Sterilizer control method and system based on cold chain fresh products

Abstract

The present disclosure provides a control method and system for a disinfection machine based on cold chain fresh-keeping products, which relates to intelligent control technology for equipment. The method includes: obtaining an image data sequence of the fresh-keeping products to be disinfected; generating a product information sequence; obtaining an optimal disinfection concentration sequence, which is generated by performing a disinfection demand analysis; obtaining a disinfection concentration deviation sequence, which is obtained by sequentially calculating adjacent disinfection concentration deviations for the optimal disinfection concentration sequence; generating a spray control optimization parameter sequence for the cold chain disinfection machine based on the disinfection concentration deviation sequence; and performing disinfection processing on the fresh-keeping products to be disinfected according to the spray control optimization parameter sequence. The method can solve the technical problem of waste of disinfectant resources caused by the use of fixed spray power for disinfection spraying when the existing cold chain disinfection machine is performing disinfection on fresh-keeping products, thereby saving the consumption of disinfectant resources when the cold chain disinfection machine is working and reducing resource waste.

Description

Technical Field

[0001] The present disclosure relates to intelligent equipment control technology, and more specifically, to a control method and system for a sterilizer based on cold chain fresh products. Background Art

[0002] A cold chain disinfection machine is a device for disinfecting cold chain products. The disinfection method of a cold chain disinfection machine is usually spray disinfection, that is, the disinfectant solution is converted into a disinfectant spray through a nozzle device to disinfect the cold chain products. Due to the different types and specifications of the products to be disinfected during spray disinfection, the required disinfection spray concentration is not exactly the same. However, traditional cold chain disinfection machines usually use a fixed spray power for disinfection spray, which easily causes waste of disinfectant resources.

[0003] The disadvantage of existing cold chain disinfection machines in disinfecting fresh products is that the use of fixed spray power for disinfection spraying causes waste of disinfectant resources. Summary of the Invention

[0004] Therefore, in order to solve the above technical problems, the technical solutions adopted in the embodiments of the present disclosure are as follows:

[0005] A control method for a sterilizer based on cold chain fresh-keeping products comprises the following steps: obtaining an image data sequence of the fresh-keeping products to be sterilized, wherein the image data sequence is obtained by sequentially capturing images of the fresh-keeping products to be sterilized by a CCD image sensor; generating a product information sequence, wherein the product information sequence is obtained by performing product identification on the image data sequence through a product identification channel; obtaining an optimal disinfection concentration sequence, wherein the optimal disinfection concentration sequence is generated by performing disinfection demand analysis on the product information sequence; obtaining a disinfection concentration deviation sequence, wherein the disinfection concentration deviation sequence is obtained by sequentially calculating adjacent disinfection concentration deviations for the optimal disinfection concentration sequence; generating a spray control optimization parameter sequence of the cold chain sterilizer based on the disinfection concentration deviation sequence; and performing disinfection treatment on the fresh-keeping products to be sterilized according to the spray control optimization parameter sequence.

[0006] A control system for a sterilizer based on cold chain fresh-keeping products includes: an image data sequence acquisition module, which is used to acquire an image data sequence of the fresh-keeping products to be sterilized, and the image data sequence is obtained by sequentially capturing images of the fresh-keeping products to be sterilized by a CCD image sensor; a product information sequence generation module, which is used to generate a product information sequence, and the product information sequence is obtained by performing product identification on the image data sequence through a product identification channel; an optimal disinfection concentration sequence acquisition module, which is used to obtain an optimal disinfection concentration sequence, and the optimal disinfection concentration sequence is generated by performing disinfection demand analysis on the product information sequence; a disinfection concentration deviation sequence acquisition module, which is used to obtain a disinfection concentration deviation sequence, and the disinfection concentration deviation sequence is obtained by sequentially calculating adjacent disinfection concentration deviations for the optimal disinfection concentration sequence; a spray control optimization parameter sequence generation module, which is used to generate a spray control optimization parameter sequence of a cold chain sterilizer based on the disinfection concentration deviation sequence; and a fresh-keeping product disinfection treatment module, which is used to perform disinfection treatment on the fresh-keeping products to be sterilized according to the spray control optimization parameter sequence.

[0007] Due to the adoption of the above technical methods, the present disclosure has achieved the following technical advances compared to the prior art:

[0008] The technical problem of the waste of disinfectant resources caused by the use of fixed spray power for disinfection spraying when the existing cold chain disinfection machine disinfects fresh products can be solved. First, the fresh products to be disinfected are sequentially imaged by the CCD image sensor to obtain an image data sequence of the fresh products to be disinfected; a product identification channel is constructed based on a convolutional neural network, and the product information of the image data sequence is identified by the product identification channel to generate a product information sequence; then, the disinfection demand analysis is performed on the product information sequence by the disinfection analysis unit to obtain an optimal disinfection concentration sequence; further, the adjacent disinfection concentration deviations are calculated in sequence according to the optimal disinfection concentration sequence to obtain a disinfection concentration deviation sequence; the disinfection concentration deviation sequence is subjected to a cold chain disinfection machine spray control parameter analysis to obtain a spray control optimization parameter sequence; finally, the disinfection treatment of the fresh products to be disinfected is performed according to the spray control optimization parameter sequence. The above method can save the consumption of disinfectant resources when the cold chain disinfection machine is working and reduce resource waste. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the following briefly introduces the drawings required for describing the embodiments.

[0010] Figure 1A flow chart of a control method for a cold chain fresh food sterilizer is provided for this application;

[0011] Figure 2 This application provides a flow chart of a method for controlling a cold chain fresh-keeping product disinfection machine, in which a disinfection analysis unit performs disinfection demand analysis on a product information sequence;

[0012] Figure 3 A structural diagram of a sterilizer control system for cold chain fresh products is provided for this application.

[0013] Explanation of the accompanying symbols: image data sequence acquisition module 11, product information sequence generation module 12, optimal disinfection concentration sequence acquisition module 13, disinfection concentration deviation sequence acquisition module 14, spray control optimization parameter sequence generation module 15, fresh product disinfection processing module 16. DETAILED DESCRIPTION

[0014] The following will be combined with the accompanying drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. Based on the embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present disclosure.

[0015] Based on the above description, if Figure 1 As shown, the present disclosure provides a control method for a sterilizer based on cold chain fresh products, comprising:

[0016] The method provided in the present application is used to optimize the spray control parameters of a cold chain disinfection machine based on the type of cold chain fresh-keeping products, so as to save the consumption of disinfectant resources when the cold chain disinfection machine is working and reduce resource waste. The cold chain disinfection machine is a spray disinfection machine, and the method is specifically implemented in a disinfection machine control system based on cold chain fresh-keeping products.

[0017] Acquire an image data sequence of the fresh-keeping product to be sterilized, wherein the image data sequence is obtained by sequentially capturing images of the fresh-keeping product to be sterilized using a CCD image sensor;

[0018] In an embodiment of the present application, first, the images of the fresh-keeping products to be sterilized are sequentially captured by a CCD image sensor, wherein the CCD image sensor is a device that uses the principle of charge transfer to capture optical images, and has the advantages of high acquisition accuracy and a wide viewing angle range. It is used to capture images of the fresh-keeping products on a conveyor belt, and the conveyor belt is connected to a cold chain sterilizer. The CCD image sensor is placed in a position area N meters in front of the sterilization area of the cold chain sterilizer, wherein N is an integer greater than 10, and the value of N can be set by technicians in this field according to actual conditions, to obtain image data of multiple fresh-keeping products to be sterilized, wherein the image data is marked with an acquisition time, and then the image data of the multiple fresh-keeping products to be sterilized are sorted in the order of the acquisition time to obtain an image data sequence. By generating an image data sequence of the fresh-keeping products to be sterilized, data support is provided for the next step of product information identification.

[0019] generating a product information sequence, wherein the product information sequence is obtained by performing product identification on the image data sequence through a product identification channel;

[0020] In an embodiment of the present application, a product identification channel is constructed based on a convolutional neural network, and the image data sequence is sequentially input into the product identification channel for product information identification to generate a product information sequence.

[0021] In one embodiment, the method further comprises:

[0022] Connecting to a cold chain warehouse management system to retrieve a plurality of fresh-keeping product image data and a plurality of fresh-keeping product product information, wherein the fresh-keeping product image data and the fresh-keeping product product information have a corresponding relationship;

[0023] Constructing a sample training data set based on the plurality of fresh food image data and the plurality of fresh food product information;

[0024] A product recognition channel is constructed based on a convolutional neural network, and supervised training and verification training are performed on the product recognition channel using the sample training data set to obtain a product recognition channel that meets expected requirements.

[0025] In an embodiment of the present application, first, a cold chain warehouse management system is connected, wherein the cold chain warehouse management system is a software system for warehousing, operating, and managing cold chain fresh-keeping products, wherein a large amount of data and information about cold chain fresh-keeping products is stored. Then, in the cold chain warehouse management system, multiple fresh-keeping product image data and multiple fresh-keeping product information are retrieved, wherein the fresh-keeping product image data and the fresh-keeping product information have a corresponding relationship, and the fresh-keeping product information includes information such as the type, specification, and appearance structure of the fresh-keeping product, and the fresh-keeping product image data and the fresh-keeping product information have a corresponding relationship. A sample training data set is constructed based on the multiple fresh-keeping product image data and the multiple fresh-keeping product information.

[0026] A product identification channel is constructed based on a convolutional neural network. The product identification channel is a neural network model that can be iteratively optimized in machine learning and is obtained through supervised training of a training data set. The input data of the product identification channel is fresh-keeping product image data, and the output data is fresh-keeping product information. Then, supervised training is performed on the product identification channel through the sample training data set. First, the sample training data set is divided into a sample training set and a sample verification set according to a preset data division ratio. The preset data division ratio can be set by those skilled in the art according to the actual data volume of the sample training data set. For example, the sample training set ratio is set to 90% and the sample verification set ratio is set to 10%; then, a sample training data is randomly selected from the sample training set as the first sample data, and supervised training is performed on the product identification channel through the first sample data to output first product identification information. Then, the first product identification information is compared with the product information of the first sample data. When the first product identification information is compared with the product information of the first sample data, the first product identification information is compared with the product information of the first sample data. When the product information of the first product identification information is consistent with the product information of the first sample data, the product identification channel is supervised and trained through the next set of sample data. When the first product identification information is inconsistent with the product information of the first sample data, the information deviation between the first product identification information and the product information of the first sample data is obtained, and the weight parameters of the product identification channel are optimized according to the information deviation. Then, supervised training of the next set of sample data is performed, and iterative training is continuously performed until the product identification channel tends to converge. Then, the output result of the product identification channel is verified by the sample verification set to obtain a preset verification accuracy index. The preset verification accuracy index can be set by those skilled in the art according to actual needs. The higher the required accuracy, the larger the preset verification accuracy index. For example, the preset verification accuracy index is set to 96%, and the accuracy of the output result of the product identification channel is judged according to the preset verification accuracy index. When the output result accuracy is greater than or equal to the verification accuracy index, the trained product identification channel is obtained. By constructing a product identification channel based on a convolutional neural network, the efficiency and accuracy of fresh-keeping product identification can be improved, and at the same time, support is provided for obtaining the optimal disinfection concentration sequence in the next step.

[0027] Obtaining an optimal disinfection concentration sequence, wherein the optimal disinfection concentration sequence is generated by performing a disinfection demand analysis on the product information sequence;

[0028] In an embodiment of the present application, first, a disinfection requirement analysis is performed on the product information sequence in sequence, and an optimal disinfection concentration sequence is obtained based on the disinfection requirement analysis results.

[0029] like Figure 2As shown, in one embodiment, the method further includes:

[0030] Perform disinfection data retrieval using the plurality of fresh-keeping product information as retrieval conditions to obtain a plurality of sets of historical qualified disinfection information corresponding to the plurality of fresh-keeping product information, wherein the historical qualified disinfection information includes a disinfectant type and a disinfectant concentration;

[0031] Randomly selecting a first piece of fresh-keeping product information from the plurality of pieces of fresh-keeping product information, and obtaining a first set of historical qualified disinfection information of the first fresh-keeping product information;

[0032] Clustering the first historical qualified disinfection information based on disinfectant type to obtain M disinfectant concentration data sets, where M is the number of disinfectant types;

[0033] Determining M disinfection concentration thresholds based on the M disinfectant concentration data sets, and determining M test concentration data sets based on a preset test step length and the M disinfection concentration thresholds;

[0034] In this embodiment of the present application, based on big data technology, a historical disinfection data search is performed using the multiple fresh-keeping product information as information retrieval conditions to obtain multiple sets of historical qualified disinfection information corresponding to the multiple fresh-keeping product information. The historical qualified disinfection information includes disinfectant type and disinfectant concentration. The disinfectant types include alcohol, peroxide disinfectants, phenolic disinfectants, etc., and the historical qualified disinfection refers to disinfection data that meets the disinfection quality requirements.

[0035] Then, a first fresh-keeping product information is randomly selected from the multiple fresh-keeping product information, where the first fresh-keeping product information is any one of the multiple fresh-keeping product information, and a first set of historical qualified disinfection information for the first fresh-keeping product information is obtained. The first historical qualified disinfection information is then partitioned according to the disinfectant type to obtain M disinfectant concentration data sets, where M is the number of disinfectant types. The maximum and minimum disinfectant concentration values in the M disinfectant concentration data sets are then extracted, and disinfection concentration thresholds are constructed based on the maximum and minimum disinfectant concentration values to obtain M disinfection concentration thresholds.

[0036] Obtain a preset test step length. Those skilled in the art can set the preset test step length based on actual needs. The higher the required accuracy, the shorter the preset test step length. Then, determine M test concentration data sets based on the preset test step length and the M disinfection concentration thresholds. For example, assuming the disinfection concentration threshold is 12-13 and the preset test step length is 0.1, the test concentration data are 12, 12.1, 12.2, etc. Then, disinfection tests are sequentially performed on the first fresh-keeping product according to the M test concentration data sets to obtain M test data sets, where the test data includes the test concentration and the corresponding disinfection quality. M optimal disinfection concentrations are generated based on the M test data sets.

[0037] By retrieving historical data based on big data technology to determine the disinfection concentration threshold, and setting the test concentration within the disinfection concentration threshold for disinfection testing, the disinfection concentration testing time can be reduced, the efficiency of the disinfection concentration testing can be improved, and disinfection testing resources can also be saved.

[0038] Performing disinfection tests on the first fresh-keeping product in sequence according to the M test concentration data sets to obtain M test data sets, and generating M optimal disinfection concentrations according to the M test data sets;

[0039] In one embodiment, the method further comprises:

[0040] Obtaining the size information and structural information of the disinfection space of the cold chain disinfection machine, and building a disinfection test scene based on the size information and structural information;

[0041] Acquire multiple speed adjustment parameters of the cold chain sterilizer conveyor belt, and generate multiple product sterilization times according to the multiple speed adjustment parameters and the size information;

[0042] Based on the multiple product disinfection times, in the disinfection test scenario, disinfection tests are sequentially performed on the first fresh-keeping product according to the M test concentration data sets to obtain M test data sets, wherein the test data includes concentration data and a disinfection quality coefficient;

[0043] The lowest concentration data in the M test data sets whose disinfection quality coefficient meets the preset quality coefficient threshold is used as the optimal disinfection concentration of the first fresh-keeping product, and M optimal disinfection concentrations are obtained.

[0044] In an embodiment of the present application, first, the size information and structural information of the disinfection space of the cold chain disinfection machine are obtained, and then a disinfection test scene is built based on the size information and the structural information. A plurality of speed adjustment parameters of the conveyor belt of the cold chain disinfection machine are obtained, wherein the speed adjustment parameter is an adjustment parameter at the lowest adjustable unit within the conveyor belt speed adjustment threshold. For example, assuming that the conveyor belt speed adjustment threshold is 3 to 6 meters per minute and the lowest adjustable unit is 0.1, the speed adjustment parameter is 3, 3.1, etc.

[0045] The length data of the disinfection space of the cold chain disinfection machine is determined based on the size information, and the passing time is calculated based on the length data and the multiple speed adjustment parameters to obtain multiple passing times, and the multiple passing times are used as product disinfection times to obtain multiple product disinfection times. Then, based on the multiple product disinfection times, in the disinfection test scenario, the first fresh-keeping product is disinfected in turn according to the M test concentration data sets to obtain M test data sets, wherein the test data includes product disinfection time, concentration data, and disinfection quality coefficient. The disinfection quality coefficient is obtained by detecting harmful substances such as viruses and bacteria on the surface of the disinfected fresh-keeping product through a detection instrument. The lower the content of harmful substances, the higher the disinfection quality, and the greater the disinfection quality coefficient.

[0046] Obtain a preset quality coefficient threshold, which can be set by a person skilled in the art based on the actual disinfection quality requirements of the fresh-keeping product. Then, the disinfection quality coefficients in the M test data sets are judged according to the preset quality coefficient threshold, and M qualified test data sets in the M test data sets whose disinfection quality coefficients meet the preset quality coefficient threshold are obtained. The lowest concentration data in the M qualified test data sets is used as the optimal disinfection concentration of the first fresh-keeping product, and M optimal disinfection concentrations are obtained, wherein each disinfectant type corresponds to an optimal disinfection concentration, and selecting the lowest concentration data can save the loss of disinfectant resources. By obtaining M optimal disinfection concentrations, data support is provided for the next step of constructing a disinfection analysis decision tree.

[0047] Building a disinfection analysis decision tree based on the first fresh-keeping product information and the M optimal disinfection concentrations, and embedding the disinfection analysis decision tree into a disinfection analysis unit, wherein the product identification channel is communicatively connected to the disinfection analysis unit;

[0048] In one embodiment, the method further comprises:

[0049] Based on the principle of decision tree, the first fresh-keeping product information is used as a child node, the disinfectant type is used as the first leaf node of the child node, the speed adjustment parameter is used as the second leaf node of the first leaf node, and the optimal disinfection concentration is used as the third leaf node of the second leaf node for data storage to generate the disinfection analysis decision tree.

[0050] In an embodiment of the present application, based on the principle of a decision tree, a disinfection analysis decision tree is used as a root node, the first fresh-keeping product information is used as a child node, the disinfectant type is used as the first leaf node of the child node, and the speed adjustment parameter is used as the second leaf node of the first leaf node. A mapping relationship exists between the speed adjustment parameter and the product disinfection time, and data can be stored based on the mapping relationship. The optimal disinfection concentration is used as the third leaf node of the second leaf node for data storage, thereby generating a disinfection analysis decision tree. By constructing a disinfection analysis decision tree based on the principle of a decision tree, the accuracy and efficiency of optimal disinfection concentration matching can be improved.

[0051] The disinfection analysis decision tree is then embedded in a disinfection analysis unit, wherein the input data of the disinfection analysis unit is the fresh product information, the disinfectant type and the conveyor belt transmission speed, and the output data is the optimal disinfection concentration, wherein the product identification channel transmits the product information to the disinfection analysis unit through signal transmission.

[0052] The disinfection analysis unit performs disinfection demand analysis on the product information sequence.

[0053] In an embodiment of the present application, the disinfectant type and conveyor belt transmission speed of the cold chain disinfection machine are obtained, and the disinfectant type, the transmission speed and the product information sequence are input into the disinfection analysis unit for optimal disinfection concentration matching, and the optimal disinfection concentration sequence is generated according to the time node order of the product information sequence.

[0054] Obtaining a disinfection concentration deviation sequence, wherein the disinfection concentration deviation sequence is obtained by sequentially calculating adjacent disinfection concentration deviations for the optimal disinfection concentration sequence;

[0055] In an embodiment of the present application, first, the deviation of the two adjacent optimal disinfection concentrations in the optimal disinfection concentration sequence is calculated in sequence according to the time node order, wherein the disinfection concentration deviation is the difference between the optimal disinfection concentration at the front of the sequence and the adjacent optimal disinfection concentration at the back, and the disinfection concentration deviation sequence is obtained. By generating the disinfection concentration deviation sequence, data support is provided for the next step of optimizing the spray control parameters of the cold chain disinfection machine.

[0056] Generate a spray control optimization parameter sequence for the cold chain disinfection machine based on the disinfection concentration deviation sequence;

[0057] In an embodiment of the present application, the spray control parameters of the cold chain disinfection machine are optimized and analyzed based on the disinfection concentration deviation sequence, wherein the spray control parameter refers to the spray power of the spray device, which can be expressed by the disinfectant consumption per unit time. The greater the disinfectant consumption per unit time, the greater the spray power, and the spray control optimization parameter sequence is obtained.

[0058] In one embodiment, the method further comprises:

[0059] Obtaining multiple spray adjustment parameters of the cold chain disinfection machine, and performing unit time concentration generation analysis based on the multiple spray adjustment parameters and the size information to obtain multiple unit time concentrations corresponding to the multiple spray adjustment parameters;

[0060] In an embodiment of the present application, first, a plurality of spray adjustment parameters of the cold chain disinfection machine are obtained, wherein the spray adjustment parameter refers to the lowest unit adjustment value adjustable within the spray adjustment parameter threshold. The volume data of the disinfection space is calculated based on the size information, and the unit time disinfectant concentration analysis is performed based on the plurality of spray adjustment parameters and the volume data. The disinfection concentration per unit time is calculated based on the unit time disinfectant consumption of the spray adjustment parameters and the volume of the disinfection space, wherein: in the actual working process of the cold chain disinfection machine, usually multiple disinfection spray devices work at the same time. Here, for the convenience of analysis, multiple spray devices can be fused and analyzed, that is, considering the overall working state of the spray device, multiple unit time concentrations corresponding to the multiple spray adjustment parameters are obtained, wherein the greater the spray adjustment parameter power, the greater the corresponding unit time concentration.

[0061] Performing a concentration decay test in the disinfection test scenario to generate a concentration decay curve;

[0062] In one embodiment, the method further comprises:

[0063] In the disinfection test scenario, different concentrations of disinfectant are sprayed by the spray device of the cold chain machine, and according to the preset time nodes, the concentration in the disinfection test scenario is detected by the concentration detector to obtain multiple concentration data sets, wherein the concentration data are marked with time node identifiers;

[0064] Based on the multiple concentration data sets, a decay ratio mean value is calculated for the concentration data at the same time node, and a concentration decay curve is constructed according to the mean value calculation result and the time node.

[0065] In an embodiment of the present application, first, in the disinfection test scene, disinfectant spraying of different concentrations is performed by the spray device of the cold chain machine to obtain a preset time node, and the preset time node is used to collect real-time concentration data in the disinfection test scene. The preset time node can be set by a technician in this field according to actual conditions, wherein the shorter the preset time node, the higher the accuracy of constructing the concentration decay curve, for example: setting the preset time node to 0.1 second; detecting the disinfectant concentration in the disinfection test scene by a concentration detector, and collecting the concentration detection data according to the preset time node to obtain multiple concentration data sets, wherein the concentration data set and the initial spray concentration of the disinfectant have a corresponding relationship, and the concentration data is marked with a collection time node.

[0066] Cluster the concentration data in the multiple concentration data sets according to the fading time nodes, that is, cluster the concentration data of the same fading time nodes into one category, wherein the fading time node refers to the time point of fading when the spraying stops, for example: 0.1 seconds after stopping the spraying, 0.2 seconds after stopping the spraying, etc. Obtain the concentration data sets corresponding to the multiple fading time nodes, and calculate the fading ratio of the concentration data in the concentration data sets, wherein the fading ratio refers to the ratio of the concentration data to the initial concentration data in the disinfection test scene, obtain the fading ratio sets corresponding to the multiple fading time nodes, and calculate the fading ratio mean of the multiple fading ratio sets in turn, and obtain the fading ratio mean corresponding to the multiple fading time nodes. Take the fading time node as the X-axis and the fading ratio mean as the Y-axis, and construct a two-dimensional rectangular coordinate system based on the corresponding relationship between the two, and distribute the multiple fading time nodes and the multiple fading ratio means in the two-dimensional rectangular coordinate system, fit and connect the distribution points from small to large according to the fading time nodes, and generate a concentration fading curve. By constructing the concentration decay curve, it provides support for the next step of concentration decay analysis, which can improve the efficiency and accuracy of concentration decay analysis.

[0067] Selecting the first fresh-keeping product on the conveyor belt of the sterilizer, and generating a first spray control optimization parameter at a first time node based on the first fresh-keeping product;

[0068] In one embodiment, the method further comprises:

[0069] Obtaining the optimal disinfection concentration and first spacing distance of the first fresh product;

[0070] generating a preset disinfection time based on the first interval distance and the conveyor belt speed;

[0071] The spray control parameters are calculated based on the optimal disinfection concentration and the preset disinfection time to generate the first spray control optimization parameters.

[0072] In an embodiment of the present application, first, the first fresh-keeping product on the conveyor belt of the sterilizer is selected, wherein the first fresh-keeping product is the first fresh-keeping product on the conveyor belt of the sterilizer, and the optimal disinfection concentration of the first fresh-keeping product is obtained based on the product identification channel and the disinfection analysis unit. Then, a first interval distance is obtained based on the distance between the CCD image sensor and the front end of the disinfection area of the cold chain sterilizer. A preset disinfection time is calculated based on the first interval distance and the conveyor belt speed, and then the disinfectant consumption per unit time, i.e., the first spray control optimization parameter, is calculated by dividing the optimal disinfection concentration by the preset disinfection time.

[0073] Obtaining the interval between adjacent fresh-keeping products on the conveyor belt, and calculating the adjacent interval duration based on the interval distance and the current conveyor belt speed;

[0074] Performing concentration prediction based on the adjacent interval duration and the concentration decay curve to generate a disinfection concentration prediction result for the next fresh-keeping product, wherein the next fresh-keeping product is marked with a time node;

[0075] When the disinfection concentration prediction result is greater than or equal to the optimal disinfection concentration of the next fresh product, the spray device of the cold chain disinfection machine is controlled to stop spraying until the next adjacent spray instruction is generated;

[0076] When the disinfection concentration prediction result is less than the optimal disinfection concentration of the next fresh-keeping product, a disinfection concentration deviation is calculated based on the disinfection concentration prediction result and the optimal disinfection concentration, and a spray optimization control parameter is generated based on the disinfection concentration deviation;

[0077] Connect the first spray control optimization parameter and the spray optimization control parameter in the order of time nodes to generate the spray control optimization parameter sequence.

[0078] In an embodiment of the present application, the interval distance between adjacent fresh-keeping products on a conveyor belt is obtained, wherein the adjacent fresh-keeping products are any group of adjacent fresh-keeping products on the conveyor belt, and the interval distance can be obtained by analyzing the image acquisition data of the fresh-keeping products, and then the adjacent interval duration is calculated based on the interval distance divided by the current conveyor belt speed, and the optimal disinfection concentration of the fresh-keeping product with the preceding sequence among the adjacent fresh-keeping products is obtained, and then the adjacent interval duration is input into the concentration decay curve to match and obtain the decay ratio, and the product of the optimal disinfection concentration of the fresh-keeping product with the preceding sequence and the decay ratio is used as the disinfection concentration prediction result of the next fresh-keeping product, wherein the next fresh-keeping product is marked with a time acquisition node.

[0079] Obtain the optimal disinfection concentration of the next fresh-keeping product, and then compare the disinfection concentration prediction result with the optimal disinfection concentration of the next fresh-keeping product. When the disinfection concentration prediction result is greater than or equal to the optimal disinfection concentration of the next fresh-keeping product, it indicates that the disinfection concentration requirement of the next fresh-keeping product can be met without continuing to spray. The spray device of the cold chain disinfection machine is controlled to stop spraying until the next adjacent spray instruction is generated, wherein the next adjacent spray instruction is obtained by comparing the optimal disinfection concentrations of the subsequent adjacent fresh-keeping products of the next fresh-keeping product.

[0080] If the predicted disinfection concentration is less than the optimal disinfection concentration for the next fresh-keeping product, the predicted disinfection concentration is subtracted from the optimal disinfection concentration to obtain a disinfection concentration deviation. Spray control parameter analysis is then performed based on the disinfection concentration deviation to generate optimized spray control parameters. Finally, the first and second optimized spray control parameters are concatenated in the order of image acquisition time nodes to generate a spray control optimization parameter sequence. By generating this spray control optimization parameter sequence, the output power of the spray device can be optimized based on the type of fresh-keeping product, thereby reducing the waste of disinfectant resources.

[0081] The disinfection treatment of the fresh-keeping products to be disinfected is performed according to the spray control optimization parameter sequence.

[0082] In an embodiment of the present application, the spray device of the cold chain disinfection machine is finally controlled to disinfect the fresh-keeping products to be disinfected according to the spray control optimization parameter sequence. The above method can solve the technical problem of waste of disinfectant resources caused by the use of fixed spray power for disinfection spray when the existing cold chain disinfection machine disinfects fresh-keeping products. It can save the consumption of disinfectant resources when the cold chain disinfection machine is working and reduce resource waste.

[0083] In one embodiment, Figure 3 The figure provides a control system for a disinfection machine based on cold chain fresh products, including: an image data sequence acquisition module 11, a product information sequence generation module 12, an optimal disinfection concentration sequence acquisition module 13, a disinfection concentration deviation sequence acquisition module 14, a spray control optimization parameter sequence generation module 15, and a fresh product disinfection treatment module 16, wherein:

[0084] An image data sequence acquisition module 11 is used to acquire an image data sequence of the fresh-keeping product to be sterilized. The image data sequence is acquired by sequentially capturing images of the fresh-keeping product to be sterilized using a CCD image sensor;

[0085] A product information sequence generation module 12, the product information sequence generation module 12 is used to generate a product information sequence, the product information sequence is obtained by performing product recognition on the image data sequence through a product recognition channel;

[0086] An optimal disinfection concentration sequence obtaining module 13 is used to obtain an optimal disinfection concentration sequence, wherein the optimal disinfection concentration sequence is generated by performing a disinfection demand analysis on the product information sequence;

[0087] A disinfection concentration deviation sequence obtaining module 14 is used to obtain a disinfection concentration deviation sequence, wherein the disinfection concentration deviation sequence is obtained by sequentially calculating adjacent disinfection concentration deviations for the optimal disinfection concentration sequence;

[0088] A spray control optimization parameter sequence generation module 15, which is used to generate a spray control optimization parameter sequence for the cold chain disinfection machine based on the disinfection concentration deviation sequence;

[0089] The fresh food disinfection processing module 16 is used to perform disinfection processing on the fresh food to be disinfected according to the spray control optimization parameter sequence.

[0090] In one embodiment, the system further comprises:

[0091] An information retrieval module, the information retrieval module is used to connect to the cold chain warehouse management system and retrieve a plurality of fresh product image data and a plurality of fresh product product information, wherein the fresh product image data and the fresh product product information have a corresponding relationship;

[0092] a sample training data set construction module, the sample training data set construction module being configured to construct a sample training data set based on the plurality of fresh-keeping product image data and the plurality of fresh-keeping product product information;

[0093] A product identification channel acquisition module is used to construct a product identification channel based on a convolutional neural network, and to perform supervised training and verification training on the product identification channel through the sample training data set to obtain a product identification channel that meets the expected requirements.

[0094] In one embodiment, the system further comprises:

[0095] a disinfection data retrieval module, the disinfection data retrieval module being configured to perform disinfection data retrieval using the plurality of fresh-keeping product information as retrieval conditions, and obtain a plurality of sets of historical qualified disinfection information corresponding to the plurality of fresh-keeping product information, wherein the historical qualified disinfection information includes a disinfectant type and a disinfectant concentration;

[0096] a first fresh-keeping product information selection module, configured to randomly select a first fresh-keeping product information from the plurality of fresh-keeping product information and obtain a first set of historical qualified disinfection information of the first fresh-keeping product information;

[0097] a disinfectant concentration data set obtaining module, the disinfectant concentration data set obtaining module being configured to cluster the first historical qualified disinfection information based on disinfectant type to obtain M disinfectant concentration data sets, where M is the number of disinfectant types;

[0098] a test concentration data set determination module, the test concentration data set determination module being configured to determine M disinfection concentration thresholds based on the M disinfectant concentration data sets, and to determine M test concentration data sets based on a preset test step size and the M disinfection concentration thresholds;

[0099] an optimal disinfection concentration generating module, the optimal disinfection concentration generating module being configured to sequentially perform disinfection tests on the first fresh-keeping product according to the M test concentration data sets to obtain M test data sets, and to generate M optimal disinfection concentrations based on the M test data sets;

[0100] a disinfection analysis decision tree construction module, the disinfection analysis decision tree construction module being configured to construct a disinfection analysis decision tree based on the first fresh-keeping product information and the M optimal disinfection concentrations, and embed the disinfection analysis decision tree into a disinfection analysis unit, wherein the product identification channel is communicatively connected to the disinfection analysis unit;

[0101] A disinfection demand analysis module is used to perform disinfection demand analysis on the product information sequence through the disinfection analysis unit.

[0102] In one embodiment, the system further comprises:

[0103] A disinfection test scene building module, which is used to obtain the size information and structural information of the disinfection space of the cold chain disinfection machine and build a disinfection test scene based on the size information and structural information;

[0104] a product disinfection time generation module, the product disinfection time generation module being configured to obtain a plurality of speed adjustment parameters of the cold chain disinfection machine conveyor belt, and generate a plurality of product disinfection times according to the plurality of speed adjustment parameters and the size information;

[0105] a test data set obtaining module, the test data set obtaining module being configured to perform a disinfection test on the first fresh-keeping product in sequence according to the M test concentration data sets within the disinfection test scenario based on the multiple product disinfection times to obtain M test data sets, wherein the test data includes concentration data and a disinfection quality coefficient;

[0106] The optimal disinfection concentration obtaining module is used to take the lowest concentration data in the M test data sets whose disinfection quality coefficient meets the preset quality coefficient threshold as the optimal disinfection concentration of the first fresh-keeping product, and obtain M optimal disinfection concentrations.

[0107] In one embodiment, the system further comprises:

[0108] A disinfection analysis decision tree generation module is used to generate the disinfection analysis decision tree based on the principle of decision tree, with the first fresh-keeping product information as a child node, the disinfectant type as the first leaf node of the child node, the speed adjustment parameter as the second leaf node of the first leaf node, and the optimal disinfection concentration as the third leaf node of the second leaf node for data storage.

[0109] In one embodiment, the system further comprises:

[0110] a unit time concentration acquisition module, which is used to obtain multiple spray adjustment parameters of the cold chain disinfection machine, and perform unit time concentration generation analysis based on the multiple spray adjustment parameters and the size information to obtain multiple unit time concentrations corresponding to the multiple spray adjustment parameters;

[0111] A concentration decay curve generation module, the concentration decay curve generation module is used to perform a concentration decay test in the disinfection test scene and generate a concentration decay curve;

[0112] a first spray control optimization parameter generation module, the first spray control optimization parameter generation module being used to select a first fresh-keeping product on the conveyor belt of the sterilizer and generate a first spray control optimization parameter at a first time node based on the first fresh-keeping product;

[0113] An adjacent interval duration calculation module is used to obtain the interval distance between adjacent fresh-keeping products on the conveyor belt, and calculate the adjacent interval duration based on the interval distance and the current conveyor belt speed;

[0114] A disinfection concentration prediction result generation module, the disinfection concentration prediction result generation module is used to perform concentration prediction based on the adjacent interval duration and the concentration decay curve, and generate a disinfection concentration prediction result for the next fresh-keeping product, wherein the next fresh-keeping product is marked with a time node;

[0115] A spray device control module, wherein the spray device control module is used to control the spray device of the cold chain disinfection machine to stop spraying when the disinfection concentration prediction result is greater than or equal to the optimal disinfection concentration of the next fresh product, until the next adjacent spray instruction is generated;

[0116] a spray optimization control parameter generation module, wherein when the disinfection concentration prediction result is less than the optimal disinfection concentration of the next fresh-keeping product, the spray optimization control parameter generation module is used to calculate a disinfection concentration deviation based on the disinfection concentration prediction result and the optimal disinfection concentration, and generate a spray optimization control parameter based on the disinfection concentration deviation;

[0117] A spray control optimization parameter sequence generation module is used to connect the first spray control optimization parameter and the spray optimization control parameter in the order of time nodes to generate the spray control optimization parameter sequence.

[0118] In one embodiment, the system further comprises:

[0119] A concentration data set acquisition module, wherein the concentration data set acquisition module is used to spray disinfectants of different concentrations through the spray device of the cold chain machine in the disinfection test scene, and perform concentration detection in the disinfection test scene through a concentration detector according to preset time nodes to obtain multiple concentration data sets, wherein the concentration data are marked with time node identifiers;

[0120] A concentration decay curve construction module is used to calculate the decay ratio mean of the concentration data at the same time node based on the multiple concentration data sets, and to construct a concentration decay curve according to the mean calculation result and the time node.

[0121] In one embodiment, the system further comprises:

[0122] An information acquisition module, configured to acquire an optimal disinfection concentration and a first spacing distance for the first fresh-keeping product;

[0123] a preset disinfection time generation module, the preset disinfection time generation module being configured to generate a preset disinfection time based on the first interval distance and the conveyor belt speed;

[0124] The first spray control optimization parameter generation module is used to calculate the spray control parameters based on the optimal disinfection concentration and the preset disinfection time to generate the first spray control optimization parameters.

[0125] In summary, compared with the prior art, the embodiments of the present disclosure have the following technical effects:

[0126] (1) It can solve the technical problem of waste of disinfectant resources caused by the use of fixed spray power for disinfection spray when the existing cold chain disinfection machine is used to disinfect fresh products. By generating a spray control optimization parameter sequence, the output power of the spray device can be optimized and controlled according to the type of fresh products, thereby saving the consumption of disinfectant resources when the cold chain disinfection machine is working and reducing resource waste.

[0127] (2) By constructing a product recognition channel based on a convolutional neural network, the efficiency and accuracy of fresh product identification can be improved, while providing support for obtaining the optimal disinfection concentration sequence in the next step.

[0128] (3) By constructing a disinfection analysis decision tree based on the decision tree principle, the accuracy and efficiency of the optimal disinfection concentration matching can be improved; by constructing a concentration decay curve, support is provided for the next step of concentration decay analysis, which can improve the efficiency and accuracy of concentration decay analysis.

[0129] The above-described embodiments merely represent several implementation methods of the present disclosure and should not be construed as limiting the scope of the present invention. Therefore, various substitutions, modifications, and variations may be made by those skilled in the art without departing from the scope of the present disclosure as defined by the appended claims, and such substitutions, modifications, and variations are all within the scope of protection of the present disclosure.

Claims

1. A control method for a sterilizer for cold chain fresh products, characterized in that: The method comprises: Acquire an image data sequence of the fresh-keeping product to be sterilized, wherein the image data sequence is obtained by sequentially capturing images of the fresh-keeping product to be sterilized using a CCD image sensor; generating a product information sequence, wherein the product information sequence is obtained by performing product identification on the image data sequence through a product identification channel; Obtaining an optimal disinfection concentration sequence, wherein the optimal disinfection concentration sequence is generated by performing a disinfection demand analysis on the product information sequence; Obtaining a disinfection concentration deviation sequence, wherein the disinfection concentration deviation sequence is obtained by sequentially calculating adjacent disinfection concentration deviations for the optimal disinfection concentration sequence; Generate a spray control optimization parameter sequence for the cold chain disinfection machine based on the disinfection concentration deviation sequence; Executing the disinfection process of the fresh-keeping products to be disinfected according to the spray control optimization parameter sequence; The step of generating a spray control optimization parameter sequence for the cold chain disinfection machine based on the disinfection concentration deviation sequence further includes: Obtaining multiple spray adjustment parameters of the cold chain disinfection machine, and performing unit time concentration generation analysis based on the multiple spray adjustment parameters and size information to obtain multiple unit time concentrations corresponding to the multiple spray adjustment parameters; Conduct concentration decay tests in the disinfection test scenario and generate concentration decay curves; Selecting the first fresh-keeping product on the conveyor belt of the sterilizer, and generating a first spray control optimization parameter at a first time node based on the first fresh-keeping product; Obtaining the interval between adjacent fresh-keeping products on the conveyor belt, and calculating the adjacent interval duration based on the interval distance and the current conveyor belt speed; Performing concentration prediction based on the adjacent interval duration and the concentration decay curve to generate a disinfection concentration prediction result for the next fresh-keeping product, wherein the next fresh-keeping product is marked with a time node; When the disinfection concentration prediction result is greater than or equal to the optimal disinfection concentration of the next fresh product, the spray device of the cold chain disinfection machine is controlled to stop spraying until the next adjacent spray instruction is generated; When the disinfection concentration prediction result is less than the optimal disinfection concentration of the next fresh-keeping product, a disinfection concentration deviation is calculated based on the disinfection concentration prediction result and the optimal disinfection concentration, and a spray optimization control parameter is generated based on the disinfection concentration deviation; Connecting the first spray control optimization parameter and the spray optimization control parameter in the order of time nodes to generate the spray control optimization parameter sequence; The method further comprises: generating a first spray control optimization parameter at a first time node according to the first fresh-keeping product; Obtaining the optimal disinfection concentration and first spacing distance of the first fresh product; generating a preset disinfection time based on the first interval distance and the conveyor belt speed; The spray control parameters are calculated based on the optimal disinfection concentration and the preset disinfection time to generate the first spray control optimization parameters.

2. The method according to claim 1, wherein The method further comprises: Connecting to a cold chain warehouse management system to retrieve a plurality of fresh-keeping product image data and a plurality of fresh-keeping product product information, wherein the fresh-keeping product image data and the fresh-keeping product product information have a corresponding relationship; Constructing a sample training data set based on the plurality of fresh food image data and the plurality of fresh food product information; A product recognition channel is constructed based on a convolutional neural network, and supervised training and verification training are performed on the product recognition channel using the sample training data set to obtain a product recognition channel that meets expected requirements.

3. The method according to claim 2, wherein The method further comprises: Perform disinfection data retrieval using the plurality of fresh-keeping product information as retrieval conditions to obtain a plurality of historical qualified disinfection information sets corresponding to the plurality of fresh-keeping product information, wherein the historical qualified disinfection information includes a disinfectant type and a disinfectant concentration; Randomly selecting a first piece of fresh food product information from the plurality of pieces of fresh food product information, and obtaining a first set of historical qualified disinfection information of the first fresh food product information; Clustering the first historical qualified disinfection information based on disinfectant type to obtain M disinfectant concentration data sets, where M is the number of disinfectant types; Determining M disinfection concentration thresholds based on the M disinfectant concentration data sets, and determining M test concentration data sets based on a preset test step length and the M disinfection concentration thresholds; Performing disinfection tests on the first fresh-keeping product in sequence according to the M test concentration data sets to obtain M test data sets, and generating M optimal disinfection concentrations according to the M test data sets; Building a disinfection analysis decision tree based on the first fresh product information and the M optimal disinfection concentrations, and embedding the disinfection analysis decision tree into a disinfection analysis unit, wherein the product identification channel is communicatively connected to the disinfection analysis unit; The disinfection analysis unit performs disinfection demand analysis on the product information sequence.

4. The method according to claim 3, wherein The method of sequentially performing disinfection tests on the first fresh-keeping product according to the M test concentration data sets to obtain M test data sets, and generating M optimal disinfection concentrations based on the M test data sets, further includes: Obtaining the size information and structural information of the disinfection space of the cold chain disinfection machine, and building a disinfection test scene based on the size information and structural information; Acquire multiple speed adjustment parameters of the cold chain sterilizer conveyor belt, and generate multiple product sterilization times according to the multiple speed adjustment parameters and the size information; Based on the multiple product disinfection times, in the disinfection test scenario, disinfection tests are sequentially performed on the first fresh-keeping product according to the M test concentration data sets to obtain M test data sets, wherein the test data includes concentration data and a disinfection quality coefficient; The lowest concentration data in the M test data sets whose disinfection quality coefficient meets the preset quality coefficient threshold is used as the optimal disinfection concentration of the first fresh-keeping product, and M optimal disinfection concentrations are obtained.

5. The method according to claim 4, wherein The step of constructing a disinfection analysis decision tree based on the first fresh product information and the M optimal disinfection concentrations further includes: Based on the principle of decision tree, the first fresh product information is used as a child node, the disinfectant type is used as the first leaf node of the child node, the speed adjustment parameter is used as the second leaf node of the first leaf node, and the optimal disinfection concentration is used as the third leaf node of the second leaf node for data storage to generate the disinfection analysis decision tree.

6. The method according to claim 4, wherein The method of performing a concentration decay test in the disinfection test scenario to generate a concentration decay curve further includes: In the disinfection test scenario, different concentrations of disinfectant are sprayed by the spray device of the cold chain machine, and according to the preset time nodes, the concentration in the disinfection test scenario is detected by the concentration detector to obtain multiple concentration data sets, wherein the concentration data are marked with time node identifiers; Based on the multiple concentration data sets, a decay ratio mean value is calculated for the concentration data at the same time node, and a concentration decay curve is constructed according to the mean value calculation result and the time node.

7. The control system of the sterilizer based on cold chain fresh products is characterized by: The system is used to perform the steps of any one of the methods for controlling a sterilizer based on cold chain fresh products described in claims 1-6, comprising: An image data sequence acquisition module is used to acquire an image data sequence of the fresh-keeping product to be sterilized, wherein the image data sequence is acquired by sequentially capturing images of the fresh-keeping product to be sterilized using a CCD image sensor; A product information sequence generation module, the product information sequence generation module is used to generate a product information sequence, the product information sequence is obtained by performing product recognition on the image data sequence through a product recognition channel; An optimal disinfection concentration sequence obtaining module, wherein the optimal disinfection concentration sequence obtaining module is used to obtain an optimal disinfection concentration sequence, wherein the optimal disinfection concentration sequence is generated by performing a disinfection demand analysis on the product information sequence; A disinfection concentration deviation sequence obtaining module, wherein the disinfection concentration deviation sequence obtaining module is used to obtain a disinfection concentration deviation sequence, wherein the disinfection concentration deviation sequence is obtained by sequentially calculating adjacent disinfection concentration deviations for the optimal disinfection concentration sequence; A spray control optimization parameter sequence generation module, which is used to generate a spray control optimization parameter sequence for the cold chain disinfection machine based on the disinfection concentration deviation sequence; A fresh food disinfection processing module is used to perform disinfection processing on the fresh food to be disinfected according to the spray control optimization parameter sequence.

Images