Intelligent control system of medical ice maker
By using high-definition cameras and depth prediction models in medical ice makers, the flip cycle and blowing parameters of medical cartridges are optimized, and the problem of high power consumption of existing medical ice makers is solved, achieving more efficient ice sludge preparation and lower power consumption.
Patent Information
- Application Number
- CN202510539856.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-27
- Publication Date
- 2025-06-10
AI Technical Summary
The existing medical ice makers have large power consumption and insufficient economical efficiency due to the short flip cycle setting and excessive blowing parameters setting.
High-definition cameras are used to capture high-definition images of normal saline in the medical cylinder, and the ice body dispersed state and dynamic volume characteristics are extracted through the convolution network, and the depth prediction model is called to predict the next flip time and refrigeration parameters of the medical cylinder, and the flip cycle and blowing parameters are optimized.
The ice sludge preparation efficiency of medical ice makers has been improved and the power consumption of medical ice makers has been significantly reduced.
Smart Images
Figure CN120120784A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of ice makers, and more particularly, to an intelligent control system for a medical ice maker. Background Art
[0002] Medical ice makers are mainly used to produce medical ice (ice slurry), and the produced medical ice can be used for attaching and freezing medical freezing scenarios (such as human organs). Medical ice makers mainly freeze through the structure of an upper freezing cavity and a lower refrigeration module. Among them, a plurality of horizontally rotating detachable medical cylinders are arranged in the upper freezing cavity, and physiological saline is contained in the medical cylinders. During the preparation of medical ice, the medical cylinders rotate around their axes, so that the physiological saline in the cylinders will not adhere. To further prevent the adhesion of water bodies, the medical cylinders will be periodically flipped, so that the ice bodies of the physiological saline in the medical cylinders remain in a dispersed state. A plurality of air ducts parallel to the medical cylinders are arranged in the refrigeration cavity, and each air duct can independently blow cold air to the corresponding medical cylinder, so that the physiological saline inside the medical cylinder freezes into ice slurry when rotating.
[0003] In actual production, to ensure that the ice bodies of the physiological saline in the medical cylinders remain in a dispersed state, the flipping period of the medical cylinders is set to be short, but this will significantly increase the power consumption of the medical ice maker; and, for the setting of the number of enabled air ducts and the blowing intensity, currently it is basically an "excessive" strategy. The above two reasons lead to poor economy of the medical ice maker, and this technical problem needs to be further improved. Summary of the Invention
[0004] In view of this, the present invention provides an intelligent control system for a medical ice maker, an electronic device, a computer storage medium, and a computer program product to solve the above technical problems.
[0005] The present invention discloses an intelligent control system for a medical ice maker. The system includes a high-definition camera, a controller, and a communicator. The controller is respectively connected to the high-definition camera and the communicator. The high-definition camera is used to capture a high-definition image of the physiological saline in the medical cylinder and transmit the high-definition image to the controller. The controller is used to call a convolutional network to extract the ice body dispersion state characteristics from the high-definition image, call a first depth prediction model to perform prediction processing on the ice body dispersion state characteristics, obtain the next flipping time of the medical cylinder, and generate a first control instruction according to the next flipping time. The controller is further used to call a convolutional network to extract the dynamic volume characteristics of the physiological saline from the high-definition image, predict the volume of the physiological saline according to the dynamic volume characteristics and the size information of the medical cylinder, call a second depth prediction model to perform prediction processing on the volume, obtain refrigeration parameters, where the refrigeration parameters include the number of air ducts opened and the cold air wind speed, and generate a second control instruction according to the refrigeration parameters. The communicator is used to send the first control instruction to the flipping mechanism of the medical cylinder for execution, and send the second control instruction to the blowing mechanism of the medical cylinder for execution.
[0006] In some embodiments, both the first depth prediction model and the second depth prediction model are constructed based on a deep learning algorithm, and the deep learning algorithm is any one of CNN, LSTM, GRU, generative adversarial network, and Transformer.
[0007] In some embodiments, the data format of the first training data for the first depth prediction model is: [ice body dispersion state characteristics, flipping duration, first ice body dispersion state evaluation label]; the data format of the second training data for the second depth prediction model is: [volume of physiological saline, number of air ducts opened, cold air wind speed, second ice body dispersion state evaluation label]; the first depth prediction model and the second depth prediction model are respectively trained using the first training data and the second training data until the training reaches the standard.
[0008] In some embodiments, the first depth prediction model and the second depth prediction model are trained using a distributed training method.
[0009] In some embodiments, the distributed training method is implemented based on a blockchain or a consortium chain. The blockchain or the consortium chain includes several auxiliary training nodes, and each of the auxiliary training nodes collects training data by itself and trains the first depth prediction model and the second depth prediction model.
[0010] In some embodiments, the calling of the convolutional network to extract the dynamic volume features of the normal saline from the high-definition image and predicting the volume of the normal saline based on the dynamic volume features and the size information of the medical cylinder includes: receiving a plurality of the high-definition images transmitted by the high-definition camera and their corresponding shooting serial numbers, predicting the target shooting serial number in a stable rotation state according to the shooting serial numbers, and calling the convolutional network to extract the dynamic volume features of the normal saline from the high-definition image corresponding to the target shooting serial number; importing the dynamic volume features into a volume prediction model, and the volume prediction model predicts the volume of the normal saline in the medical cylinder.
[0011] In some embodiments, the predicting the target shooting serial number in a stable rotation state according to the shooting serial numbers includes: statistically analyzing the historical ice-making data of the medical ice maker to obtain the regular volume of the normal saline, and predicting the target shooting serial number in a stable rotation state of the medical cylinder according to the regular volume, the rated parameters of the self-rotation mechanism of the medical ice maker, and the shooting frequency of the high-definition camera.
[0012] The present invention also discloses an electronic device, which is applied to the intelligent control system of the medical ice maker described in any one of the foregoing; the electronic device includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor.
[0013] The present invention also discloses a computer storage medium, which is applied to the intelligent control system of the medical ice maker described in any one of the foregoing; the computer-readable storage medium stores a computer program.
[0014] The present invention also discloses a computer program product, which is applied to the intelligent control system of the medical ice maker described in any one of the foregoing; the computer program product is pre-packaged with computer program codes.
[0015] The beneficial effects of the present invention are as follows: The solution of the present invention optimizes the next flipping time of the medical cylinder, the number of air duct openings during blowing, and the cold air speed, improves the ice sludge preparation efficiency of the medical ice maker, and significantly reduces the power consumption of the medical ice maker. Description of the Drawings
[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.
[0017] Figure 1It is a schematic structural diagram of an intelligent control system for a medical ice maker according to an embodiment of the present invention.
[0018] Figure 2 It is a schematic diagram of the internal structure of the controller according to an embodiment of the present invention. Specific embodiments
[0019] The following specific embodiments illustrate the implementation manners of the present application. Those skilled in the art can easily understand other advantages and effects of the present application from the content disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts belong to the scope of protection of the present application.
[0020] In addition, the technical features involved in different implementation manners of the present application described below can be combined with each other as long as they do not conflict with each other.
[0021] As Figure 1 shown, an embodiment of the present invention discloses an intelligent control system for a medical ice maker. The system includes a high-definition camera, a controller, and a communicator. The controller is respectively connected to the high-definition camera and the communicator. The high-definition camera is configured to capture a high-definition image of the physiological saline in the medical cylinder and transmit the high-definition image to the controller. The controller is configured to call a convolutional network to extract the ice dispersion state characteristics of the physiological saline from the high-definition image, call a first depth prediction model to perform a prediction process on the ice dispersion state characteristics, obtain the next flipping time of the medical cylinder, and generate a first control instruction according to the next flipping time. The controller is further configured to call a convolutional network to extract the dynamic volume characteristics of the physiological saline from the high-definition image, predict the volume of the physiological saline according to the dynamic volume characteristics and the size information of the medical cylinder, call a second depth prediction model to perform a prediction process on the volume, obtain refrigeration parameters, where the refrigeration parameters include the number of opened air ducts and the cold air velocity, and generate a second control instruction according to the refrigeration parameters. The communicator is configured to send the first control instruction to the flipping mechanism of the medical cylinder for execution, and send the second control instruction to the blowing mechanism of the medical cylinder for execution.
[0022] As mentioned in the background art, in the existing medical ice maker, because the flipping cycle is set too short, and the number of opened air ducts and the cold air velocity of the blowing mechanism are set according to the "excessive" strategy, the power consumption of the medical ice maker is relatively large and the economy is insufficient. To solve this technical problem, as Figure 2As shown, the present invention pre - constructs a first depth prediction model and a second depth prediction model, which are respectively used to predict the optimal next flipping time of the medical cylinder, the optimal number of air duct openings for blowing air on the medical cylinder, and the cold air velocity, so as to prepare ice mud that meets the requirements with as low power consumption as possible.
[0023] Specifically: A high - definition camera is set at a suitable position inside or outside the medical cylinder of the medical ice maker to capture the state of the physiological saline ice body inside the medical cylinder in real time.
[0024] The controller first calls the convolutional network to extract the characteristics of the dispersed state of the physiological saline ice body from the high - definition image, calls the first depth prediction model to predict the next flipping time of the medical cylinder based on this, and accordingly controls the flipping mechanism of the medical cylinder to flip the medical cylinder when reaching this next flipping time to keep the ice body inside the medical cylinder in a dispersed state.
[0025] The controller also calls the convolutional network to extract the dynamic volume characteristics of the physiological saline from the high - definition image, that is, the characteristics of the physiological saline relative to the cylinder body of the medical cylinder when the medical cylinder is in a self - rotating state, such as the average height reached by the physiological saline, the depth of the middle vortex, etc. According to this dynamic volume characteristic and the known size information of the medical cylinder, the volume of the physiological saline can be predicted. Then, the second depth prediction model is called to perform prediction processing on the volume to obtain the optimal refrigeration parameters, including the number of air duct openings and the cold air velocity. Accordingly, the blowing mechanism of the medical cylinder is controlled to open the appropriate number of air ducts according to this refrigeration parameter and control the air velocity of each air duct to the above - mentioned cold air velocity.
[0026] Therefore, the solution of the present invention optimizes the next flipping time of the medical cylinder, the number of air duct openings during blowing, and the cold air velocity, improves the ice mud preparation efficiency of the medical ice maker, and significantly reduces the power consumption of the medical ice maker.
[0027] It should be noted that the order of generation of the above - mentioned first control instruction and second control instruction may not be specifically limited. Either the first control instruction can be generated first, or the second control instruction can be generated first. Generally speaking, the first control instruction is generated multiple times during the preparation process of the ice mud, that is, it is necessary to intermittently control the medical cylinder to flip multiple times, while the second control instruction is basically generated at the initial stage of the ice mud preparation and controls the blowing mechanism to execute after generation until the ice mud preparation is completed.
[0028] In some embodiments, both the first depth prediction model and the second depth prediction model are constructed based on deep - learning algorithms, and the deep - learning algorithm is any one of CNN, LSTM, GRU, generative adversarial network, and Transformer.
[0029] In an embodiment of the present invention, the present invention uses CNN, LSTM, GRU, generative adversarial network or Transformer to construct the above-mentioned first depth prediction model and second depth prediction model. The two models can use the same deep learning algorithm or different deep learning algorithms, and the present invention does not limit this.
[0030] In some embodiments, the data format of the first training data of the first depth prediction model is: [ice body dispersion state feature, opening and flipping duration, first ice body dispersion state evaluation label]; the data format of the second training data of the second depth prediction model is: [volume of physiological saline, number of air ducts opened, cold air velocity, second ice body dispersion state evaluation label]; the first depth prediction model and the second depth prediction model are trained respectively using the first training data and the second training data until the training reaches the standard.
[0031] In an embodiment of the present invention, since the first depth prediction model is used to determine the next flipping time of the medical cylinder, its training data includes ice body dispersion state features, opening and flipping duration, and the first ice body dispersion state evaluation label. Among them, the ice body dispersion state features and the opening and flipping duration are the entity data of the training data, which are used to directly train the first depth prediction model, while the first ice body dispersion state evaluation label is used to analyze the deviation between the input data and the predicted output data, and then guide the training direction of the first depth prediction model. The first ice body dispersion state evaluation label is evaluated by the staff based on the actual test results. Specifically, when the ice body in the medical cylinder is in a certain dispersion state, the flipping is started after the first duration, the second duration,..., the Nth duration respectively. The flipping duration is preset (the single flipping duration of the flipping mechanism is also the same when the medical ice maker prepares ice mud). After the flipping is completed, the dispersion state of the ice body in the medical cylinder is evaluated. This evaluation includes at least several aspects such as the ice body particle size, ice body distribution density, ice body uniformity, etc. The ice body dispersion state score is obtained by comprehensive scoring. The higher the ice body dispersion state score, the better the dispersion state of the ice body in the medical cylinder after flipping, and vice versa. Then, the corresponding first ice body dispersion state evaluation label is determined based on the ice body dispersion state score. Repeat the above steps to obtain multiple pieces of first training data.
[0032] Since the second depth prediction model is used to determine the number of air duct openings and the cold air velocity of the air blowing mechanism of the medical cylinder, the training data thereof includes the volume of physiological saline, the number of air duct openings, the cold air velocity, and the second ice mass dispersion state evaluation label. Among them, the number of air duct openings and the cold air velocity are the entity data of the training data, which are used to directly train the second depth prediction model, while the second ice mass dispersion state evaluation label is used to analyze the deviation between the input data and the predicted output data, and then guide the training direction of the second depth prediction model. The second ice mass dispersion state evaluation label can also be evaluated by the staff based on the actual test results. Specifically, a certain volume of physiological saline is filled into the medical cylinder, and multiple groups of the number of air duct openings and the cold air velocity are determined. The physiological saline in the medical cylinder is blown with cold air according to each group of the number of air duct openings and the cold air velocity. The cold air blowing duration is a preset value (the single blowing duration of the flipping mechanism can also be the same when the medical ice maker prepares ice mud). After reaching the cold air blowing duration, the dispersion state of the ice mass, that is, the ice mud, in the medical cylinder is evaluated. This evaluation also includes at least several aspects such as the size of the ice mass particles, the distribution density of the ice mass, and the uniformity of the ice mass. The ice mass dispersion state score is obtained by comprehensive scoring. The higher the ice mass dispersion state score, the better the dispersion state of the ice mass in the medical cylinder after cold air blowing, and vice versa. Then, the corresponding second ice mass dispersion state evaluation label is determined according to the ice mass dispersion state score. Repeating the above steps, multiple pieces of second training data are obtained.
[0033] Regarding the evaluation indexes of the dispersion state of the ice mud, the explanations are as follows: Size of the ice mass particles: During the condensation process of the ice mass, particles of different sizes may be formed. Larger particles may indicate that the ice mass condenses more loosely, while smaller particles may indicate that the ice mass condenses more tightly.
[0034] Distribution density of the ice mass: The distribution density of the ice mass in the medical cylinder is also an important feature. A higher distribution density may mean that the gaps between the ice masses are smaller, while a lower distribution density may indicate that there are larger voids between the ice masses.
[0035] Uniformity of the ice mass: Whether the distribution of the ice mass in the medical cylinder is uniform is also one of its features. A uniform distribution may indicate that the refrigeration system works better, while a non-uniform distribution may indicate problems such as uneven refrigeration or ice mass adhesion.
[0036] Correspondingly, the above-mentioned ice mass dispersion state features also include features such as the size of the ice mass particles, the distribution density of the ice mass, and the uniformity of the ice mass.
[0037] In some embodiments, the first depth prediction model and the second depth prediction model are trained in a distributed training manner.
[0038] In some embodiments, the distributed training method is implemented based on a blockchain or a consortium blockchain, and several auxiliary training nodes are included in the blockchain or the consortium blockchain. Each of the auxiliary training nodes collects training data by itself and trains the first depth prediction model and the second depth prediction model.
[0039] In the embodiments of the present invention, in view of the certain difficulty in obtaining the first training data and the second training data, therefore, the present invention adopts a distributed training method, that is, the first depth prediction model and the second depth prediction model are published to a blockchain or a consortium blockchain, and the auxiliary training nodes therein collect training data by themselves according to the above format and perform separate training respectively. Each auxiliary training node feeds back the key parameters of the trained model to the local node, and the local node performs fusion processing on each key parameter, that is, obtains preliminary key parameters, and then fills the preliminary key parameters into the blank first depth prediction model and the second depth prediction model locally, and then uses the above first training data and the second training data locally to train the two models until the training reaches the standard. Among them, the key parameters of the model vary according to the different deep learning algorithms adopted by the model, and will not be elaborated here.
[0040] In some embodiments, the method of calling a convolutional network to extract the dynamic volume feature of physiological saline from the high-definition image and predicting the volume of physiological saline according to the dynamic volume feature and the size information of the medical cylinder includes: receiving a plurality of the high-definition images transmitted by the high-definition camera and their corresponding shooting serial numbers, predicting a target shooting serial number in a stable rotation state according to the shooting serial numbers, and calling a convolutional network to extract the dynamic volume feature of physiological saline from the high-definition image corresponding to the target shooting serial number; importing the dynamic volume feature into a volume prediction model, and the volume prediction model predicts the volume of physiological saline in the medical cylinder.
[0041] In the embodiments of the present invention, after the medical cylinder starts to rotate, the high-definition camera takes a plurality of high-definition images at a preset shooting frequency, and each high-definition image corresponds to a shooting serial number. The controller can predict a target shooting serial number in a stable rotation state according to the size of the shooting serial number, and call a convolutional network to extract the dynamic volume feature of physiological saline only from one high-definition image in a stable rotation state, which is subsequently used to analyze the volume of physiological saline contained in the medical cylinder. The reason for such a setting is that before the medical cylinder reaches a stable rotation, the dynamic characteristics of the physiological saline in the medical cylinder are unstable, while after the medical cylinder reaches a stable rotation, the dynamic characteristics of the physiological saline are basically stable. At this time, the extracted dynamic volume feature has a higher correlation with the volume of the physiological saline in the medical cylinder, and the accuracy of the subsequently predicted volume is also higher.
[0042] Meanwhile, asFigure 2 As shown, a volume prediction model can also be pre - constructed, which can be constructed based on conventional classification algorithms, such as decision trees, SVM, etc. Since the models constructed by such classification algorithms are small in volume and faster in classification, they are suitable for quickly predicting the volume of normal saline in the medical cylinder.
[0043] In some embodiments, predicting the target shooting serial number in a stable rotation state according to the shooting serial number includes: statistically analyzing the historical ice - making data of the medical ice - maker to obtain the regular volume of normal saline, and predicting the target shooting serial number of the medical cylinder in a stable rotation state according to the regular volume, the rated parameters of the self - rotating mechanism of the medical ice - maker, and the shooting frequency of the high - definition camera.
[0044] In the embodiments of the present invention, the time from the start of the medical cylinder to the stable rotation state is mainly affected by the rated parameters of the self - rotating mechanism of the medical ice - maker and the volume of normal saline contained in the medical cylinder. The above - mentioned rated parameters are known, and the volume of normal saline contained in the medical cylinder can be statistically obtained according to multiple historical ice - making data of the ice - maker in a recent period. Of course, a certain adjustment coefficient (greater than 1, such as 1.2, 1.5) can also be multiplied by the statistical value to improve its accuracy. At the same time, since the shooting frequency of the high - definition camera is also known, the target shooting serial number of the medical cylinder in a stable rotation state can be predicted based on the above three factors.
[0045] Of course, the prediction of the target shooting serial number can also be realized based on a specific model, which can also be constructed based on the aforementioned deep - learning algorithm, and the present invention does not make specific limitations on this.
[0046] The embodiments of the present invention also disclose an electronic device, which is applied to the intelligent control system of the medical ice - maker described in any one of the foregoing; the electronic device includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor.
[0047] The embodiments of the present invention also disclose a computer storage medium, which is applied to the intelligent control system of the medical ice - maker described in any one of the foregoing; the computer - readable storage medium stores a computer program.
[0048] The embodiments of the present invention also disclose a computer program product, which is applied to the intelligent control system of the medical ice - maker described in any one of the foregoing; the computer program product is pre - packaged with computer program code.
[0049] It should be noted that in the description of the present invention, the terms "first", "second", etc. are only used for descriptive purposes and cannot be construed as indicating or implying relative importance. In addition, in the description of the present invention, unless otherwise specified, the meaning of "a plurality of" is two or more.
[0050] Any process or method description shown in a flowchart or described otherwise herein can be understood to represent a module, segment, or portion of code including one or more executable instructions for implementing a specific logical function or process. The scope of the preferred embodiments of the present invention includes additional implementations, where the functions may be executed in a substantially simultaneous manner or in an order opposite to that shown or discussed, according to the functions involved, which should be understood by those skilled in the technical field of the embodiments of the present invention.
[0051] It should be understood that each part of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits with logic gate circuits for implementing logical functions on data signals, application specific integrated circuits with appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.
[0052] Those of ordinary skill in the art of the present technology can understand that all or part of the steps carried by the method of the above embodiments can be completed by instructing relevant hardware through a program. The program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiments.
[0053] In addition, each functional unit in various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically alone, or two or more units can be integrated into one module. The above integrated module can be implemented in the form of hardware or in the form of a software functional module. When the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.
[0054] The above-mentioned storage medium can be a read-only memory, a magnetic disk, an optical disk, etc.
[0055] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "examples", "specific examples", or "some examples", etc., mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic descriptions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.
[0056] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.
Claims
1. An intelligent control system for a medical ice maker, characterized in that: The system includes a high-definition camera, a controller and a communicator, wherein the controller is connected to the high-definition camera and the communicator respectively; the high-definition camera is used to take a high-definition image of the physiological saline in the medical cylinder and transmit the high-definition image to the controller; the controller is used to call a convolutional network to extract the ice dispersion state characteristics of the physiological saline from the high-definition image, call a first depth prediction model to predict the ice dispersion state characteristics, obtain the next flip time of the medical cylinder, and generate a first control instruction according to the next flip time; the controller is also used to call a convolutional network to extract the dynamic volume characteristics of the physiological saline from the high-definition image, predict the volume of the physiological saline according to the dynamic volume characteristics and the size information of the medical cylinder, call a second depth prediction model to predict the volume, obtain refrigeration parameters, the refrigeration parameters include the number of air ducts opened and the cold wind speed, and generate a second control instruction according to the refrigeration parameters; the communicator is used to send the first control instruction to the flip mechanism of the medical cylinder for execution, and send the second control instruction to the blower mechanism of the medical cylinder for execution.
2. The intelligent control system for a medical ice machine according to claim 1, characterized in that: The first depth prediction model and the second depth prediction model are both constructed based on a deep learning algorithm, and the deep learning algorithm is any one of CNN, LSTM, GRU, generative adversarial network, and Transformer.
3. The intelligent control system for a medical ice maker according to claim 2, characterized in that: The data format of the first training data of the first depth prediction model is: [ice body dispersion state characteristics, opening and flipping time, first ice body dispersion state evaluation label]; the data format of the second training data of the second depth prediction model is: [volume of physiological saline, number of air ducts opened, cold air speed, second ice body dispersion state evaluation label]; the first training data and the second training data are used to train the first depth prediction model and the second depth prediction model respectively until the training meets the standards.
4. The intelligent control system for a medical ice maker according to claim 1, characterized in that: The first depth prediction model and the second depth prediction model are trained in a distributed training manner.
5. The intelligent control system for a medical ice maker according to claim 4, characterized in that: The distributed training method is implemented based on a blockchain or a consortium chain. The blockchain or the consortium chain includes a number of auxiliary training nodes. Each of the auxiliary training nodes collects training data on its own and trains the first depth prediction model and the second depth prediction model.
6. The intelligent control system for a medical ice maker according to claim 1, characterized in that: The calling of the convolutional network to extract the dynamic volume feature of the saline solution from the high-definition image, and predicting the volume of the saline solution according to the dynamic volume feature and the size information of the medical cylinder, comprises: receiving a plurality of the high-definition images and their corresponding shooting sequence numbers transmitted by the high-definition camera, predicting the shooting sequence number of a target in a stable rotation state according to the shooting sequence number, calling the convolutional network to extract the dynamic volume feature of the saline solution from the high-definition image corresponding to the target shooting sequence number; and importing the dynamic volume feature into a volume prediction model, wherein the volume prediction model predicts the volume of the saline solution in the medical cylinder.
7. The intelligent control system for a medical ice maker according to claim 6, characterized in that: The predicting of the target shooting sequence number in a stable rotation state according to the shooting sequence number comprises: performing statistical analysis on historical ice-making data of the medical ice-making machine to obtain a regular volume of physiological saline, and predicting the target shooting sequence number of the medical cylinder in a stable rotation state according to the regular volume, rated parameters of the rotation mechanism of the medical ice-making machine and the shooting frequency of the high-definition camera.
8. An electronic device, characterized in that: Applicable to a medical ice-making machine intelligent control system as described in any one of claims 1 to 7; the electronic device comprises: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor.
9. A computer storage medium, characterized in that: Applicable to an intelligent control system for a medical ice maker as described in any one of claims 1 to 7; the computer readable storage medium stores a computer program.
10. A computer program product, characterized in that: Applicable to an intelligent control system for a medical ice-making machine as claimed in any one of claims 1 to 7; the computer program product is pre-packaged with computer program code.