Dry product storage quality intelligent supervision system and method based on AI

Abstract

The invention relates to the technical field of dry product storage supervision, and discloses an AI-based dry product storage quality intelligent supervision system and method, and the system comprises a dry product information management module, a dry product storage quality recognition module, and a dry product storage abnormity management module. Searching and processing a target dry product storage deterioration image library by combining different types of dry product storage deterioration image libraries stored based on big data with an intelligent search algorithm and dry product type identification parameters, so as to realize personalized accurate analysis of the storage quality of the dry products; according to the dry product storage live image parameters, the intelligent identification algorithm and the target dry product storage deterioration image library are combined to carry out dry product storage quality intelligent and efficient evaluation, so that the storage quality state of the stored dry product is scientifically and accurately analyzed, and the efficiency and quality of dry product storage quality supervision are improved. And the workload of warehouse management of dry goods and the deterioration probability of the dry goods are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of dry goods product storage supervision, and specifically to an intelligent supervision system and method for the storage quality of dry goods products based on AI. Background Art

[0002] The storage of dried agricultural products refers to the process of storing and preserving dried agricultural products in a warehouse. As an important part of the logistics of agricultural products, the storage of dried goods has the following characteristics: storage environment requirements: the temperature in the dry warehouse is generally between 5°C and 20°C, and the humidity must be maintained at 70%. It is mainly used for storing grains, oils, noodles, and dried foods, etc. The dry warehouse must also maintain a well-ventilated and naturally dry environment. Food cannot be directly exposed to sunlight, and the distance between the goods and the walls, the ground, and the ceiling should be at least 15 cm. Warehouse management: The storage management of agricultural products includes multiple aspects such as the warehousing, outwarehousing, inventory control, and maintenance of the storage environment of agricultural products. Reasonable warehouse management can ensure the quality and quantity of agricultural products, reduce losses, and improve logistics efficiency. Processing and sorting: In some cases, the storage of dried goods also undertakes the functions of processing and sorting. For example, selecting, sorting, processing, and packaging longan dried fruits, bird's nests, fish maws, etc. to meet market demands. During the storage process of dry goods products, it is usually necessary for warehouse management personnel to regularly and quantitatively inspect the dry goods products and record the information of the dry goods products with storage deterioration problems. The existing storage quality supervision of dry goods products cannot intelligently and efficiently detect the storage quality of dry goods products, nor can it digitally feedback and supervise the deteriorated dry goods products in storage, reducing the efficiency and reliability of the storage quality supervision of dry goods products.

[0003] The Chinese patent application for invention with the publication number CN119106962A discloses a quality supervision system for the agricultural product supply chain based on blockchain. By setting up a management unit, a contract management unit, a quality management unit, a supervision management unit, and a quality traceability unit, it respectively realizes the management of the blockchain system; provides intelligent contract services to collect and supervise the factors affecting product quality in the entire supply chain of branded agricultural products; realizes the transparency and integrity of the standardized production of agricultural products, and enhances the public credibility of the quality and brand of agricultural products. However, the above technical solutions cannot intelligently supervise the storage quality during the storage process of agricultural products, reducing the applicability of the quality supervision system for the agricultural product supply chain. Summary of the Invention

[0004] (I) Technical Problems to be Solved

[0005] To solve the problem that the existing warehousing quality supervision of dry goods products cannot achieve intelligent and efficient detection of the warehousing quality of dry goods products, nor can it digitally feedback and supervise the deteriorated dry goods products in storage, which reduces the efficiency and reliability of the warehousing quality supervision of dry goods products, and to achieve the purpose of accurately collecting the actual situation images of dry goods products in storage and the image information of dry goods product storage labels, intelligently identifying the information of dry goods product types, accurately collecting the record information of dry goods product types, scientifically judging the consistency of dry goods product information, efficiently feedbacking the abnormal information of dry goods product type records, accurately searching the image database of deteriorated dry goods products in storage, intelligently analyzing the warehousing quality status of dry goods products, and efficiently and dynamically executing the cleaning operation for deteriorated dry goods products in storage.

[0006] (II)Technical Solution

[0007] The present invention is achieved through the following technical solutions: An AI-based intelligent supervision method for the storage quality of dry goods products, the method comprising the following steps:

[0008] S1. Collect the actual situation image data of dry goods products in storage and the image data of dry goods product storage labels;

[0009] S2. Perform type recognition processing on the dry goods products in storage according to the actual situation image data of dry goods products in storage and the image data of different types of dry goods products to generate dry goods product type recognition data;

[0010] S3. Perform text information collection processing on the dry goods product storage label records according to the image data of dry goods product storage labels to generate dry goods product storage label record text data, and perform consistency judgment processing on the type recognition information and storage label record information of dry goods products with the dry goods product type recognition data to generate dry goods product storage record information consistency judgment data; when they are different, construct dry goods product type storage record abnormal data and execute the feedback operation of dry goods product type record abnormal information;

[0011] S4. When the dry goods product storage record information consistency judgment data is the same or when the feedback operation of dry goods product type record abnormal information is completed, perform a search process on the image database of deteriorated dry goods products in storage required for specific dry goods product storage quality analysis based on the dry goods product type recognition data and different types of dry goods product storage deteriorated image libraries to generate a target dry goods product storage deteriorated image library;

[0012] S5. Perform dry goods product storage quality analysis processing according to the actual situation image data of dry goods products in storage and the target dry goods product storage deteriorated image library to construct target dry goods product storage quality analysis data;

[0013] S6. Search and process the dry product storage quality deterioration information based on the target dry product storage quality analysis data, generate the dry product storage quality deterioration analysis data, perform the judgment process on the dry product storage quality deterioration status, and generate the dry product storage quality deterioration status judgment data. When it is determined that there is no such situation, directly end the current dry product storage quality supervision operation;

[0014] S7. When it exists, construct the dry product storage deterioration management data and execute the dry product storage deterioration cleaning operation.

[0015] Preferably, the operation steps for collecting the dry product storage actual situation image data and the dry product storage label image data are as follows:

[0016] S11. Online collect the on-site image information of the dry product to be measured in the unpacked state in the storage plant through the cloud camera, and generate the dry product storage actual situation image data.

[0017] Online collect the product information record label image information of the dry product to be measured on the storage shelf in the storage plant through the cloud camera, and generate the dry product storage label image data.

[0018] Preferably, the operation steps for identifying the type of dry product in storage by comparing the dry product storage actual situation image data with the image data of different types of dry products and generating the dry product type identification data are as follows:

[0019] S21. Establish a set of image data of different types of dry products , ; where represents the image data of different types of dry products corresponding to the th type of dry product, represents the maximum value of the number of dry product types; the dry product types include chestnuts, Chinese chestnuts, pili nuts, hazelnuts, cashews, walnuts, melon seeds, pine nuts, almonds, ginkgo nuts, pistachios, pecans, desert nuts, torreya nuts, white melon seeds, pumpkin seeds, peanuts, almonds, macadamia nuts, agaric, laver, shiitake mushrooms, red dates, cinnamon, chili peppers, Chinese prickly ash, star anise, fennel, pepper, wolfberries, longans, peanuts, dried tangerine peels, raisins, dried green vegetables, dried cowpeas, dried string beans, dried eggplants, dried lettuce, dried pakchoi, dried cucumbers, dried eggplants, preserved mustard greens, dried daylilies, dried bamboo shoots, and dried radishes, sorghum, millet, buckwheat, sweet buckwheat, bitter buckwheat, oats, naked oats, barley, broomcorn millet, millet, coix seeds, amaranth, kidney beans, mung beans, adzuki beans, red beans, broad beans, peas, cowpeas, lentils, and black beans. The image data of different types of dry products represents the product image data corresponding to different types of dry products;

[0020] S22. Use the K-D tree nearest neighbor search algorithm to compare the dry product storage actual situation image data with the set of image data of different types of dry goods products the image data of different types of dry goods products in perform image feature matching to search for the image data of different types of dry goods products that match the actual storage image data of the dry goods products the corresponding text information of the dry goods product type, and generate dry goods product type recognition data .

[0021] Preferably, perform text information acquisition and processing on the dry goods product storage label record according to the dry goods product storage label image data, generate dry goods product storage label record text data, and perform consistency judgment processing on the type recognition information and storage label record information of the dry goods product with the dry goods product type recognition data to generate dry goods product storage record information consistency judgment data; when they are different, construct dry goods product type storage record abnormal data and execute the following operation steps of the dry goods product type record abnormal information feedback operation:

[0022] S31. Use an optical character recognition software to collect and process the text information of the dry goods product storage label record in the dry goods product storage label image data and generate dry goods product storage label record text data , the optical character recognition software includes any one of AWS Text and Adobe Acrobat DC, and the dry goods product storage label record text data includes the name text data, specification quantity text data, production date text data, and storage shelf location text data of the dry goods product;

[0023] S32. Perform keyword matching of the dry goods product type between the dry goods product type recognition data and the dry goods product storage label record text data to generate dry goods product storage record information consistency judgment data ;

[0024] When and perform successful keyword matching of the dry goods product type, indicating that the dry goods product type recognition information is consistent with the dry goods product type information recorded in the label, then output the dry goods product storage record information consistency judgment data as the same;

[0025] When and perform unsuccessful keyword matching of the dry goods product type, indicating that the dry goods product type recognition information is inconsistent with the dry goods product type information recorded in the label, then output the dry goods product storage record information consistency judgment data​ are different; at this time, the dry product type identification data , the dry product storage label record text data , the consistency judgment data of the dry product storage record information are combined to construct abnormal data of the dry product type storage record , and the abnormal data of the dry product type storage record is pushed online through the Internet of Things communication network to the dry product storage quality supervision platform to execute the feedback operation of abnormal information of the dry product type record.

[0026] Preferably, when the consistency judgment data of the dry product storage record information is the same or when the feedback operation of the abnormal information of the dry product type record is completed, based on the dry product type identification data and the different types of dry product storage deterioration image libraries, the search process of the dry product storage deterioration image library required for the specific dry product storage quality analysis is as follows:

[0027] S41. When the consistency judgment data of the dry product storage record information is the same or when the feedback operation of the abnormal information of the dry product type record is completed, a set of different types of dry product storage deterioration image libraries is established , where represents the different types of dry product storage deterioration image library corresponding to the th type of dry product, , where represents the th image data of different types of dry product storage deterioration in the different types of dry product storage deterioration image library, represents the th image data of different types of dry product storage deterioration in the different types of dry product storage deterioration image library; the different types of dry product storage deterioration image data represent the corresponding dry product appearance image data after different types of dry products deteriorate during storage; the dry product storage deterioration includes black spots, mildew spots, holes and appearance color changes on the dry products;

[0028] S42. The width-first search algorithm is used to match the dry product type identification data with the different types of dry product storage deterioration image libraries in the set of different types of dry product storage deterioration image libraries to perform character matching of the dry product type, and search out the dry product type identification data ​The corresponding image library of different types of dry goods products in storage deterioration and generate a target image library of dry goods products in storage deterioration through data identification wherein represents the target image library of dry goods products in storage deterioration the th target dry goods product storage deterioration image data in the library represents the target image library of dry goods products in storage deterioration the th target dry goods product storage deterioration image data in the library; the target dry goods product storage deterioration image data represents the corresponding dry goods product appearance image data after the dry goods product type to be measured deteriorates during storage.

[0029] Preferably, the operation steps of performing quality analysis and processing on the storage of dry goods products based on the actual image data of the storage of dry goods products and the target image library of dry goods products in storage deterioration to construct the target quality analysis data of the storage of dry goods products are as follows:

[0030] S51. Obtain the actual image data of the storage of dry goods products and the target image library of dry goods products in storage deterioration ;

[0031] S52. Match the actual image data of the storage of dry goods products with the target dry goods product storage deterioration image data in the target image library of dry goods products in storage deterioration to perform image feature matching, and generate target quality analysis data of the storage of dry goods products based on the image feature matching results ; The specific operation steps for generating the target quality analysis data of the storage of dry goods products are as follows:

[0032] S521. Initialize, define relevant structure parameters as vectors, and in the search space of the target image library of dry goods products in storage deterioration the target dry goods product storage deterioration image data to constitute dimensional optimization problem, where the dry goods product storage deterioration recognition sand cat represents a 1× array of the solution of the problem, and each variable value is a floating-point number, and each variable value to represents being between the lower bound and the upper bound in the search space of the target image library of dry goods products in storage deterioration ;

[0033] S522. Search for prey. The final parameters and main parameters for controlling the transition between the exploration and exploitation phases are , when > 1, the sand cat for identifying spoilage in dry - goods storage searches in the search space of the target image library for spoilage in dry - goods storage for the target image data of spoilage in dry - goods storage that matches the actual situation image data of dry - goods storage . The search process of the sand cat for identifying spoilage in dry - goods storage depends on the release of low - frequency noise to search for the target image data of spoilage in dry - goods storage that matches the actual situation image data of dry - goods storage . Assuming that the sensitivity range of the sand cat for identifying spoilage in dry - goods storage is from 0 to 2 kHz, which represents being inspired by the auditory characteristics of the sand cat for identifying spoilage in dry - goods storage, and assuming its value is 2, is the current iteration number, is the maximum iteration number, and rand(0, 1) represents a random number with a value ranging from 0 to 1; , , where represents the sensitivity vector; each sand cat for identifying spoilage in dry - goods storage updates its position according to the best candidate position and the current position and its sensitivity range , that is, updates the position of the target image data of spoilage in dry - goods storage that matches the actual situation image data of dry - goods storage in the search space of the target image library for spoilage in dry - goods storage . The position calculation formula is as follows: , where represents the current position of the individual sand cat for identifying spoilage in dry - goods storage in the search space of the target image library for spoilage in dry - goods storage at the (t + 1) - th iteration, represents the best candidate position of the sand cat for identifying spoilage in dry - goods storage in the search space of the target image library for spoilage in dry - goods storage at the t - th iteration, which is the most matching target image data of spoilage in dry - goods storage for the actual situation image data of dry - goods storage ; represents the position of the sand cat for identifying spoilage in dry - goods storage in the search space of the target image library for spoilage in dry - goods storage at the t - th iteration, searching for the target image data of spoilage in dry - goods storage that matches the actual situation image data of dry - goods storage ;

[0034] S523. Attack prey. When ≤1, the sand cat for identifying spoilage in dry - goods storage searches in the search space of the target image library for spoilage in dry - goods storage , and uses the best candidate position and the current position ​Generate a random position, that is, in the target dry goods product storage deterioration image library Randomly search in the search space for the target dry goods product storage deterioration image data that matches the actual dry goods product storage image data The random position of the target dry goods product storage deterioration image data that matches. Assume that the sensitivity range of the dry goods product storage deterioration recognition sand cat is a circle, and use the roulette method to randomly select an angle for each dry goods product storage deterioration recognition sand cat , according to the random position calculation formula: , perform a random position search for the target dry goods product storage deterioration image data that matches the actual dry goods product storage image data , where represents the sensitivity vector, represents the angle the cosine value of represents the random position of the dry goods product storage deterioration recognition sand cat in the target dry goods product storage deterioration image library search space. The random position enables the dry goods product storage deterioration recognition sand cat to approach and attack the prey, that is, in the target dry goods product storage deterioration image library search space to search for the target dry goods product storage deterioration image data that matches the actual dry goods product storage image data ;

[0035] S524. When the search algorithm reaches the maximum number of iterations T, output the target dry goods product storage deterioration image data that matches the actual dry goods product storage image data ;

[0036] S525. Generate target dry goods product storage quality analysis data based on the image feature matching result of the actual dry goods product storage image data output in step S524 and the target dry goods product storage deterioration image data ;

[0037] When matches or successfully in image feature matching, it indicates that the dry goods have deteriorated during storage, then it means that the output target dry goods product storage quality analysis data is deteriorated;

[0038] When and or do not match successfully in image features, it indicates that the dry goods have not deteriorated during storage, then it means that the output target dry goods product storage quality analysis data is normal.

[0039] Preferably, based on the target dry goods product storage quality analysis data, search and process the dry goods product storage quality deterioration information, generate the dry goods product storage quality deterioration analysis data, and perform the dry goods product storage quality deterioration status judgment process, and generate the dry goods product storage quality deterioration status judgment data. When it does not exist, the operation steps to directly end the current dry goods product storage quality supervision operation are as follows:

[0040] S61. Use the breadth-first search algorithm to search and process the dry goods product storage quality deterioration information in the target dry goods product storage quality analysis data according to the deterioration keywords and mark the target dry goods product storage quality analysis data with the analysis result of the dry goods product storage quality deterioration to generate the dry goods product storage quality deterioration analysis data ;

[0041] S62. Search for the characters of the dry goods product storage quality deterioration information in the dry goods product storage quality deterioration analysis data and generate the dry goods product storage quality deterioration status judgment data based on the search result of the characters of the dry goods product storage quality deterioration information ;

[0042] When there are no characters, indicating that the quality of the dry goods product has not deteriorated during the target dry goods product storage quality supervision process, then output the dry goods product storage quality deterioration status judgment data as non-existent, and directly end the current dry goods product storage quality supervision operation at this time;

[0043] When there are characters, indicating that the quality of the dry goods product has deteriorated during the target dry goods product storage quality supervision process, then output the dry goods product storage quality deterioration status judgment data as existent.

[0044] Preferably, when it exists, the operation steps to construct the dry goods product storage deterioration management data and perform the dry goods product storage deterioration cleaning operation are as follows:

[0045] S71. When the dry goods product storage quality deterioration status judgment data is existent, combine the dry goods product type identification data , the dry goods product storage label record text data , and the dry goods product storage quality deterioration status judgment data to construct the dry goods product storage deterioration management data , where ;

[0046] S72. Push the dry product warehousing deterioration management data online to the dry product storage quality supervision platform through the Internet of Things communication network and feedback it online to the warehousing management personnel to perform the dry product warehousing deterioration cleaning operation.

[0047] An AI-based intelligent supervision system for dry product storage quality, used to implement the AI-based intelligent supervision method for dry product storage quality. The system includes a dry product information management module, a dry product warehousing quality identification module, and a dry product warehousing exception management module;

[0048] The dry product information management module includes a dry product warehousing actual situation image acquisition unit, a dry product warehousing label image acquisition unit, a different type dry product image storage unit, a dry product type identification unit, a dry product type record information acquisition unit, a dry product information consistency judgment unit, and a dry product type record exception information feedback unit;

[0049] The dry product warehousing actual situation image acquisition unit acquires dry product warehousing actual situation image data through a cloud camera; the dry product warehousing label image acquisition unit acquires dry product warehousing label image data through a cloud camera; the different type dry product image storage unit is used to store different type dry product image data; the dry product type identification unit performs warehousing dry product type identification processing based on the dry product warehousing actual situation image data and different type dry product image data to generate dry product type identification data; the dry product type record information acquisition unit performs text information acquisition processing of dry product warehousing label records according to the dry product warehousing label image data in combination with text recognition software to generate dry product warehousing label record text data; the dry product information consistency judgment unit performs dry product type identification information and warehousing label record information consistency judgment processing based on the dry product warehousing label record text data and the dry product type identification data to generate dry product warehousing record information consistency judgment data; the dry product type record exception information feedback unit is used to construct dry product type warehousing record exception data and perform dry product type record exception information feedback operations in combination with the dry product storage quality supervision platform;

[0050] The dry product warehousing quality identification module includes a different type dry product warehousing deterioration image library storage unit, a target dry product warehousing deterioration image library search unit, and a dry product warehousing quality analysis unit;

[0051] The storage unit of the storage deterioration image library for different types of dry goods products is used to store the storage deterioration image library for different types of dry goods products; the search unit of the target storage deterioration image library for dry goods products performs a search process for the storage deterioration image library required for the specific storage quality analysis of dry goods products based on the dry goods product type identification data and the storage deterioration image library for different types of dry goods products, and generates the target storage deterioration image library for dry goods products; the storage quality analysis unit of dry goods products performs a storage quality analysis process for dry goods products based on the actual storage image data of dry goods products and the target storage deterioration image library for dry goods products, and constructs the target storage quality analysis data for dry goods products.

[0052] The abnormal management module for the storage of dry goods products includes a search unit for the storage quality deterioration information of dry goods products, a judgment unit for the storage quality deterioration state of dry goods products, and a cleaning unit for the storage deterioration of dry goods products.

[0053] The search unit for the storage quality deterioration information of dry goods products performs a search process for the storage quality deterioration information of dry goods products according to the target storage quality analysis data of dry goods products, and generates the analysis data for the storage quality deterioration of dry goods products; the judgment unit for the storage quality deterioration state of dry goods products performs a judgment process for the storage quality deterioration state of dry goods products based on the analysis data for the storage quality deterioration of dry goods products, and generates the judgment data for the storage quality deterioration state of dry goods products; the cleaning unit for the storage deterioration of dry goods products is used to construct the management data for the storage deterioration of dry goods products and perform the cleaning operation for the storage deterioration of dry goods products in combination with the storage quality supervision platform for dry goods products.

[0054] (III) Beneficial effects

[0055] The present invention provides an intelligent supervision system and method for the storage quality of dry goods products based on AI. It has the following beneficial effects:

[0056] 1. Accurately collect the in - warehouse actual situation image parameters of dry - goods products and the in - warehouse label image parameters of dry - goods products online through cloud cameras, providing real - data support for intelligent analysis of dry - goods product types and judgment of the consistency of in - warehouse information records of dry - goods products; scientifically store the image parameters of different types of dry - goods products, combine intelligent search algorithms with the in - warehouse actual situation image parameters of dry - goods products for online intelligent identification of in - warehouse dry - goods product types, improving the intelligence of dry - goods product storage management; according to the in - warehouse label image parameters of dry - goods products, combine with text - image recognition software to achieve efficient and accurate collection of the text information recorded on the in - warehouse labels of dry - goods products; based on the text information recorded on the in - warehouse labels of dry - goods products and the type - identification information of dry - goods products, independently and accurately analyze the consistency between the type - identification information of dry - goods products and the information recorded on the in - warehouse labels, realize scientific judgment of the consistency between dry - goods product type identification and the information recorded on the in - warehouse labels, and give online feedback on the abnormality of the in - warehouse record information of dry - goods products, ensuring the complete consistency between dry - goods product identification information and in - warehouse record information, and improving the scientific nature and functional diversity of intelligent supervision of dry - goods product storage.

[0057] 2. Through the combination of the in - warehouse deterioration image libraries of different types of dry - goods products based on big - data storage, intelligent search algorithms, and dry - goods product type - identification parameters, conduct search processing on the target in - warehouse deterioration image library of dry - goods products, realizing personalized and precise analysis of the storage quality of dry - goods products; based on the in - warehouse actual situation image parameters of dry - goods products, combine with intelligent recognition algorithms and the target in - warehouse deterioration image library of dry - goods products to conduct intelligent and efficient evaluation of the storage quality of dry - goods products, realizing scientific and accurate analysis of the storage quality status of in - warehouse dry - goods products, improving the efficiency and quality of storage - quality supervision of dry - goods products, and reducing the workload of dry - goods product storage management and the probability of dry - goods deterioration.

[0058] 3. Through accurate collection of in - warehouse quality deterioration information of dry - goods products according to the analysis data of the storage quality of target dry - goods products, realizing accurate collection of information on deteriorated in - warehouse dry - goods products; based on the analysis parameters of in - warehouse quality deterioration of dry - goods products, conduct digital and efficient identification of the in - warehouse quality deterioration status of dry - goods products, improving the scientific nature of in - warehouse quality management of dry - goods products; scientifically construct in - warehouse deterioration management data of dry - goods products based on dry - goods product type - identification information, in - warehouse label record information, and in - warehouse quality deterioration status judgment information, realizing accurate collection of information on deteriorated in - warehouse dry - goods products and storage location information, and at the same time, through the Internet of Things communication network, timely feedback to the storage - quality supervision platform of dry - goods products and execute the cleaning operation of deteriorated in - warehouse dry - goods products, realizing accurate feedback of information on deteriorated in - warehouse dry - goods products, and improving the convenience and applicability of dry - goods product storage management. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a schematic diagram of the modules of the intelligent storage - quality supervision system for dry - goods products based on AI provided by the present invention;

[0060] Figure 2Flow chart of the intelligent supervision method for the storage quality of dry goods products based on AI provided by the present invention. Detailed implementation manners

[0061] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0062] Embodiments of the intelligent supervision system and method for the storage quality of dry goods products based on AI are as follows:

[0063] Embodiment 1:

[0064] Please refer to Figure 1 - Figure 2 , the intelligent supervision method for the storage quality of dry goods products based on AI, the method includes the following steps:

[0065] S1. Collect the actual situation image data of the dry goods products in storage and the label image data of the dry goods products in storage;

[0066] S2. Perform type recognition processing on the dry goods products in storage according to the actual situation image data of the dry goods products in storage and the image data of different types of dry goods products, and generate type recognition data of the dry goods products;

[0067] S3. Perform text information collection processing on the storage label records of the dry goods products according to the label image data of the dry goods products in storage, generate the text data of the storage label records of the dry goods products, and perform consistency judgment processing on the type recognition information and storage label record information of the dry goods products with the type recognition data of the dry goods products to generate consistency judgment data of the storage records of the dry goods products; when they are different, construct abnormal data of the type storage records of the dry goods products and execute the feedback operation of abnormal information of the type records of the dry goods products;

[0068] S4. When the consistency judgment data of the storage records of the dry goods products is the same or when the feedback operation of the abnormal information of the type records of the dry goods products is completed, perform a search process on the image library of the deteriorated dry goods products in storage required for the specific storage quality analysis of the dry goods products based on the type recognition data of the dry goods products and the image library of the deteriorated dry goods products in storage of different types, and generate the target image library of the deteriorated dry goods products in storage;

[0069] S5. Perform storage quality analysis processing on the dry goods products according to the actual situation image data of the dry goods products in storage and the target image library of the deteriorated dry goods products in storage, and construct the target storage quality analysis data of the dry goods products;

[0070] S6. Search and process the information on the deterioration of the storage quality of dry goods products based on the analysis data of the storage quality of target dry goods products, generate the analysis data on the deterioration of the storage quality of dry goods products, perform the judgment process on the deterioration status of the storage quality of dry goods products, and generate the judgment data on the deterioration status of the storage quality of dry goods products. When it is determined that there is no deterioration, directly end the current storage quality supervision operation of dry goods products;

[0071] S7. When there is deterioration, construct the management data on the deterioration of the storage of dry goods products and perform the cleaning operation on the deterioration of the storage of dry goods products.

[0072] Furthermore, please refer to Figure 1 - Figure 2 , and the operating steps for collecting the on-site image data of the storage situation of dry goods products and the image data of the storage labels of dry goods products are as follows:

[0073] S11. Online collect the on-site image information of the dry goods products to be tested in the unpacked state in the storage warehouse through the cloud camera, and generate the on-site image data of the storage situation of dry goods products ;

[0074] Online collect the image information of the product information record labels on the storage shelves of the dry goods products to be tested in the storage warehouse through the cloud camera, and generate the image data of the storage labels of dry goods products .

[0075] The operating steps for identifying the types of dry goods products in the storage based on the on-site image data of the storage situation of dry goods products and the image data of different types of dry goods products, and generating the type identification data of dry goods products are as follows:

[0076] S21. Establish a set of image data of different types of dry goods products , where represents the image data of different types of dry goods products corresponding to the th type of dry goods product, represents the maximum value of the number of dry goods product types; the dry goods product types include chestnuts, Chinese chestnuts, pili nuts, hazelnuts, cashews, walnuts, melon seeds, pine nuts, almonds, ginkgo nuts, pistachios, pecans, desert nuts, torreya nuts, white melon seeds, pumpkin seeds, peanuts, almonds, macadamia nuts, black fungus, laver, shiitake mushrooms, red dates, cinnamon, chili peppers, Chinese prickly ash, star anise, fennel, pepper, goji berries, longans, peanuts, dried tangerine peels, raisins, dried green vegetables, dried cowpeas, dried string beans, dried eggplants, dried asparagus lettuce, dried pakchoi, dried cucumbers, dried eggplants, preserved mustard greens, dried daylilies, dried bamboo shoots and dried radishes, sorghum, millet, buckwheat, sweet buckwheat, bitter buckwheat, oats, naked oats, barley, broomcorn millet, millet, coix seeds, amaranth, kidney beans, mung beans, adzuki beans, red beans, broad beans, peas, cowpeas, lentils and black beans, and the image data of different types of dry goods products represent the product image data corresponding to different types of dry goods products;

[0077] S22. Use the K-D tree nearest neighbor search algorithm to compare the real-time image data of the dry goods product warehouse with the image data sets of different types of dry goods products in the different types of dry goods product image data for image feature matching, and search for the different types of dry goods product image data matching the real-time image data of the dry goods product warehouse to obtain the corresponding dry goods product type text information, and generate dry goods product type recognition data .

[0078] Collect and process the text information of the dry goods product warehouse label records based on the dry goods product warehouse label image data, generate the dry goods product warehouse label record text data, and perform a consistency judgment process on the type recognition information and warehouse label record information of the dry goods product with the dry goods product type recognition data to generate the consistency judgment data of the dry goods product warehouse record information; when they are different, construct the abnormal data of the dry goods product type warehouse record and execute the following operation steps of the abnormal information feedback operation for the dry goods product type record:

[0079] S31. Use an optical character recognition software to collect and process the text information of the dry goods product warehouse label records in the dry goods product warehouse label image data, and generate the dry goods product warehouse label record text data , and the optical character recognition software includes any one of AWS Text and Adobe Acrobat DC. The dry goods product warehouse label record text data includes the name text data, specification quantity text data, production date text data, and warehouse shelf location text data of the dry goods product;

[0080] S32. Perform keyword matching for the dry goods product type between the dry goods product type recognition data and the dry goods product warehouse label record text data to generate the consistency judgment data of the dry goods product warehouse record information based on the keyword matching result of the dry goods product type ;

[0081] When and successfully perform keyword matching for the dry goods product type, indicating that the dry goods product type recognition information is consistent with the dry goods product type information in the label record, then output the consistency judgment data of the dry goods product warehouse record information as the same;

[0082] When and The keyword matching for the dry product type fails, indicating that the dry product type identification information is inconsistent with the dry product type information recorded in the label. Then, the consistency judgment data of the dry product storage record information is output are different; at this time, the dry product type identification data , the dry product storage label record text data , and the consistency judgment data of the dry product storage record information are combined to construct the abnormal data of the dry product type storage record , and the abnormal data of the dry product type storage record is pushed online through the Internet of Things communication network and fed back to the dry product storage quality supervision platform to execute the feedback operation of the abnormal information of the dry product type record.

[0083] Through the cooperation of the dry product storage live image acquisition unit and the dry product storage label image acquisition unit, the live image parameters of the dry product storage and the label image parameters of the dry product storage are accurately collected online by the cloud lens, providing real data support for the intelligent analysis of the dry product type and the judgment of the consistency of the dry product storage information record; the image storage units of different types of dry products and the dry product type identification unit cooperate with each other to scientifically store the image parameters of different types of dry products, combine the intelligent search algorithm with the live image parameters of the dry product storage to perform online intelligent identification of the dry product type in the storage, and improve the intelligence of the dry product storage management; the dry product type record information acquisition unit efficiently and accurately acquires the text information of the dry product storage label record according to the dry product storage label image parameters combined with the text and image recognition software; the dry product information consistency judgment unit and the dry product type record abnormal information feedback unit cooperate with each other, and based on the text information of the dry product storage label record and the dry product type identification information, the consistency of the type identification information of the dry product and the storage label record information is analyzed independently and accurately, realizing the scientific judgment of the consistency of the dry product type identification and the storage label record information, and online feedback of the abnormal dry product storage record information to ensure that the dry product identification information and the storage record information are completely consistent, and improving the scientific nature and functional diversity of the dry product storage intelligent supervision.

[0084] Further, please refer to Figure 1 - Figure 2 , when the consistency judgment data of the dry product storage record information is the same or when the feedback operation of the dry product type record abnormal information is completed, based on the dry product type identification data and the different types of dry product storage deterioration image libraries, the search process of the dry product storage deterioration image library required for the specific dry product storage quality analysis is carried out. The operation steps for generating the target dry product storage deterioration image library are as follows:

[0085] S41. When the consistency judgment data of the dry product storage record information For the same or when the dry goods product type records abnormal information feedback job execution is completed, establish a collection of different types of dry goods product storage deterioration image libraries ,in Indicates Different types of dry goods product storage deterioration image libraries corresponding to the dry goods product types, ,in Image library representing different types of dry goods deterioration during storage Middle Image data of different types of dry goods deterioration during storage. Image library representing different types of dry goods deterioration during storage Middle The storage deterioration image data of different types of dry goods products; the storage deterioration image data of different types of dry goods products represent the appearance image data of different types of dry goods products after they deteriorate during the storage process; the storage deterioration of dry goods products includes the appearance of black spots, mold spots, holes and color changes on the dry goods products;

[0086] S42, using breadth-first search algorithm to identify dry goods product type data Image gallery collection with different types of dry goods products going bad in warehouse Image library of different types of dry goods products deteriorating during storage Perform dry goods product type character matching to search for dry goods product type identification data Corresponding image library of different types of dry goods products deteriorating during storage , and generate the target dry goods product storage deterioration image library through data identification ,in Represents the target dry goods product storage deterioration image library Middle The target dry goods product storage deterioration image data, Represents the target dry goods product storage deterioration image library Middle The target dry goods product storage deterioration image data represents the dry goods product appearance image data corresponding to the dry goods product type to be tested after deterioration during the storage process.

[0087] The steps for analyzing and processing the storage quality of dry goods products based on the actual storage image data of dry goods products and the storage deterioration image library of target dry goods products to construct the storage quality analysis data of target dry goods products are as follows:

[0088] S51. Acquiring real-time image data of dry goods storage and Target Dry Goods Product Storage Deterioration Image Library ;

[0089] S52. Match the real-time image data of the dry goods product storage with the target dry goods product storage deterioration image library for the target dry goods product storage deterioration image data to perform image feature matching, and generate target dry goods product storage quality analysis data based on the image feature matching results ; Execute the specific operation steps for generating the target dry goods product storage quality analysis data as follows:

[0090] S521. Initialize, define the relevant structure parameters as vectors, and in the search space of the target dry goods product storage deterioration image library the target dry goods product storage deterioration image data to constitute dimensional optimization problem, where the dry goods product storage deterioration recognition sand cat represents the solution of the problem as a 1× array, and each variable value is a floating-point number, and each variable value to represents being between the lower bound and the upper bound in the search space of the target dry goods product storage deterioration image library ;

[0091] S522. Search for prey. The final parameter and the main parameter for controlling the transition between the exploration and exploitation phases are , when > 1, the dry goods product storage deterioration recognition sand cat searches in the search space of the target dry goods product storage deterioration image library for the target dry goods product storage deterioration image data that matches the real-time image data of the dry goods product storage . The search process of the dry goods product storage deterioration recognition sand cat depends on the release of low-frequency noise to search for the target dry goods product storage deterioration image data that matches the real-time image data of the dry goods product storage . Assume that the sensitivity range of the dry goods product storage deterioration recognition sand cat is from 0 to 2 kHz, represents being inspired by the auditory characteristics of the dry goods product storage deterioration recognition sand cat, and assume its value is 2, is the current iteration number, is the maximum iteration number, and rand(0,1) represents a random number with a value from 0 to 1; , , where represents the sensitivity vector; Each dry goods product storage deterioration recognition sand cat is based on the best candidate position and the current position and its sensitivity range Update its own position, that is, in the target dry product storage deterioration image library Update the position of the dry product storage actual situation image data in the search space The position of the target dry product storage deterioration image data that matches; the position calculation formula is as follows: , where Represents the current position of the sand cat individual for dry product storage deterioration recognition in the target dry product storage deterioration image library at the (t + 1)-th iteration in the search space, Represents the position of the sand cat for dry product storage deterioration recognition at the t-th iteration in the target dry product storage deterioration image library The best candidate position of the target dry product storage deterioration image data that is most matched with the dry product storage actual situation image data searched in the search space ;

[0092] S523. Attack the prey. When ≤ 1, the sand cat for dry product storage deterioration recognition searches in the target dry product storage deterioration image library in the search space, and uses the best candidate position and the current position to generate a random position, that is, randomly search for the random position of the target dry product storage deterioration image data that matches the dry product storage actual situation image data in the target dry product storage deterioration image library Assume that the sensitivity range of the sand cat for dry product storage deterioration recognition is a circle, and use the roulette method to randomly select an angle for each sand cat for dry product storage deterioration recognition According to the random position calculation formula: , perform a random position search for the target dry product storage deterioration image data that matches the dry product storage actual situation image data , where represents the sensitivity vector, represents the angle represents the cosine value of , represents the random position of the sand cat for dry product storage deterioration recognition in the target dry product storage deterioration image library The random position makes the sand cat for dry product storage deterioration recognition approach and attack the prey, that is, search for the target dry product storage deterioration image data that matches the dry product storage actual situation image data in the target dry product storage deterioration image library ;

[0093] S524. When the search algorithm reaches the maximum number of iterations T, output the data that matches the dry product storage actual situation image data​ Matched target dry goods product storage deterioration image data;

[0094] S525. Generate target dry goods product storage quality analysis data based on the image feature matching result between the actual situation image data of the dry goods product storage output in step S524 and the target dry goods product storage deterioration image data; ;

[0095] When matches with or successfully in image feature matching, indicating that deterioration has occurred during the storage process of the dry goods product, then it means to output the target dry goods product storage quality analysis data is deterioration;

[0096] When does not match with or successfully in image feature matching, indicating that no deterioration has occurred during the storage process of the dry goods product, then it means to output the target dry goods product storage quality analysis data is normal.

[0097] Through the mutual cooperation of the storage units of different types of dry goods product storage deterioration image libraries and the search units of the target dry goods product storage deterioration image library, based on the different types of dry goods product storage deterioration image libraries stored in large data, combined with intelligent search algorithms and dry goods product type recognition parameters, the search processing of the target dry goods product storage deterioration image library is carried out to realize personalized and accurate analysis of the storage quality of dry goods products; the dry goods product storage quality analysis unit, based on the actual situation image parameters of the dry goods product storage, combines intelligent recognition algorithms and the target dry goods product storage deterioration image library to conduct intelligent and efficient evaluation of the storage quality of dry goods products, realizing scientific and accurate analysis of the storage quality status of the stored dry goods products, improving the efficiency and quality of the storage quality supervision of dry goods products, reducing the workload of dry goods product storage management and the probability of dry goods deterioration.

[0098] Further, please refer to Figure 1 - Figure 2 , perform search processing on the dry goods product storage quality deterioration information according to the target dry goods product storage quality analysis data, generate dry goods product storage quality deterioration analysis data and conduct judgment processing on the dry goods product storage quality deterioration state, and generate dry goods product storage quality deterioration state judgment data. When it is non-existent, the operation steps to directly end the current dry goods product storage quality supervision operation are as follows:

[0099] S61. Use the breadth-first search algorithm to search the target dry goods product storage quality analysis data according to the deterioration keywords Search and process the information on the deterioration of the quality of dry goods products in storage, and obtain the target dry goods product storage quality analysis data with the analysis result of deterioration in the quality of dry goods products in storage After data identification, generate the analysis data on the deterioration of the quality of dry goods products in storage ;

[0100] S62. For the analysis data on the deterioration of the quality of dry goods products in storage Conduct a character search on the information of the deterioration of the quality of dry goods products in storage, and generate the judgment data on the deterioration status of the quality of dry goods products in storage based on the result of the character search on the information of the deterioration of the quality of dry goods products in storage ;

[0101] When There are no characters, indicating that the quality of the dry goods products has not deteriorated during the supervision process of the target dry goods product storage quality, then output the judgment data on the deterioration status of the quality of dry goods products in storage as non - existent, and directly end the current supervision operation of the storage quality of dry goods products at this time;

[0102] When There are characters, indicating that the quality of the dry goods products has deteriorated during the supervision process of the target dry goods product storage quality, then output the judgment data on the deterioration status of the quality of dry goods products in storage as existent.

[0103] When it is existent, the operation steps for constructing the management data on the deterioration of dry goods products in storage and performing the cleaning operation on the deterioration of dry goods products in storage are as follows:

[0104] S71. When the judgment data on the deterioration status of the quality of dry goods products in storage is existent, combine the dry goods product type identification data , the text data of the dry goods product storage label record , and the judgment data on the deterioration status of the quality of dry goods products in storage through data combination to construct the management data on the deterioration of dry goods products in storage , where ;

[0105] S72. Push the management data on the deterioration of dry goods products in storage to the storage quality supervision platform of dry goods products through the online feedback of the Internet of Things communication network, and online feedback to the warehouse management personnel to perform the cleaning operation on the deterioration of dry goods products in storage.

[0106] Through the dry product storage quality deterioration information search unit, accurate collection of dry product storage quality deterioration information is carried out according to the target dry product storage quality analysis data, realizing accurate collection of information on deteriorated dry products in storage; the dry product storage quality deterioration status judgment unit digitally and efficiently identifies the dry product storage quality deterioration status based on the dry product storage quality deterioration analysis parameters, improving the scientific nature of dry product storage quality management; the dry product storage deterioration cleaning unit scientifically constructs dry product storage deterioration management data based on the dry product type identification information, storage label record information, and storage quality deterioration status judgment information, realizing accurate collection of information on deteriorated dry products in storage and their storage location information, and at the same time timely feedback to the dry product storage quality supervision platform through the Internet of Things communication network and execute the dry product storage deterioration cleaning operation, realizing accurate feedback of information on deteriorated dry products in storage, and improving the convenience and applicability of dry product storage management.

[0107] Embodiment 2:

[0108] Please refer to Figure 1 - Figure 2 , an AI-based intelligent supervision system for dry product storage quality, used to implement an AI-based intelligent supervision method for dry product storage quality. The system includes a dry product information management module, a dry product storage quality identification module, and a dry product storage anomaly management module;

[0109] The dry product information management module includes a dry product storage live image acquisition unit, a dry product storage label image acquisition unit, a storage unit for images of different types of dry products, a dry product type identification unit, a dry product type record information acquisition unit, a dry product information consistency judgment unit, and a dry product type record anomaly information feedback unit;

[0110] The dry product storage live image acquisition unit acquires dry product storage live image data through a cloud lens; the dry product storage label image acquisition unit acquires dry product storage label image data through a cloud lens; the different type dry product image storage unit is used to store different type dry product image data; the dry product type recognition unit performs storage dry product type recognition processing based on the dry product storage live image data and different type dry product image data, and generates dry product type recognition data; the dry product type record information acquisition unit performs text information acquisition processing of the dry product storage label record according to the dry product storage label image data in combination with an optical character recognition software, and generates dry product storage label record text data; the dry product information consistency judgment unit performs dry product type recognition information and storage label record information consistency judgment processing based on the dry product storage label record text data and the dry product type recognition data, and generates dry product storage record information consistency judgment data; the dry product type record abnormal information feedback unit is used to construct dry product type storage record abnormal data and execute dry product type record abnormal information feedback operations in combination with the dry product storage quality supervision platform;

[0111] The dry product storage quality recognition module includes a different type dry product storage deterioration image library storage unit, a target dry product storage deterioration image library search unit, and a dry product storage quality analysis unit;

[0112] The different type dry product storage deterioration image library storage unit is used to store different type dry product storage deterioration image libraries; the target dry product storage deterioration image library search unit performs target dry product storage deterioration image library search processing required for specific dry product storage quality analysis based on the dry product type recognition data and different type dry product storage deterioration image libraries, and generates a target dry product storage deterioration image library; the dry product storage quality analysis unit performs dry product storage quality analysis processing based on the dry product storage live image data and the target dry product storage deterioration image library, and constructs target dry product storage quality analysis data;

[0113] The dry product storage abnormal management module includes a dry product storage quality deterioration information search unit, a dry product storage quality deterioration state judgment unit, and a dry product storage deterioration cleaning unit;

[0114] The dry product storage quality deterioration information search unit performs dry product storage quality deterioration information search processing based on the target dry product storage quality analysis data, and generates dry product storage quality deterioration analysis data; the dry product storage quality deterioration status judgment unit performs dry product storage quality deterioration status judgment processing based on the dry product storage quality deterioration analysis data, and generates dry product storage quality deterioration status judgment data; the dry product storage deterioration cleaning unit is used to construct dry product storage deterioration management data and perform dry product storage deterioration cleaning operations in combination with the dry product storage quality supervision platform.

[0115] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An AI-based intelligent supervision method for storage quality of dry goods products, characterized in that: The method comprises the following steps: S1. Collecting real-time image data of dry goods product storage and image data of dry goods product storage labels; S2, performing storage dry goods product type recognition processing based on the dry goods product storage real-time image data and the image data of different types of dry goods products to generate dry goods product type recognition data; S3, collecting and processing the text information of the dry goods product storage label record according to the dry goods product storage label image data, generating dry goods product storage label record text data, and performing consistency judgment processing on the dry goods product type identification information and storage label record information with the dry goods product type identification data, generating dry goods product storage record information consistency judgment data; when they are not the same, constructing dry goods product type storage record abnormal data and executing dry goods product type record abnormal information feedback operation; S4. When the dry goods product storage record information consistency judgment data is the same or when the dry goods product type record abnormal information feedback operation is completed, a dry goods product storage deterioration image library search process required for specific dry goods product storage quality analysis is performed based on the dry goods product type identification data and different types of dry goods product storage deterioration image libraries to generate a target dry goods product storage deterioration image library; S5, performing dry goods product storage quality analysis processing based on the dry goods product storage real-time image data and the target dry goods product storage deterioration image library to construct target dry goods product storage quality analysis data; S6. Perform a search process for storage quality deterioration information of dry goods products according to the storage quality analysis data of the target dry goods products, generate storage quality deterioration analysis data of dry goods products, perform a storage quality deterioration state judgment process of dry goods products, and generate storage quality deterioration state judgment data of dry goods products. If the judgment data does not exist, directly end the storage quality supervision operation of dry goods products. S7. If it exists, construct the dry goods product storage deterioration management data and perform the dry goods product storage deterioration cleaning operation.

2. The AI-based intelligent supervision method for storage quality of dry goods products according to claim 1 is characterized by: The S1 comprises the following steps: S11. Collect the on-site image information of the dry goods products to be tested in the unpacked state in the storage plant online through the cloud camera, and generate the real-time image data of the dry goods product storage ; The cloud lens collects the product information of the dry goods products to be tested on the storage shelves in the warehouse and records the label image information, and generates the dry goods product storage label image data .

3. The AI-based intelligent supervision method for storage quality of dry goods products according to claim 2 is characterized by: The S2 comprises the following steps: S21. Establish image data sets of different types of dry goods products ; S22, using the KD tree nearest neighbor search algorithm to With the Middle Different types of dry goods product image data corresponding to the dry goods product types Perform image feature matching and search for Matching the Corresponding dry goods product type text information, and generate dry goods product type identification data .

4. The AI-based intelligent supervision method for storage quality of dry goods products according to claim 3 is characterized by: The S3 comprises the following steps: S31, using image and text recognition software to identify the image data of the dry goods product storage label Collect and process the text information of dry goods product storage label records, and generate dry goods product storage label record text data ; S32, the With the Perform keyword matching of dry goods product types, and generate consistency judgment data of dry goods product storage record information based on the keyword matching results of dry goods product types ; when and If the dry goods product type keyword match is successful, the dry goods product storage record information consistency judgment data is output for the same; when and If the dry goods product type keyword is not matched successfully, the dry goods product storage record information consistency judgment data is output are different; at this time, , , Combine data to construct abnormal data of dry goods product type warehouse records , the Online feedback is pushed to the dry goods product storage quality supervision platform through the Internet of Things communication network to execute the feedback operation of abnormal information on dry goods product type records.

5. The AI-based intelligent supervision method for storage quality of dry goods products according to claim 4 is characterized by: The S4 comprises the following steps: S41, when the For the same or when the dry goods product type records abnormal information feedback job execution is completed, establish a collection of different types of dry goods product storage deterioration image libraries ; S42, using a breadth-first search algorithm to With the Middle Different types of dry goods product storage deterioration image library corresponding to different types of dry goods product types Perform dry goods product type character matching and search for the The corresponding , and generate the target dry goods product storage deterioration image library through data identification .

6. The AI-based intelligent supervision method for storage quality of dry goods products according to claim 5 is characterized by: The S5 comprises the following steps: S51, obtaining the and stated ; S52, the With the middle to Perform image feature matching on the target dry goods product storage deterioration image data, and generate the target dry goods product storage quality analysis data based on the image feature matching results ; Execute and generate the storage quality analysis data of the target dry goods product The specific steps are as follows: S521, initialization, defining the relevant structural parameters as vectors, in the The target dry goods product storage deterioration image data in the search space to constitute In the dimensional optimization problem, the sand cat represents the 1× Array, each variable value is a floating point number, each variable value to Indicates that the between lower and upper bounds in the search space; S522, Searching for prey, the final and main parameters controlling the transition between exploration and exploitation phases are ,when >1, dry goods storage spoilage identification sand cat in the Search the search space for The dry goods product storage deterioration image data that matches the target dry goods product storage deterioration recognition sand cat search process relies on the release of low-frequency noise to search for the dry goods product storage deterioration recognition sand cat search process that matches the target ... The target dry goods product storage spoilage image data matched, assuming that the sensitivity range of the dry goods product storage spoilage identification sand cat From 0 to 2kHz, It means that it is inspired by the auditory characteristics of sand cats in the identification of dry goods deterioration during storage. Assume that its value is 2. is the current iteration number, is the maximum number of iterations, rand(0,1) represents a random number between 0 and 1; each dry goods product storage spoilage identification sand cat is based on the best candidate position and current location And its sensitivity range Update your location, that is, The search space is updated with the The location of the matched storage deterioration image data of the target dry goods product; S523, attack prey, when ≤1, the dry goods storage spoilage identification sand cat is described in Search the search space and use the best candidate position With current location Generate a random position, that is, Randomly search the search space for The random position of the target dry goods product storage deterioration image data that matches the target dry goods product storage deterioration image data, assuming that the sensitivity range of the dry goods product storage deterioration identification sand cat is a circle, and using the roulette method to randomly select an angle for each dry goods product storage deterioration identification sand cat , according to the random position and the The random position search of the target dry goods product storage spoilage image data that matches the random position makes the dry goods product storage spoilage identification sand cat approach and attack the prey, that is, in the Search the search space for Matching storage deterioration image data of the target dry goods product; S524. When the search algorithm reaches the maximum number of iterations T, output Matching storage deterioration image data of the target dry goods product; S525: output according to step S524 The image feature matching result with the image data of the target dry goods product storage deterioration is used to generate the storage quality analysis data of the target dry goods product ; when and or If the image feature matching is successful, it means the storage quality analysis data of the target dry goods product is output. for deterioration; when and or If no image features are matched successfully, the target dry goods product storage quality analysis data is output. is normal.

7. The AI-based intelligent supervision method for storage quality of dry goods products according to claim 6 is characterized by: The S6 comprises the following steps: S61, using a breadth-first search algorithm to search the Search and process the storage quality deterioration information of dry goods products, and report the storage quality analysis results of dry goods products that are deteriorated After data identification, the storage quality deterioration analysis data of dry goods products is generated ; S62. Perform a character search for the storage quality deterioration information of dry goods products, and generate the storage quality deterioration status judgment data of dry goods products based on the search results of the character search results of the storage quality deterioration information of dry goods products ; when If there is no character in the , the dry goods product storage quality deterioration status judgment data is output If it does not exist, the dry goods product storage quality supervision operation will be terminated directly; when If there are characters in the , the dry goods product storage quality deterioration status judgment data is output For existence.

8. The AI-based intelligent supervision method for storage quality of dry goods products according to claim 7 is characterized by: The S7 comprises the following steps: S71, when the When exists, the , , Through data combination, dry goods storage deterioration management data is constructed ; S72, the Online feedback is pushed to the dry goods product storage quality supervision platform through the Internet of Things communication network, and online feedback is given to warehouse management personnel to perform the cleanup of dry goods storage spoilage.

9. An AI-based intelligent supervision system for storage quality of dry goods products, used to implement the AI-based intelligent supervision method for storage quality of dry goods products described in any one of 1-8, characterized in that: The system includes a dry goods product information management module, a dry goods product storage quality identification module, and a dry goods product storage abnormality management module.

Images