Automatic production method and system for spleen-tonifying eight-treasure pastry

Abstract

The present application discloses an automatic production method and system for a spleen-tonifying eight-treasure pastry. When a semi-finished product is dried by means of microwave heating, the real-time change of water content of the semi-finished product is monitored on the basis of a machine vision technique, so as to timely stop drying. In this way, drying can be timely stopped on the basis of the real-time change condition of water content of the pastry, so as to ensure that the water content of eight-treasure pastry products can be controlled within a preset range, thereby improving the quality consistency of the produced eight-treasure pastry products.

Description

Automated production method and system for spleen-strengthening eight-treasure cake Technical Field

[0001] The present application relates to the field of intelligent production, and more specifically, to an automated production method and system for spleen-strengthening eight-treasure cake. Background Art

[0002] Jianpi Bazhen Cake is a famous health-care pastry, mainly made of eight Chinese herbal medicines such as Codonopsis pilosula, Atractylodes macrocephala, Poria cocos, and rice flour. It is often used to treat symptoms such as anorexia, malnutrition, indigestion, and chronic diarrhea in children. It is also suitable for dietary therapy of indigestion in the elderly.

[0003] The traditional Bazhen Cake production process typically involves grinding the various ingredients into a powder, mixing with water, molding, and steaming. However, this processing method presents several challenges. First, the product has a high moisture content, making it susceptible to moisture, mold, and spoilage. Second, because it's simply mixed and steamed, the various ingredients in the cake aren't fully bonded, resulting in a poor texture, a dry taste, and a tendency to crumble. This makes it unappetizing for those with weak spleens and stomachs.

[0004] Another traditional rice cake method involves soaking and grinding rice. Initially, this rice cake has a delicate, soft, and rich rice aroma. However, over time, starch retrogradation occurs, resulting in a shorter shelf life or requiring the addition of large amounts of gums or modified starches to improve the texture. Furthermore, rice cakes produced using wet rice grinding are too sticky and difficult to digest, making them unsuitable for the elderly and young children. Therefore, simply using wet rice grinding cannot be used for this purpose.

[0005] In addition, in the traditional Bazhen Cake processing, various Chinese medicinal materials are simply ground into powder and mixed. This simple addition method makes the absorption rate of various Chinese medicinal materials slow, the bioavailability is low, and the quality control is difficult.

[0006] Therefore, expect a kind of optimized automatic production method scheme of spleen-strengthening eight-treasure cake.

[0007] Summary of the Invention

[0008] The present application is proposed to address the aforementioned technical issues. The embodiments of the present application provide an automated production method and system for a spleen-strengthening eight-treasure cake. While drying the semi-finished product via microwave heating, the method uses machine vision technology to monitor the real-time changes in the moisture content of the semi-finished product and to stop the drying process appropriately. This allows the drying process to be stopped appropriately based on the real-time changes in the moisture content of the cake, ensuring that the moisture content of the eight-treasure cake product is controlled within a predetermined range, thereby improving the quality consistency of the produced eight-treasure cake product.

[0009] According to one aspect of the present application, there is provided an automated production method for spleen-strengthening eight-treasure cake, comprising:

[0010] Adding codonopsis pilosula into distilled water and performing reflux extraction at a predetermined temperature, and centrifuging to obtain the supernatant to obtain a first mixed solution, and at the same time obtaining the remaining medicinal residue;

[0011] Adding distilled water to the remaining medicinal residue and placing it in an ultrasonic generator container for ultrasonic extraction, and centrifuging to obtain the supernatant to obtain a second mixed solution;

[0012] mixing and concentrating the first mixed solution and the second mixed solution to obtain an extract;

[0013] Adding Atractylodes macrocephala, Poria cocos, white hyacinth bean, coix seed, Chinese yam, Euryale ferox, and lotus seed into purified water for soaking and grinding, then adding sucrose to obtain a mixture, and placing the mixture into a chopper for chopping to obtain a slurry;

[0014] Mixing rice flour with the extract and the slurry, adding magnesium stearate and purified water, and placing the mixture into the chopping machine for processing to obtain a semi-finished product; and

[0015] The semi-finished product is placed in a mold for shaping, and the shaped semi-finished product is dried by microwave heating to obtain the eight-treasure cake.

[0016] According to another aspect of the present application, an automated production system for spleen-strengthening eight-treasure cake is provided, comprising:

[0017] A video acquisition module is used to collect drying monitoring videos of semi-finished products within a predetermined time period through a camera;

[0018] a temporal semantic analysis module, configured to analyze the drying monitoring video to obtain temporal semantic features of the drying state; and

[0019] The control result generating module is used to determine whether to stop microwave heating based on the temporal semantic features of the drying state.

[0020] Compared to existing technologies, the automated production method and system for spleen-strengthening eight-treasure cake provided in this application utilizes machine vision technology to monitor the real-time changes in the moisture content of the semi-finished product during microwave drying, allowing for timely cessation of the drying process. This allows for timely cessation of the drying process based on the real-time changes in the moisture content of the cake, ensuring that the moisture content of the eight-treasure cake product remains within a predetermined range, thereby improving the quality consistency of the resulting eight-treasure cake. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The above and other purposes, features, and advantages of the present application will become more apparent through a more detailed description of the embodiments of the present application in conjunction with the accompanying drawings. The accompanying drawings are intended to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0022] FIG1 is a flow chart of an automated production method of a spleen-strengthening eight-treasure cake according to an embodiment of the present application;

[0023] FIG2 is a system architecture diagram of an automated production method for the spleen-strengthening eight-treasure cake according to an embodiment of the present application;

[0024] FIG3 is a flow chart of the training phase of the automated production method of the spleen-strengthening eight-treasure cake according to an embodiment of the present application;

[0025] FIG4 is a flow chart of sub-step S6 of the automated production method of the spleen-strengthening eight-treasure cake according to an embodiment of the present application;

[0026] FIG5 is a flow chart of sub-step S62 of the automated production method of the spleen-strengthening eight-treasure cake according to an embodiment of the present application;

[0027] FIG6 is a block diagram of an automated production system for the spleen-strengthening eight-treasure cake according to an embodiment of the present application;

[0028] FIG7 is a schematic diagram of a scenario of an automated production method of the spleen-strengthening eight-treasure cake according to an embodiment of the present application. DETAILED DESCRIPTION

[0029] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described herein.

[0030] As used in this application and the claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not intended to refer to the singular but may include the plural. Generally speaking, the terms "comprises" and "include" only indicate the inclusion of the steps and elements specifically identified, and these steps and elements do not constitute an exclusive list. A method or apparatus may also include other steps or elements.

[0031] Although the present application makes various references to certain modules in the system according to embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are illustrative only, and different aspects of the system and method can use different modules.

[0032] Flowcharts are used in this application to illustrate the operations performed by the systems according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, the various steps may be processed in reverse order or simultaneously, as needed. Furthermore, other operations may be added to these processes, or one or more operations may be removed from these processes.

[0033] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described herein.

[0034] Another method for making traditional rice cakes is to soak and grind rice. This type of rice cake has a delicate, soft and glutinous texture during initial processing, with a rich rice aroma. However, over time, starch retrogradation is likely to occur, resulting in a shorter shelf life of the product, or requiring the addition of large amounts of gums or modified starches to improve the poor texture. At the same time, this type of rice cake produced using wet rice grinding is too sticky and difficult to digest, making it unsuitable for the elderly and young children. Therefore, it cannot be simply processed using the wet rice grinding method. In addition, in the processing of traditional Bazhen Cake, various Chinese medicinal materials are simply ground into powder and mixed. This simple addition method results in slow absorption of the various Chinese medicinal materials, low bioavailability, and difficulty in quality control. Therefore, an optimized automated production method for spleen-strengthening Bazhen Cake is desired.

[0035] In the technical solution of the present application, an automated production method for the spleen-strengthening eight-treasure cake is proposed. Figure 1 is a flow chart of the automated production method for the spleen-strengthening eight-treasure cake according to the embodiment of the present application. As shown in Figure 1, the automated production method for the spleen-strengthening eight-treasure cake according to the embodiment of the present application includes the following steps: S1, adding codonopsis pilosula to distilled water and performing reflux extraction at a predetermined temperature, and centrifuging to obtain the supernatant to obtain a first mixed liquid, and at the same time obtaining the remaining medicinal residue; S2, adding the remaining medicinal residue to distilled water and placing it in an ultrasonic generating container for ultrasonic extraction, and centrifuging to obtain the supernatant to obtain a second mixed liquid; S3, mixing and concentrating the first mixed liquid and the second mixed liquid to obtain an extract; S4, adding Atractylodes macrocephala and Poria cocos. Poria, white lentil, coix seed, yam, gorgon fruit, and lotus seed are added to purified water for soaking and pulping, and then sucrose is added to obtain a mixture, and the mixture is placed in a chopper for chopping to obtain a slurry; S5, rice flour, the extract and the slurry are mixed, magnesium stearate and purified water are added, and the mixture is placed in the chopper for processing to obtain a semi-finished product; and, S6, the semi-finished product is placed in a mold for shaping, and the shaped semi-finished product is dried by microwave heating to obtain the eight-treasure cake.

[0036] Specifically, in step S1, after adding codonopsis pilosula into distilled water, reflux extraction is performed at a predetermined temperature, and the supernatant is centrifuged to obtain a first mixed solution, and the remaining medicinal residue is obtained at the same time. After adding codonopsis pilosula into distilled water, reflux extraction is performed at a predetermined temperature, which can help extract the effective ingredients in codonopsis pilosula. This process utilizes the principle of reflux extraction. By controlling the temperature and time, the active ingredients in codonopsis pilosula are dissolved in distilled water, and effective extraction is achieved as the distilled water circulates and refluxes. Among them, codonopsis pilosula, also known as polygonatum, is a Chinese herbal medicine, and its scientific name is Codonopsis pilosula. It is one of the traditional Chinese medicinal materials and is widely used in the field of traditional Chinese medicine. The root of codonopsis pilosula contains a variety of active ingredients, including polysaccharides, saponins, amino acids, etc. It is believed to have the effects of nourishing and strengthening, replenishing qi and blood, improving immunity, regulating and maintaining.

[0037] It is worth noting that reflux extraction is a commonly used extraction technology suitable for extracting active ingredients from plant materials. During the reflux extraction process, the plant material is mixed with an extraction solvent (such as distilled water), heated to a set temperature, and then the solvent is circulated through a reflux device so that the active ingredients in the solvent are fully in contact with and dissolved in the plant material. This can increase the extraction efficiency and improve the concentration of the target ingredient. For Codonopsis pilosula, reflux extraction can help extract the polysaccharides, saponins and other active ingredients therein. These ingredients have the effects of nourishing and strengthening, invigorating qi and nourishing blood, and are beneficial to physical health. Through reflux extraction, these active ingredients in Codonopsis pilosula can be effectively extracted for the preparation of herbal beverages, Chinese medicine preparations, etc.

[0038] Accordingly, in a possible implementation, the following steps can be performed to add distilled water to the Codonopsis pilosula and then perform reflux extraction at a predetermined temperature, and centrifuge to obtain the supernatant to obtain a first mixed solution, and at the same time obtain the remaining medicinal residue, for example: add an appropriate amount of distilled water to the reflux extractor; add the Codonopsis pilosula medicinal material to the reflux extractor; set a suitable extraction temperature and time according to the characteristics of the Codonopsis pilosula and the solubility characteristics of the required components; ensure that the reflux extractor is well sealed to avoid loss of solvent and entry of external impurities; turn on the heating device to heat the solvent in the reflux extractor to the set temperature; when the solvent begins to boil, add The heating device will condense the steam into liquid and reflux it into the reflux extractor, where it will fully contact the Codonopsis pilosula medicinal material; continue the reflux extraction for a period of time to ensure that the active ingredients in the Codonopsis pilosula are fully dissolved in the solvent; turn off the heating device and stop the reflux extraction; pour the mixed liquid in the reflux extractor into a centrifuge tube; use a centrifuge to perform centrifugal operation to precipitate solid particles and impurities to the bottom of the centrifuge tube to obtain a supernatant; transfer the supernatant in the centrifuge tube to a volumetric flask or other container as the mixed liquid obtained from the first extraction; the remaining medicinal materials can be subjected to secondary extraction or other treatments to fully extract the active ingredients therein.

[0039] Specifically, in step S2, the remaining medicinal residues are added to distilled water and placed in an ultrasonic generating container for ultrasonic extraction, and the supernatant is centrifuged to obtain a second mixed solution. During the reflux extraction process, Codonopsis pilosula and distilled water are added to an ultrasonic generating container. Through the action of ultrasound, the dissolution of the effective ingredients in Codonopsis pilosula can be accelerated and transferred to the solvent. This can improve the extraction efficiency and shorten the extraction time. Among them, extraction is a commonly used separation and extraction technique for separating a target substance from a mixture. In the fields of chemistry and biochemistry, extraction generally refers to using a solvent to extract a target substance from a solid, liquid or gas.

[0040] It is worth noting that ultrasonic generators are commonly used experimental equipment to accelerate mass transfer between the solvent and the medicinal material during reflux extraction. They typically consist of a container and an ultrasonic generator. The operating principle of ultrasonic generators is to use the high-frequency vibrations of ultrasound to produce cavitation and microfluidics, thereby increasing the contact area and mass transfer rate between the solvent and the medicinal material. The ultrasonic vibrations form a large number of tiny bubbles in the liquid, and when the bubbles burst, they generate violent liquid flow, promoting the dissolution and diffusion of solutes in the solvent.

[0041] Accordingly, in one possible implementation, the remaining medicinal residue can be added to distilled water and then placed in an ultrasonic generator container for ultrasonic extraction. The supernatant can then be centrifuged to obtain a second mixed solution. For example, ensure that the ultrasonic generator container and centrifuge are clean and the work surface is tidy. Prepare the required amount of remaining medicinal residue and distilled water; add an appropriate amount of remaining medicinal residue to the ultrasonic generator container. The ratio of medicinal residue to distilled water can be adjusted as needed. Add enough distilled water to the ultrasonic generator container to cover the remaining medicinal residue. Ensure that the medicinal residue is fully immersed in the distilled water. Start the ultrasonic generator and set the appropriate ultrasonic frequency and power. Place the ultrasonic generator in the container to ensure that the ultrasonic waves are fully transmitted into the mixture of medicinal residue and distilled water. The ultrasonic treatment time can be adjusted as needed. After the ultrasonic treatment is complete, place the container in a centrifuge and perform high-speed centrifugation to separate the supernatant and solid residue. The centrifugation time and speed can be adjusted as needed. Carefully remove the container from the centrifuge, taking care not to disturb the sediment. Use a pipette or straw to carefully transfer the supernatant to another container. Thus, a second mixed solution is obtained, which contains the target component extracted from the remaining medicinal residue.

[0042] Specifically, in step S3, the first mixed liquid and the second mixed liquid are mixed and concentrated to obtain an extract. Among them, the extract is a form of pharmaceutical preparation, also known as an extract or extract liquid. It is a liquid preparation made by soaking and extracting medicinal materials and solvents. When preparing an extract, the medicinal materials are usually soaked in a suitable solvent so that the active ingredients in the medicinal materials are dissolved in the solvent to form a concentrated liquid medicine. Extracts are commonly used in traditional Chinese medicine preparations and can be used for oral administration, external use or other appropriate purposes. Compared with other preparation forms, extracts have the following characteristics: Concentration: The extract contains a higher concentration of drug ingredients, so when used, it needs to be appropriately diluted according to the doctor's orders or pharmacist's guidance; Easy to use: The extract can be used directly without further processing, which is convenient for taking or external use; Rapid absorption: Since the drug ingredients in the extract have been dissolved in the solvent, they can be absorbed and utilized by the human body more quickly.

[0043] Accordingly, in one possible implementation, the first and second mixed liquids can be mixed and concentrated to obtain an extract by the following steps: for example, preparing a container suitable for mixing and concentrating, such as a beaker or pot, and a heating device (such as a hot plate or stovetop). A stirrer, thermometer, and timer are also required. The first and second mixed liquids are poured into the prepared container and gently stirred with the stirrer to thoroughly mix. The container is placed on the heating device and the mixture is gradually heated. The temperature should be determined based on the solvent used and the characteristics of the medicinal material. Typically, the temperature is between 50 and 80 degrees Celsius. During the heating process, the mixture is continuously stirred to ensure uniform heating and mixing. A thermometer is used to monitor the temperature to avoid excessive temperatures. As heating proceeds, the solvent gradually evaporates, and the mixture becomes concentrated. The duration of heating can be determined as needed based on the desired concentration. Once the desired concentration is reached, the heating device is turned off and the mixture is allowed to cool to room temperature. The extract is then transferred to a suitable container and properly sealed for storage.

[0044] Specifically, in step S4, Atractylodes macrocephala, Poria cocos, white lentils, coix seed, Chinese yam, Euryale ferox, and lotus seeds are added to purified water for soaking and grinding, and then sucrose is added to obtain a mixture. The mixture is then placed in a chopper and chopped to obtain a slurry. Atractylodes macrocephala is a common Chinese medicinal material belonging to the Asteraceae family. It is a perennial herb that primarily grows in China, Japan, and South Korea. According to Traditional Chinese Medicine theory, Atractylodes macrocephala has the effects of strengthening the spleen and replenishing qi, warming the middle and stopping diarrhea. It is commonly used to treat symptoms such as loss of appetite, abdominal distension, loose stools, and fatigue caused by spleen and stomach deficiency. In addition, Atractylodes macrocephala is also used to regulate menstrual disorders and postpartum weakness in women. Poria cocos, also known as Poria cocos peel or Poria cocos flesh, is a common Chinese medicinal material. It is made from the dried sclerotia and rhizome of the Poria cocos plant, a member of the Poriaceae family. Poria cocos has the effects of strengthening the spleen and eliminating dampness, calming the mind and promoting sleep, and is commonly used to treat symptoms such as spleen and stomach deficiency, edema, and insomnia. It is widely used in Traditional Chinese Medicine. White lentils are known to strengthen the spleen and replenish qi, promote diuresis, and reduce swelling, and have a certain therapeutic effect on symptoms such as spleen and stomach deficiency and edema. Coix seed, also known as coix seed or coix seed, is a common Chinese medicinal herb and edible grain. It is the round seed of the grass plant Coix lachryma-jobi, typically pale yellow or white in color. It has the effects of strengthening the spleen and eliminating dampness, promoting diuresis, and reducing swelling. It can be used to treat symptoms such as spleen and stomach discomfort, edema, and frequent urination caused by damp heat. Coix seed is also believed to have skin-whitening, heat-clearing and detoxifying properties, and is often used in skincare products and beauty foods. Chinese yam, also known as yam or yam, is a common edible plant and Chinese medicinal herb. It is the tuberous portion of the yam plant, also known as the Dioscorea family. It is oblong, smooth, and typically pale yellow or light brown in color. It has the effects of tonifying the lungs and kidneys, nourishing yin and moistening dryness, and strengthening the spleen and relieving diarrhea. It is commonly used to treat symptoms of lung, kidney, and spleen deficiency. Chinese yam can be used alone or in combination with other Chinese medicinal herbs to create traditional Chinese medicine prescriptions. Gorgon fruit, also known as gorgon fruit seeds, gorgon seeds, and lotus seeds, is a common Chinese medicinal herb and edible plant. It is the round or oval, brown seed of the gorgon fruit plant, a member of the Nymphaeaceae family. It has various uses in traditional Chinese medicine, primarily for tonifying the spleen and stomach, and for strengthening and relieving diarrhea. It is believed to strengthen the spleen and kidneys, tonify the lungs, and relieve coughs, and is commonly used to treat symptoms such as spleen and stomach deficiency and coughs caused by lung deficiency. Gorgon fruit can also be used to treat symptoms such as spleen and stomach dampness stagnation, diarrhea, and enuresis.

[0045] It's worth noting that a blender is also known as a food processor or blender. It's typically used to chop, blend, and mix ingredients for cooking and other food preparation. A blender typically consists of a container and an electric motor with blades. When the motor is running, the blades spin rapidly, chopping or blending the ingredients to the desired consistency.

[0046] Accordingly, in a possible implementation method, the following steps can be performed: adding Atractylodes macrocephala, Poria cocos, white lentils, coix seed, Chinese yam, Euryale ferox, and lotus seeds to pure water for soaking and pulping, and then adding sucrose to obtain a mixture, and placing the mixture into a chopper for chopping to obtain a slurry, for example: preparing materials: Atractylodes macrocephala, Poria cocos, white lentils, coix seed, Chinese yam, Euryale ferox, lotus seeds, pure water and an appropriate amount of sucrose; washing Atractylodes macrocephala, Poria cocos, white lentils, coix seed, Chinese yam, Euryale ferox, and lotus seeds separately; placing the washed Atractylodes macrocephala, Poria cocos, white lentils, coix seed, Chinese yam, Euryale ferox, and lotus seeds into a container; adding enough pure water to completely soak the materials therein. The soaking time can be adjusted based on needs and personal preference; 4-6 hours is generally recommended to allow the ingredients to fully absorb the water. Pour the soaked ingredients into a blender or food processor. Start the blender and blend until the ingredients become a paste. The blending time can be adjusted based on personal taste; 1-2 minutes is generally sufficient. While blending, gradually add an appropriate amount of sucrose to adjust the sweetness of the paste. The amount of sucrose can be adjusted to suit your taste. Once blending is complete, pour the resulting paste into a container and use or store it.

[0047] Specifically, in step S5, rice flour is mixed with the extract and the slurry, and magnesium stearate and purified water are added. The mixture is then placed in the chopper for processing to obtain a semi-finished product. Magnesium stearate, with the chemical formula Mg(C18H35O2)2, is also known as magnesium stearate or magnesium stearic acid. It is a white solid and is commonly used as a medicine, food additive, etc. In medicine, magnesium stearate is often used to treat constipation. It can promote intestinal peristalsis by increasing moisture and lubrication in the intestines, thereby alleviating constipation symptoms. In the food industry, magnesium stearate is used as an emulsifier, stabilizer, and anti-caking agent. It helps mix oil and water, increases food stability, and prevents food from clumping. In addition, purified water plays a role in dissolving, diluting, and adjusting concentration during the mixing process. It makes the mixture more fluid, making it easier to mix the rice flour and slurry. Purified water also plays a role in adjusting the overall taste and texture.

[0048] Accordingly, in a possible implementation, the rice flour can be mixed with the extract and the slurry, and then magnesium stearate and pure water can be added, and the mixture can be placed in the chopper for processing to obtain a semi-finished product, for example: pouring the rice flour into a container; pouring the extract and the slurry into the rice flour, stirring evenly to ensure that all ingredients are fully mixed; adding an appropriate amount of magnesium stearate to the mixture. The amount of magnesium stearate can be adjusted according to personal preference and the desired consistency; slowly adding pure water while using a chopper for processing. The amount of pure water can be adjusted as needed to achieve the desired consistency and texture; continue to use the chopper to process the mixture until the desired semi-finished product state is reached. The processing time can be determined according to personal preference and the desired mouthfeel.

[0049] Specifically, in step S6, the semi-finished product is placed in a mold and formed, and then dried by microwave heating to obtain the eight-treasure cake. In particular, in a specific example of the present application, as shown in Figures 2 and 4, step S6 includes: S61, using a camera to collect a drying monitoring video of the semi-finished product for a predetermined time period; S62, analyzing the drying monitoring video to obtain a drying state temporal semantic feature; and S63, determining whether to stop microwave heating based on the drying state temporal semantic feature.

[0050] Correspondingly, the S61 is to collect a drying monitoring video of the semi-finished product for a predetermined time period through a camera. Considering that in the actual preparation process of the Eight Treasures Cake, when the semi-finished product after forming is dried by microwave heating, it is particularly important to monitor the moisture content to control the progress of microwave heating. This is because if the semi-finished product is over-dried, its moisture content will decrease, making the produced Eight Treasures Cake taste too dry and the quality difference between products will also be large. If the semi-finished product is not dried enough, it will lead to a high moisture content, making the Eight Treasures Cake easily affected by moisture, moldy and deteriorating, thereby affecting the product quality of the Eight Treasures Cake. Therefore, in the technical solution of the present application, the moisture content of the semi-finished product is monitored by a camera in real time to stop the drying process in a timely manner. In this way, the moisture content can be controlled within a predetermined range, thereby improving the consistency of the product.

[0051] Accordingly, S62 analyzes the drying monitoring video to obtain a drying state temporal semantic feature. In particular, in a specific example of the present application, as shown in FIG5 , S62 includes: S621 sparsely sampling the drying monitoring video to obtain a plurality of drying monitoring key frames; S622 , performing feature extraction on each of the plurality of drying monitoring key frames using a shallow feature extractor based on a deep neural network model to obtain a plurality of drying state feature vectors; and S623 , performing semantic association encoding on the plurality of drying state feature vectors to obtain a drying state temporal semantic feature vector as the drying state temporal semantic feature.

[0052] Said S621 performs sparse sampling on the drying monitoring video to obtain a plurality of drying monitoring key frames. Considering that the drying monitoring video of the semi-finished product is usually continuous and contains multiple frames of images per second. Such a large amount of video data will be very time-consuming and complicated to process. In the drying process, only images at certain time points can provide useful information, such as the moisture distribution on the surface of the cake. Therefore, in the technical solution of the present application, the drying monitoring video is further sparsely sampled to obtain a plurality of drying monitoring key frames. By performing sparse sampling processing on the original video, a small number of key frames can be selected from the continuous video to better capture key information such as the moisture distribution status on the surface of the cake, while ignoring redundant images at other time points and simplifying the subsequent processing process.

[0053] It is worth noting that sparse sampling is a sampling method used in generative models. Its purpose is to increase the diversity of the model by randomly selecting generated samples. When generating text or other types of data, sparse sampling can prevent the model from being overly biased towards generating high-frequency words or common patterns, thereby producing richer and more diverse results. In sparse sampling, each generated sample has a probability threshold that determines whether it is selected as the output. A higher threshold will result in fewer samples being selected, thereby increasing diversity, but may also cause the quality of the generated samples to decrease. Conversely, a lower threshold will result in more samples being selected, thereby improving the quality of the generated samples, but may lack diversity.

[0054] Accordingly, in one possible implementation, the drying monitoring video can be sparsely sampled to obtain multiple drying monitoring key frames by the following steps: for example, loading the drying monitoring video into a computer or video editing software; determining how many key frames you wish to extract from the video. This will determine the frequency of sparse sampling. For example, if you wish to extract one key frame per second, the sampling frequency will be 1 frame / second; calculating the sampling interval based on the sampling frequency. For example, if the sampling frequency is 1 frame / second, the sampling interval will be 1 second; starting from the start time of the video, extract key frames at the set sampling interval. This can be achieved using the capture function of the video editing software or the video processing library in the programming language; saving the extracted key frames as separate image files or video files. Ensure that each key frame is assigned a unique file name for subsequent use and management; review the extracted key frames to ensure that they accurately represent the content of the drying monitoring video. If necessary, you can adjust the sampling frequency or sampling interval to obtain better results.

[0055] In step S622, a shallow feature extractor based on a deep neural network model is used to extract features from each of the multiple drying monitoring key frames to obtain multiple drying state feature vectors. In the technical solution of the present application, a convolutional neural network model, which excels in extracting implicit features from images, is used to process the multiple drying monitoring key frames to extract characteristic information about the drying state of the cake in each key frame. In particular, when monitoring the drying state of cakes, more attention should be paid to characteristic information about the moisture distribution state on the cake surface, i.e., shallow features of the cake surface. However, when encoding a convolutional neural network, these shallow features become blurred or even overwhelmed by noise as the depth of the convolutional neural network increases. Therefore, in the technical solution of the present application, the multiple drying monitoring key frames are subjected to feature extraction using a shallow feature extractor based on a convolutional neural network model to extract characteristic information about the moisture distribution state on the cake surface from each drying monitoring key frame, thereby obtaining multiple drying state feature vectors.

[0056] It is worth noting that the Convolutional Neural Network (CNN) is a deep learning model widely used in computer vision and image processing. Its design is inspired by the workings of the visual cortex in biology. CNN is primarily used to process data with a grid structure, such as images and videos. It extracts features and performs classification or regression tasks by applying a series of convolution operations and nonlinear activation functions to the input data. The advantage of convolutional neural networks is that they can automatically learn features in the input data and have the characteristics of translation invariance and local connectivity, which make them excellent in image processing tasks. It has achieved many important breakthroughs in fields such as image classification, object detection, and semantic segmentation.

[0057] S623 performs semantic association encoding on the multiple drying state feature vectors to obtain a drying state temporal semantic feature vector as the drying state temporal semantic feature. Considering that during the drying process of the semi-finished product via microwave heating, the drying state features of the cake body will continuously change as the drying process progresses, that is, the moisture distribution state feature information on the cake body surface in each keyframe has a temporal correlation relationship. Therefore, in order to further improve the monitoring of the drying state moisture content during the cake drying process, thereby timely stopping the drying process to control the product moisture content within a predetermined range, thereby improving the product quality and consistency of the eight-treasure cake, in the technical solution of the present application, the multiple drying state feature vectors are further subjected to feature mining through a drying state temporal change feature extractor based on an RNN model to extract global temporal correlation feature information between the moisture distribution state features on the cake body surface in each drying monitoring keyframe, that is, the temporal dynamic change feature information of the moisture state of the cake body surface in the time dimension, thereby obtaining a drying state temporal semantic feature vector.

[0058] It is worth noting that the Recurrent Neural Network (RNN) is a deep learning model suitable for processing sequential data such as text, speech, and time series data. The RNN model is characterized by its ability to capture temporal information in a sequence and pass previous information to subsequent steps. The basic structure of the RNN model is a recurrent unit that receives input and the hidden state of the previous moment and outputs the hidden state of the current moment. This recurrent structure enables the RNN to have memory when processing sequential data, allowing it to use previous information to influence subsequent predictions.

[0059] Accordingly, in one possible implementation, the multiple drying state feature vectors can be passed through a drying state temporal change feature extractor based on an RNN model to obtain the drying state temporal semantic feature vector through the following steps, for example: using a recurrent neural network (RNN) model as a drying state temporal change feature extractor. The RNN model can process sequence data and capture the temporal information of the data; the prepared multiple drying state feature vectors are input into the RNN model to extract the temporal change characteristics of the drying state; when processing sequence data, the RNN model will gradually update the internal state information based on the previous input and the current input, and output a temporal semantic feature vector, which contains the temporal change information of the drying state.

[0060] It is worth mentioning that in other specific examples of the present application, the drying monitoring video can also be analyzed in other ways to obtain the temporal semantic features of the drying state, for example: first, it is necessary to collect monitoring videos containing the drying process. Ensure that the video quality is good and the details of the drying process are clearly visible; pre-process the video, including removing noise, adjusting brightness and contrast, etc., to improve the accuracy of subsequent analysis; decompose the video into a series of frame images. Video processing libraries or software can be used to extract images of each frame; feature extraction is performed on each frame image to capture the temporal semantic features of the drying state. Computer vision techniques, such as convolutional neural networks (CNN) or optical flow methods, can be used to extract image features; temporal analysis is performed on the extracted features to obtain the temporal semantic features of the drying state. Time series analysis methods, such as sliding windows or recurrent neural networks (RNNs), can be used to model and analyze feature sequences; the temporal semantic features are represented in visual or numerical form for further analysis and application. The temporal features of the drying state can be represented in the form of charts, curves, or statistical indicators.

[0061] Correspondingly, the S63 determines whether to stop microwave heating based on the temporal semantic features of the drying state. In particular, in a specific example of the present application, the S63 includes: passing the temporal semantic feature vector of the drying state through a classifier to obtain a classification result, and the classification result is used to indicate whether to stop microwave heating. That is, the classification process is performed based on the temporal change features of the drying state of the cake, so as to stop the drying process in a timely manner based on the real-time changes in the moisture content of the cake, so as to control the moisture content of the product within a predetermined range and improve the quality consistency of the product. Specifically, the drying state temporal semantic feature vector is fully connected encoded using multiple fully connected layers of the classifier to obtain an encoded classification feature vector; and the encoded classification feature vector is passed through the Softmax classification function of the classifier to obtain the classification result.

[0062] A classifier is a machine learning model or algorithm used to classify input data into different categories or labels. Classifiers are part of supervised learning and perform classification tasks by learning the mapping from input data to output categories.

[0063] A fully connected layer is a common layer type in neural networks. In a fully connected layer, each neuron is connected to all neurons in the previous layer, and each connection has a weight. This means that each neuron in a fully connected layer receives input from all neurons in the previous layer, performs a weighted sum of these inputs using the weights, and then passes the result to the next layer.

[0064] The Softmax classification function is a commonly used activation function for multi-classification problems. It converts each element of the input vector into a probability value between 0 and 1, where the sum of these probabilities equals 1. The Softmax function is often used in the output layer of neural networks and is particularly well-suited for multi-classification problems because it maps the network output into a probability distribution for each category. During training, the output of the Softmax function is used to calculate the loss function and update the network parameters through the backpropagation algorithm. It is worth noting that the output of the Softmax function does not change the relative size of the elements; it simply normalizes them. Therefore, the Softmax function does not change the characteristics of the input vector; it simply converts it into a probability distribution.

[0065] It is worth mentioning that in other specific examples of the present application, it is also possible to determine whether to stop microwave heating based on the temporal semantic features of the drying state in other ways, for example: determining the conditions for stopping microwave heating based on specific needs and experimental experience. This may include reaching a certain degree of dryness, reaching a specific drying time, the temporal semantic feature vector of the drying state reaching a certain threshold, etc.; during the microwave heating process, continuously monitoring the changes in the temporal semantic feature vector of the drying state. The latest feature vector can be obtained by collecting data in real time and inputting it into a feature extractor; the monitored temporal semantic feature vector of the drying state is judged according to the set stop condition. If the stop condition is met, microwave heating is stopped; if the stop condition is not met, microwave heating is continued; it can be determined whether to stop microwave heating based on the temporal semantic features of the drying state. In this way, the time of microwave heating can be automatically controlled according to the changes in the drying state to achieve the ideal drying effect.

[0066] It is worth mentioning that in other specific examples of the present application, the semi-finished product can be placed in a mold for shaping by other methods, and the semi-finished product after shaping can be dried by microwave heating to obtain the Eight Treasures Cake, for example: prepare a mold suitable for making Eight Treasures Cake, which can be a silicone mold, a glass container or a microwave-compatible ceramic bowl, etc.; pour the prepared glutinous rice paste into the mold, so that it is fully filled to the appropriate height of the mold; use a scraper or a flat utensil to smooth the surface of the glutinous rice paste; put the mold containing the glutinous rice paste into the microwave; set the appropriate heating time and power according to the power of the microwave oven and the thickness of the glutinous rice paste; generally speaking, you can choose medium power for heating first, and then gradually increase or decrease the heating time according to the actual situation; according to the actual situation , usually the heating time can be about 1-2 minutes each time, and then check the condition of the glutinous rice paste; if the surface of the glutinous rice paste is found to be dry, the heating time can be reduced or the power can be lowered; if the glutinous rice paste is still moist, the heating time can be increased or the power can be increased; after each heating, use a toothpick or bamboo stick to insert the center of the glutinous rice paste to check if there is any adhesion; if the inserted toothpick is clean, it means that the glutinous rice paste is dry and you can proceed to the next step; if there is adhesion on the toothpick, it means that the glutinous rice paste needs to be heated further; when the glutinous rice paste is completely dry, remove the mold and let it cool naturally; during the cooling process, the glutinous rice cake will become more solidified and cured; turn the mold over, tap gently or cut along the edge with a knife to remove the eight-treasure cake from the mold; the eight-treasure cake is ready and can be eaten directly or stored.

[0067] It should be understood that before using the above-mentioned neural network model for inference, it is necessary to train the shallow feature extractor based on the convolutional neural network model, the drying state temporal change feature extractor based on the RNN model, and the classifier. In other words, the automated production method of the spleen-strengthening eight-treasure cake of the present application also includes a training phase for training the shallow feature extractor based on the convolutional neural network model, the drying state temporal change feature extractor based on the RNN model, and the classifier.

[0068] FIG3 is a flow chart of the training phase of the automated production method of the spleen-strengthening eight-treasure cake according to the embodiment of the present application. As shown in FIG3 , the automated production method of the spleen-strengthening eight-treasure cake according to the embodiment of the present application includes: the training phase, including: S110, acquiring training data, the training data including a training drying monitoring video of a semi-finished product in a predetermined time period, and the true value of whether to stop microwave heating; S120, sparsely sampling the training drying monitoring video to obtain a plurality of training drying monitoring key frames; S130, passing the plurality of training drying monitoring key frames through the shallow feature extractor based on the convolutional neural network model to obtain a plurality of training drying state feature vectors; S140, passing the plurality of training drying state feature vectors through the shallow feature extractor based on the convolutional neural network model to obtain a plurality of training drying state feature vectors; The dry state temporal change feature extractor based on the RNN model is passed through to obtain a training dry state temporal semantic feature vector; S150, the training dry state temporal semantic feature vector is passed through the classifier to obtain a classification loss function value; and, S160, based on the classification loss function value and through the direction propagation of gradient descent, the shallow feature extractor based on the convolutional neural network model, the dry state temporal change feature extractor based on the RNN model and the classifier are trained, wherein, in each round of iteration of the training, the weight matrix of the classifier is optimized for cross-domain attention transfer of feature distribution.

[0069] In particular, in the technical solution of the present application, when the multiple drying monitoring key frames are respectively passed through a shallow feature extractor based on a convolutional neural network model to obtain multiple drying state feature vectors, each drying state feature vector expresses the image semantic features of the drying monitoring key frame, so that the temporal correlation between the image semantic features of the drying monitoring key frames at different time points can be further extracted through a drying state temporal change feature extractor based on an RNN model. However, considering that in the case of sparse sampling, the multiple drying monitoring key frames are uneven in temporal distribution, and the image semantic distribution of each drying monitoring key frame itself in the temporal direction is also uneven, therefore, although the temporal correlation is extracted, the drying state temporal semantic feature vector still has an uneven local feature distribution. In this way, when the dry state temporal semantic feature vector is classified by a classifier, the difference in distribution transferability of the uneven local feature distribution in the domain transfer process of classification is taken into account. For example, when the weight matrix of the classifier is adapted to the local temporal association feature representation of the image semantic feature with high temporal correlation, it will have better distribution transferability than the local temporal association feature representation of the image semantic feature with low temporal correlation, and vice versa. Therefore, it is necessary to adaptively optimize the feature vector to be classified with respect to the weight matrix of the classifier, so as to improve the training effect of the classification training of the dry state temporal semantic feature vector by the classifier, that is, to improve the classification speed and the accuracy of the classification results obtained. Therefore, the applicant of the present application performs feature transfer optimization based on feature distribution cross-domain attention on the weight matrix M during each iteration of the weight matrix of the classifier, which is specifically expressed as:

[0070] Where M is the weight matrix of the classifier, the scale of M is m×m, V1 to V m is the m row vector of the weight matrix M, ||·||2 represents the bi-norm of the eigenvector, (:∑ j m i，j :) is a row vector obtained by arranging the sum of each row vector of the weight matrix M, and cov1(·) and cov2(· both represent single-layer convolution operations. Represents matrix multiplication, and M' represents the weight matrix of the classifier after iteration. Here, the feature transfer optimization based on feature distribution cross-domain attention targets the different representations of the feature distribution of the dry state temporal semantic feature vector in the feature space domain and the classification target domain. Based on the cross-domain diversity feature representation of the weight matrix M of the classifier relative to the dry state temporal semantic feature vector to be classified, attention is paid to the spatial structured feature distribution of the weight matrix M through a convolution operation to enhance the transferability of the cross-domain gaps of the good transfer feature distribution in the diversified feature distribution, while suppressing the negative transfer of the poor transfer feature distribution. Based on the distribution structure of the weight matrix M itself relative to the dry state temporal semantic feature vector to be classified, the unsupervised domain transfer adaptive optimization of the weight matrix M is achieved, thereby improving the training effect of the dry state temporal semantic feature vector through the classifier for classification training. In this way, the chopping speed value of the chopping machine can be adaptively adjusted based on the changes in the slurry state of the mixture, thereby optimizing the chopping efficiency and quality of the raw material mixture, thereby improving the absorbability of the raw material ingredients and the taste of the eight-treasure cake product. In this way, the drying process can be stopped in a timely manner based on the real-time change of the moisture content of the cake body to ensure that the moisture content of the eight-treasure cake product can be controlled within a predetermined range, thereby improving the quality consistency of the produced eight-treasure cake products.

[0071] In summary, the automated production method for the spleen-strengthening eight-treasure cake according to the embodiment of the present application is described. During the drying process of the semi-finished product by microwave heating, machine vision technology is used to monitor the real-time changes in the moisture content of the semi-finished product and to stop the drying process at the appropriate time. In this way, the drying process can be stopped at the appropriate time based on the real-time changes in the moisture content of the cake, ensuring that the moisture content of the eight-treasure cake product is controlled within a predetermined range, thereby improving the quality consistency of the produced eight-treasure cake products.

[0072] Furthermore, an automated production system for the spleen-strengthening eight-treasure cake is also provided.

[0073] Figure 6 is a block diagram of an automated production system for a spleen-strengthening eight-treasure cake according to an embodiment of the present application. As shown in Figure 6, the automated production system 300 for a spleen-strengthening eight-treasure cake according to an embodiment of the present application includes: a video acquisition module 310 for capturing, via a camera, drying monitoring videos of semi-finished products over a predetermined time period; a temporal semantic analysis module 320 for analyzing the drying monitoring videos to obtain temporal semantic features of the drying state; and a control result generation module 330 for determining whether to stop microwave heating based on the temporal semantic features of the drying state.

[0074] As described above, the automated production system 300 for the spleen-strengthening eight-treasure cake according to the embodiment of the present application can be implemented in various wireless terminals, such as a server having an automated production algorithm for the spleen-strengthening eight-treasure cake. In one possible implementation, the automated production system 300 for the spleen-strengthening eight-treasure cake according to the embodiment of the present application can be integrated into a wireless terminal as a software module and / or a hardware module. For example, the automated production system 300 for the spleen-strengthening eight-treasure cake can be a software module in the operating system of the wireless terminal, or can be an application developed for the wireless terminal; of course, the automated production system 300 for the spleen-strengthening eight-treasure cake can also be one of the many hardware modules of the wireless terminal.

[0075] Alternatively, in another example, the automated production system 300 for the spleen-strengthening eight-treasure cake and the wireless terminal may be separate devices, and the automated production system 300 for the spleen-strengthening eight-treasure cake may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0076] FIG7 is a schematic diagram of a scenario for an automated production method for a spleen-strengthening eight-treasure cake according to an embodiment of the present application. As shown in FIG7 , in this application scenario, a camera (e.g., C as shown in FIG7 ) is used to obtain a drying monitoring video of a semi-finished product for a predetermined time period. Subsequently, the monitoring video is input into a server (e.g., S in FIG7 ) that is equipped with an automated production algorithm for the spleen-strengthening eight-treasure cake. The server is capable of processing the input monitoring video using the automated production algorithm for the spleen-strengthening eight-treasure cake to generate a classification result indicating whether to stop microwave heating.

[0077] While various embodiments of the present disclosure have been described above, the above descriptions are illustrative, non-exhaustive, and not intended to be limiting of the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, their practical applications, or improvements to existing technologies, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. An automated production method for spleen-strengthening eight-treasure cake, characterized in that: include: After adding the dang ginseng into distilled water, reflux extraction is performed at a predetermined temperature, and the supernatant is taken by centrifugation to obtain a first mixed solution, and the remaining medicinal residue is obtained at the same time; Adding distilled water to the remaining medicinal residues and placing them in an ultrasonic generator container for ultrasonic extraction, and centrifuging to obtain a supernatant to obtain a second mixed solution; Mixing and concentrating the first mixed solution and the second mixed solution to obtain an extract; Adding Atractylodes macrocephala, Poria cocos, white hyacinth bean, coix seed, yam, Euryale ferox, and lotus seed into purified water for soaking and grinding, then adding sucrose to obtain a mixture, and putting the mixture into a chopping machine for chopping to obtain a slurry; The rice flour is mixed with the extract and the slurry, and magnesium stearate and purified water are added thereto, and the mixture is placed in the chopping machine for processing to obtain a semi-finished product; as well as The semi-finished product is placed in a mold for forming, and the formed semi-finished product is dried by microwave heating to obtain the eight-treasure cake.

2. The automated production method of the spleen-strengthening eight-treasure cake according to claim 1, characterized in that: The semi-finished product is placed in a mold for forming, and the semi-finished product is dried by microwave heating to obtain the eight-treasure cake, comprising: Collect the drying monitoring video of the semi-finished product in a predetermined time period through the camera; Analyzing the drying monitoring video to obtain a drying state temporal semantic feature; and Based on the drying state temporal semantic features, it is determined whether to stop microwave heating.

3. The automated production method of the spleen-strengthening eight-treasure cake according to claim 2, characterized in that: The drying monitoring video is analyzed to obtain the drying state temporal semantic features, including: Sparsely sampling the drying monitoring video to obtain a plurality of drying monitoring key frames; Performing feature extraction on the plurality of drying monitoring key frames respectively by a shallow feature extractor based on a deep neural network model to obtain a plurality of drying state feature vectors; and The plurality of dry state feature vectors are semantically associated encoded to obtain a dry state temporal semantic feature vector as the dry state temporal semantic feature.

4. The automated production method of the spleen-strengthening eight-treasure cake according to claim 3, characterized in that: The deep neural network model is a convolutional neural network model.

5. The automated production method of the spleen-strengthening eight-treasure cake according to claim 4, characterized in that: The plurality of dry state feature vectors are semantically associated encoded to obtain a dry state temporal semantic feature vector as the dry state temporal semantic feature, comprising: passing the plurality of dry state feature vectors through a dry state temporal change feature extractor based on an RNN model to obtain the dry state temporal semantic feature vector.

6. The automated production method of the spleen-strengthening eight-treasure cake according to claim 5, characterized in that: Determining whether to stop microwave heating based on the drying state temporal semantic feature includes: passing the drying state temporal semantic feature vector through a classifier to obtain a classification result, wherein the classification result is used to indicate whether to stop microwave heating.

7. The automated production method of the spleen-strengthening eight-treasure cake according to claim 6, characterized in that: It also includes a training step: for training the shallow feature extractor based on the convolutional neural network model, the dry state temporal change feature extractor based on the RNN model, and the classifier.

8. The automated production method of the spleen-strengthening eight-treasure cake according to claim 7, characterized in that: The training step comprises: Acquiring training data, the training data comprising a training drying monitoring video of a semi-finished product in a predetermined time period, and a true value of whether to stop microwave heating; Sparsely sampling the training drying monitoring video to obtain a plurality of training drying monitoring key frames; Passing the plurality of training drying monitoring key frames through the shallow feature extractor based on the convolutional neural network model respectively to obtain a plurality of training drying state feature vectors; Passing the plurality of training dry state feature vectors through the dry state temporal change feature extractor based on the RNN model to obtain a training dry state temporal semantic feature vector; Passing the training dry state temporal semantic feature vector through the classifier to obtain a classification loss function value; and The shallow feature extractor based on the convolutional neural network model, the dry state temporal change feature extractor based on the RNN model and the classifier are trained based on the classification loss function value and propagated in the direction of gradient descent, wherein in each round of iteration of the training, the weight matrix of the classifier is optimized for cross-domain attention transfer of feature distribution.

9. The automated production method of the spleen-strengthening eight-treasure cake according to claim 8, characterized in that: In each iteration of the training, the weight matrix of the classifier is optimized for cross-domain attention transfer of feature distribution using the following optimization formula; Wherein, the optimization formula is: Where M is the weight matrix of the classifier, the scale of M is m×m, V1 to V m is the The m row vectors of the weight matrix M, ||·||2 represents the bi-norm of the eigenvector, (:∑ j m i,j :) is a row vector obtained by arranging the sum of each row vector of the weight matrix M, and cov1(·) and cov2(·) both represent single-layer convolution operations, represents matrix multiplication, and M' represents the weight matrix of the classifier after iteration.

10. An automated production system for spleen-strengthening eight-treasure cake, characterized in that: include: A video acquisition module, used to collect the drying monitoring video of the semi-finished product in a predetermined time period through a camera; A temporal semantic analysis module, used for analyzing the drying monitoring video to obtain the temporal semantic features of the drying state; as well as The control result generation module is used to determine whether to stop microwave heating based on the drying state temporal semantic features.

Images