i-th independent airtight zone of the overall storage area, where i≥1. The storage item parameters include the item category and the target preservation time. The first strategy generation module inputs the acquired storage item parameters into the pre-built first item preservation strategy model and outputs the first item preservation strategy, which includes storage temperature and storage humidity. The second data acquisition module acquires the maturity characteristic parameters of the stored items and obtains the maturity coefficient of the items based on the maturity characteristic parameters of the stored items. The second strategy generation module inputs the obtained maturity coefficient, the first item preservation strategy, and the stored item parameters into the pre-built second item preservation strategy model, and outputs the second item preservation strategy, which is the gas parameter ratio. The preservation control module controls the parameter ratio, temperature, and humidity of the circulating gas in the i-th independent airtight zone based on the generated first and second item preservation strategies. The third data acquisition module acquires comprehensive environmental parameters of the i-th independent airtight zone in real time using a set of sensors based on a preset time interval. These comprehensive environmental parameters include ethylene concentration and TVOC gas content. Feature extraction is performed on the acquired comprehensive environmental parameters to obtain the ethylene concentration change rate. and the rate of change of TVOC gas content ; The strategy optimization module is based on the rate of change in ethylene concentration. and the rate of change of TVOC gas content Obtain the spoilage level of stored items, obtain the current second item preservation strategy, and trigger different atmosphere optimization strategies for different spoilage levels of stored items. The degradation levels include a first degradation level, a second degradation level, and a third degradation level, and the degree of degradation increases sequentially from the first degradation level to the second degradation level and the third degradation level. Methods for obtaining the spoilage level of stored items include: Preset the threshold values ​​for the rate of change of ethylene concentration (Ymax) and the threshold values ​​for the rate of change of TVOC gas content (Tmax); when >Ymax, and When Tmax is greater than Tmax, the third mutation level is determined to be generated; when <Ymax, and When Tmax < Tmax, determine that the first mutation level has been generated; The remaining cases are judged to generate a second deterioration level.

2. The AI ​​physical preservation technology control system with multi-zone adaptive gas circulation according to claim 1, characterized in that, The gas parameter ratio refers to the ratio of nitrogen, oxygen, and carbon dioxide content.

3. The AI ​​physical preservation technology control system with multi-zone adaptive gas circulation according to claim 1, characterized in that, The training method for the first item preservation strategy model includes: G sets of training data are collected in advance, where G is a positive integer greater than 0. The training data includes the item category and the target shelf life, as well as the first item preservation strategy corresponding to the item category and the target shelf life. A multi-output gradient boosting regression model was selected. The collected data is divided into training set, validation set and test set according to a preset ratio; The model is trained using the training set, the optimizer and learning rate are set, the model parameters are updated using the backpropagation algorithm, the loss function is minimized, and the model is validated using the validation set to achieve the best performance on the validation set. The trained model is evaluated using a test set, and the mean squared error index of the model is calculated to assess its performance. Once the model's performance evaluation meets the standards, the trained first-item preservation strategy model is deployed and applied.

4. The AI ​​physical preservation technology control system with multi-region adaptive gas circulation according to claim 1, characterized in that, The method for obtaining the maturity characteristic parameters of stored items includes: By installing industrial cameras on the inbound conveyor belt, image data of items being transported to the i-th independent airtight zone can be obtained; The acquired images are subjected to resolution unification, color correction, and exposure correction to obtain preprocessed images. Then, an image segmentation model is used to extract the object contours and generate feature images. Convert the feature image from the RGB color space to the HSV color space; Kernel density estimation is used to obtain the target color peaks in HSV images; Calculate and obtain the hue center value and standard deviation.

5. The AI ​​physical preservation technology control system with multi-zone adaptive gas circulation according to claim 4, characterized in that, The method for obtaining the maturity coefficient of stored items based on maturity characteristic parameters includes: Define hue value A piecewise linear function leading to maturity scores; Assign a weight to each hue value in turn; The maturity coefficient is calculated by weighting the maturity score and the set hue value.

6. The AI ​​physical preservation technology control system with multi-zone adaptive gas circulation according to claim 2, characterized in that, The training method for the second item preservation strategy model includes: F sets of training data were collected in advance, where F is a positive integer greater than 0. The training data included maturity coefficient, storage temperature and storage humidity, and parameters of stored items, as well as the second item preservation strategy corresponding to maturity coefficient, storage temperature and storage humidity, and parameters of stored items. The F sets of training data are divided into a training set and a validation set, wherein the training set is used to train the second item preservation strategy model, and the validation set is used to evaluate the generalization performance of the second item preservation strategy model. During the training of the second item preservation strategy model, minimizing the cross-entropy loss function is used as the optimization objective. An early stopping strategy is used to monitor the performance of the validation set. The model performance is optimized by continuously adjusting the network parameters. When the prediction accuracy on the validation set reaches the expected accuracy, the second item preservation strategy model is considered to have converged and training is stopped. The second item preservation strategy model is trained using a deep neural network based on a multilayer perceptron. The maturity coefficient, storage temperature, storage humidity, and stored item parameters are converted into feature vectors. The input layer of the second item preservation strategy model receives the feature vectors and extracts nonlinear relationships in the data through multiple hidden layers. Finally, the output layer of the second item preservation strategy model calculates the probability distribution of gas parameter ratios through a softmax activation function and outputs the gas parameter ratio corresponding to the highest probability as the final prediction result.

7. The AI ​​physical preservation technology control system with multi-zone adaptive gas circulation according to claim 1, characterized in that, Obtain the current preservation strategy for the second item. For different spoilage levels of stored items, trigger different atmosphere optimization strategies, including: When the spoilage level of stored items is the first spoilage level, no atmosphere optimization strategy is generated; When the spoilage level of the stored goods is the second or third spoilage level, the second item preservation strategy and the spoilage level of the stored goods are input into the pre-built gas parameter optimization model, and the corrected gas parameter ratio is output as the atmosphere optimization strategy.

8. The AI ​​physical preservation technology control system with multi-zone adaptive gas circulation according to claim 7, characterized in that, The training method for the gas parameter optimization model includes: Q sets of training data are collected in advance, where Q is a positive integer greater than 0. The training data includes the second item preservation strategy and the spoilage level of the stored item, as well as the gas parameter ratios corresponding to the second item preservation strategy and the spoilage level of the stored item. A deep learning model was selected as the gas parameter optimization model. Training data was used to train the gas parameter optimization model. The second item preservation strategy and the spoilage level of the stored items were used as inputs to the gas parameter optimization model, and the gas parameter ratio was used as the output. Stochastic gradient descent was employed, and the weights and biases of the gas parameter optimization model were adjusted using the backpropagation algorithm to minimize the error between the predicted and actual results. A loss function, the mean squared error, was set. When the loss function value converged, the training of the gas parameter optimization model was stopped, and the gas parameter optimization model corresponding to the convergence of the loss function value was taken as the trained gas parameter optimization model.

9. The AI ​​physical preservation technology control system with multi-zone adaptive gas circulation according to claim 1, characterized in that, The sensor group includes an optical ethylene sensor and a PID-type TVOC sensor. The optical ethylene sensor is used to collect the ethylene concentration in the i-th independent airtight zone, and the PID-type TVOC sensor is used to collect the TVOC gas content in the i-th independent airtight zone.

Images