Cleaning control method for mixer, processor, mixing tank and mixer

Abstract

The present invention relates to the technical field of engineering machinery, and discloses a cleaning control method, a processor, a mixing tank and a mixing tank for a mixer. The cleaning control method comprises: determining a first cleaning time according to the total time of the last cleaning; obtaining a first image of the mixing tank after cleaning the mixing tank for the first cleaning time; inputting the first image into a trained image classification model to obtain a classification level of the first image, the classification level corresponding to the classification cleaning time; and re-cleaning the mixing tank according to the classification cleaning time corresponding to the classification level. The mixing tank cleaning method based on deep learning image classification can replace manual judgment of the cleaning degree, and automatically set the cleaning time according to the cleaning degree, and at the same time adjust the current first cleaning time according to the total time of the last cleaning, which can shorten the number of judgments. The present invention realizes intelligent cleaning, improves cleaning efficiency, delays shaft seizure, saves water, and is safer and greener.

Description

Technical Field

[0001] The present invention relates to the technical field of engineering machinery, and in particular to a cleaning control method, a processor, a mixing tank and a mixer for a mixer. Background Art

[0002] Concrete is a readily solidifying material, so the interior of the mixer tank needs to be cleaned after unloading. Mixers are typically equipped with a spray cleaning device to clean the tank immediately after unloading to prevent the concrete from solidifying and affecting the mixer's operation. However, achieving precise and efficient control of the spray device to achieve automated, thorough, and intelligent cleaning has become a pressing issue.

[0003] Currently, mixers primarily flush residual concrete using high-pressure cleaning devices, often using a fixed flushing time set based on operational experience. Because the amount of residual concrete after each operation varies, the required flushing time varies. This can result in a fixed flushing time that is too long, increasing wastewater and incompatible with green and healthy development. Alternatively, the fixed flushing time can be too short, resulting in incomplete and incomplete flushing of the mixer tank, causing residual concrete to adhere to the agitator shaft or the inner walls of the tank, exacerbating the risk of shaft seizure. Summary of the Invention

[0004] In order to overcome the deficiencies in the prior art, embodiments of the present invention provide a cleaning control method for a blender, a processor, a blending tank, and a blender.

[0005] In order to achieve the above object, the first aspect of the present invention provides a cleaning control method for a blender, wherein the blender includes a blending tank, and the cleaning control method includes:

[0006] Determine the first cleaning time according to the total time of the previous cleaning;

[0007] After cleaning the stirring tank for a first cleaning time, acquiring a first image of the stirring tank;

[0008] Inputting the first image into the trained image classification model to obtain a classification level of the first image, where the classification level corresponds to the classification cleaning time;

[0009] Reclean the mixing tank according to the classification cleaning time corresponding to the classification level.

[0010] In an embodiment of the present invention, the classification level includes at least one of the following:

[0011] Complete cleaning, residue cleaning and agglomeration cleaning;

[0012] The classified cleaning time corresponding to completed cleaning is the first time, the classified cleaning time corresponding to residual cleaning is the second time, and the classified cleaning time corresponding to agglomeration cleaning is the third time. The first time includes zero, the second time is greater than the first time, and the third time is greater than the second time.

[0013] In an embodiment of the present invention, the cleaning control method further includes:

[0014] After cleaning the mixing tank according to the classified cleaning time, reacquiring a second image of the mixing tank;

[0015] Inputting the second image into the trained image classification model to obtain a classification level of the second image;

[0016] When the classification level of the second image is cleaning completed, the cleaning of the mixing tank is completed;

[0017] If the classification level of the second image is residual cleaning, re-cleaning the mixing tank according to the second time;

[0018] When the classification level of the second image is agglomeration cleaning, the mixing tank is re-cleaned according to the third time.

[0019] In an embodiment of the present invention, determining the first cleaning time according to the total time of the previous cleaning includes:

[0020] Determine the number of residual washes and the number of agglomeration washes from the previous wash process;

[0021] When the number of agglomeration cleaning is not zero, the first cleaning time is obtained according to formula (1):

[0022]

[0023] in, Indicates that the quotient of T and T3 is rounded up, T0 represents the first cleaning time, T3 represents the third time, and T represents the total time of the last cleaning;

[0024] When the number of agglomeration cleaning is zero and the number of residual cleaning is not zero, the first cleaning time is obtained according to formula (2):

[0025]

[0026] in, It means that the quotient of T and T2 is rounded up, T2 represents the second time, and the first cleaning time is less than the total time of the previous cleaning.

[0027] In an embodiment of the present invention, the image classification model is established by:

[0028] Acquire a test image of the mixing tank;

[0029] Determine the image label corresponding to the test image;

[0030] Preprocess the test image;

[0031] According to the pre-processed test image and the pre-trained image classification model, the classification level corresponding to the test image is obtained;

[0032] According to the classification level and image label corresponding to the test image, the pre-trained image classification model is adjusted to obtain the trained image classification model.

[0033] In an embodiment of the present invention, the cleaning control method further includes:

[0034] When the classification level corresponding to the test image is consistent with the image label, the test image is determined to be the correct output image;

[0035] When the number of correct output images is greater than a first preset threshold and the ratio of the number of correct output images to the total number of test images is greater than a second preset threshold, the training of the image classification model is completed.

[0036] In an embodiment of the present invention, inputting the first image into the trained image classification model includes:

[0037] Setting a region of interest (ROI) for the first image;

[0038] The preprocessed first image is input into the trained image classification model.

[0039] A second aspect of the present invention provides a processor configured to execute the above-mentioned cleaning control method for a blender.

[0040] A third aspect of the present invention provides a stirring tank comprising:

[0041] An image acquisition device, used for acquiring images of the mixing tank;

[0042] A light source for supplementing light to the mixing tank; and

[0043] The processor mentioned above.

[0044] A fourth aspect of the present invention provides a blender comprising the above-mentioned blending tank.

[0045] A fifth aspect of the present invention provides a machine-readable storage medium storing instructions for enabling a machine to execute the above-mentioned cleaning control method for a blender.

[0046] Because various operating conditions within the same mixing tank often share similarities, to improve cleaning efficiency, the first cleaning time is determined based on the total time of the previous cleaning. This ensures that the current first cleaning time is close to the total time of the previous cleaning, reducing the number of judgments made by the image classification model and resulting in more efficient cleaning. Furthermore, the current first cleaning time can be set to be shorter than the total time of the previous cleaning, thus avoiding water waste and increasing wastewater volume. After the first cleaning time, an image of the mixing tank is captured to determine the degree of cleaning completion. The next cleaning time is automatically set based on this degree of completion. This method, based on deep learning image classification, replaces manual cleaning completion assessments, enabling intelligent cleaning. It also improves cleaning efficiency, conserves water, and is safer and more environmentally friendly than traditional cleaning methods. If the image classification model detects that the image is classified as complete, the tank cleaning process is terminated. This results in a more thorough and comprehensive tank rinse, minimizing residual concrete adhering to the agitator shaft or the inner wall of the tank, delaying shaft seizure and paving the way for shaft seizure detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following detailed description, they are used to explain the embodiments of the present invention, but do not constitute a limitation of the embodiments of the present invention. In the accompanying drawings:

[0048] Figure 1 Schematically shows one of the flow charts of a cleaning control method for a blender according to an embodiment of the present invention;

[0049] Figure 2 Schematically shows a hardware layout diagram of a blender according to an embodiment of the present invention;

[0050] Figure 3 The second flowchart of the cleaning control method for a blender according to an embodiment of the present invention is schematically shown.

[0051] Description of Reference Numerals

[0052] 10-Industrial computer; 11-Mixing tank;

[0053] 12- Programmable Logic Controller PLC 13- Cleaning device;

[0054] 14-hopper; 15-image acquisition device;

[0055] 16-Material viewing device port. DETAILED DESCRIPTION

[0056] The following describes the specific implementation of the embodiment of the present invention in detail with reference to the accompanying drawings. It should be understood that the specific implementation described herein is only used to illustrate and explain the embodiment of the present invention and is not used to limit the embodiment of the present invention.

[0057] It should be noted that if the implementation methods of this application involve directional indications (such as up, down, left, right, front, back, etc.), such directional indications are only used to explain the relative position relationship, movement status, etc. between the various components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0058] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of this application, the descriptions of "first", "second", etc. are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features specified as "first" and "second" may explicitly or implicitly include at least one of such features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the fact that they can be implemented by ordinary technicians in this field. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0059] Figure 1 The following schematically shows a flow chart of a method for controlling the cleaning of a blender according to an embodiment of the present invention. The blender includes a blending tank. Figure 1 As shown, in one embodiment of the present invention, a cleaning control method for a blender is provided, comprising the following steps:

[0060] Step 101, determining a first cleaning time according to the total time of the last cleaning;

[0061] Step 102, after the mixing tank is cleaned for a first cleaning time, obtaining a first image of the mixing tank;

[0062] Step 103: input the first image into the trained image classification model to obtain a classification level of the first image, where the classification level corresponds to the classification cleaning time;

[0063] Step 104 : Re-clean the mixing tank according to the classification cleaning time corresponding to the classification level.

[0064] Because various operating conditions within the same mixing tank often share similarities, to improve cleaning efficiency, the first cleaning time is determined based on the total time of the previous cleaning. This ensures that the current first cleaning time is close to the total time of the previous cleaning, reducing the number of judgments made by the image classification model and resulting in more efficient cleaning. Furthermore, the current first cleaning time can be set to be shorter than the total time of the previous cleaning, thus avoiding water waste and increasing wastewater volume. After the first cleaning time, an image of the mixing tank is captured to determine the degree of cleaning completion. The next cleaning time is automatically set based on this degree of completion. This method, based on deep learning image classification, replaces manual cleaning completion assessments, enabling intelligent cleaning. It also improves cleaning efficiency, conserves water, and is safer and more environmentally friendly than traditional cleaning methods. If the image classification model detects that the image is classified as complete, the tank cleaning process is terminated. This results in a more thorough and comprehensive tank rinse, minimizing residual concrete adhering to the agitator shaft or the inner wall of the tank, delaying shaft seizure and paving the way for shaft seizure detection.

[0065] The present invention is based on the existing material-watching device of the mixing main unit of the mixing station, and relates to an intelligent cleaning control method for the mixing main unit based on deep learning image classification. Routine maintenance and cleaning of concrete mixers is an important task. Usually, the cleaning of the mixer is divided into two situations: (1) During a working cycle, a high-pressure water gun is used to clean the mixing tank to prevent the residual concrete from solidifying. (2) At the end of the work, an appropriate amount of sand and water is added to the mixing tank for stirring, so that the solidified concrete falls off in the collision with the sand and gravel. If it is found that the concrete is seriously stuck to the shaft, the shaft agglomeration needs to be cleaned.

[0066] In an embodiment of the present invention, intelligent cleaning of the mixing tank is achieved by replacing the traditional fixed-time cleaning method with multiple short-time cleanings (dynamically setting the cleaning time for each time). The cleaning device can be manually controlled in real time in the control room and the cleaning effect can be observed; the mixing tank can also be automatically cleaned and the cleaning effect can be automatically determined; and the mixing shaft can be delayed. The cleaning control method for the mixer according to the embodiment of the present invention can achieve efficient control of the cleaning device, achieve the purpose of saving water, green production and delaying the shaft sticking, and can observe the situation of the mixing shaft sticking in real time, so as to carry out daily maintenance of the mixer.

[0067] In one embodiment, the classification level includes at least one of the following:

[0068] Complete cleaning, residue cleaning and agglomeration cleaning;

[0069] The classified cleaning time corresponding to completed cleaning is the first time, the classified cleaning time corresponding to residual cleaning is the second time, and the classified cleaning time corresponding to agglomeration cleaning is the third time. The first time includes zero, the second time is greater than the first time, and the third time is greater than the second time.

[0070] In one embodiment, the cleaning control method further includes:

[0071] After cleaning the mixing tank according to the classified cleaning time, reacquiring a second image of the mixing tank;

[0072] Inputting the second image into the trained image classification model to obtain a classification level of the second image;

[0073] When the classification level of the second image is cleaning completed, the cleaning of the mixing tank is completed;

[0074] If the classification level of the second image is residual cleaning, re-cleaning the mixing tank according to the second time;

[0075] When the classification level of the second image is agglomeration cleaning, the mixing tank is re-cleaned according to the third time.

[0076] After each cleaning, an image of the mixing tank is obtained, and then the mixing tank is cleaned again according to the classification cleaning time corresponding to the classification level of the image, until the classification level of the image is completed, and the cleaning of the mixing tank is terminated.

[0077] Classifying cleaning images into different levels involves creating a training dataset, collecting images from various cleaning stages, and then qualitatively analyzing each image to determine the approximate time required for cleaning to complete based on the cleaning experience of the engineer. The images are then classified into different levels based on the cleaning time. In one embodiment, the image classification levels include completed cleaning, residual cleaning, and agglomerated cleaning.

[0078] Cleaning completed: There is no residual concrete in the mixing tank, and the cleaning operation is completed. The cleaning completed image is classified as one category and named as cleaning completed. When the cleaning time corresponding to the cleaning completed category is the first time, the first time can be 0.

[0079] Residual cleaning: The mixing shaft and mixing blades contain a small amount of residual concrete. The time required for thorough cleaning is 0<t≤T2. The images in this interval are classified as the same category and named as residual cleaning. The cleaning time threshold T2 (second time) is set.

[0080] Agglomerate cleaning: The mixing shaft and mixing blades contain a lot of concrete agglomerates. The time required for thorough cleaning is T2<t≤T3. The images in this interval are classified as the same category and named as agglomerate cleaning. The cleaning time threshold T3 (the third time) is set.

[0081] In summary, based on engineers' experience, we can categorize images into three different cleaning times. Since images requiring residual cleaning and agglomeration cleaning require further cleaning, corresponding cleaning times are set: T2 for residual cleaning and T3 for agglomeration cleaning. It should be noted that categorizing images into these three categories is only one implementation of the present invention. Images can be further categorized into different levels based on actual needs, with each level corresponding to a specific cleaning time.

[0082] In one embodiment, determining the first cleaning time according to the total time of the previous cleaning includes:

[0083] Determine the number of residual washes and the number of agglomeration washes from the previous wash process;

[0084] When the number of agglomeration cleaning is not zero, the first cleaning time is obtained according to formula (1):

[0085]

[0086] in, Indicates that the quotient of T and T3 is rounded up, T0 represents the first cleaning time, T3 represents the third time, and T represents the total time of the last cleaning;

[0087] When the number of agglomeration cleaning is zero and the number of residual cleaning is not zero, the first cleaning time is obtained according to formula (2):

[0088]

[0089] in, It means that the quotient of T and T2 is rounded up, T2 represents the second time, and the first cleaning time is less than the total time of the previous cleaning.

[0090] For example, during the last cleaning process, the total time of the last cleaning is T, the number of agglomeration cleanings during the last cleaning process is not zero, and T3 represents the cleaning time corresponding to the agglomeration cleaning during the last cleaning process. Assuming T / T3 = 3.2, then the current first cleaning time T0 = (4-1)*T3 = 3*T3, and at this time T = 3.2*T3. Assuming T / T3 = 3, then the current first cleaning time T0 = (3-1)*T3 = 2*T3, and at this time T = 3*T3. It can be seen that the current first cleaning time T0 is close to the total time T of the last cleaning, but the first cleaning time T0 is less than the total time T of the last cleaning.

[0091] Regarding the first cleaning time setting for the mixing tank, since various operating conditions of the same mixing tank are similar, in order to improve cleaning efficiency, the first cleaning time is determined based on the total time of the previous cleaning, so that the current first cleaning time is close to the total time of the previous cleaning, reducing the number of judgments made by the image classification model and making cleaning more efficient. On the other hand, the current first cleaning time can be set to be less than the total time of the previous cleaning to avoid wasting water resources and increasing the amount of wastewater.

[0092] In one embodiment, the image classification model is established by:

[0093] Acquire a test image of the mixing tank;

[0094] Determine the image label corresponding to the test image;

[0095] Preprocess the test image;

[0096] According to the pre-processed test image and the pre-trained image classification model, the classification level corresponding to the test image is obtained;

[0097] According to the classification level and image label corresponding to the test image, the pre-trained image classification model is adjusted to obtain the trained image classification model.

[0098] In one embodiment, the cleaning control method further includes:

[0099] When the classification level corresponding to the test image is consistent with the image label, the test image is determined to be the correct output image;

[0100] When the number of correct output images is greater than a first preset threshold and the ratio of the number of correct output images to the total number of test images is greater than a second preset threshold, the training of the image classification model is completed.

[0101] In one embodiment, inputting the first image into the trained image classification model includes:

[0102] Setting a region of interest (ROI) for the first image;

[0103] The preprocessed first image is input into the trained image classification model.

[0104] The present invention relates to an automatic cleaning control method based on deep learning image classification. Since the present invention adopts a deep learning classification model, the degree of cleaning is mainly judged by the collected images, thereby setting the cleaning time. Therefore, the learning of the image classification model is completed first, and then the automatic cleaning system is constructed through the classification model.

[0105] Data collection is required to train image classification models. Figure 2 The hardware layout diagram of the mixer according to the embodiment of the present invention is shown schematically. Figure 2 As shown, an image acquisition device 15 is installed above the main mixing unit. An industrial camera can capture image data from the mixing tank at each cleaning stage. Because the mixing tank is an enclosed space, a supplemental light source is required on the acquisition device 15. The captured images can be categorized into three categories and labeled (complete cleaning, residual cleaning, and agglomeration cleaning). The amount of data collected is consistent across all classification levels.

[0106] Preprocess the image. When building a sample database, deep learning models require a large amount of sample data. Data augmentation can be performed through methods such as image flipping, rotation, translation, and brightness shifting. Furthermore, when training the model offline, the collected offline data may be relatively simple and incomplete. Therefore, in later practical applications, additional data can be collected to expand the sample library and retrain the image classification model to improve classification accuracy.

[0107] Offline training model. In one embodiment, the ResNet101 model in the deep learning classification model is selected, the number of model training is set, and the accuracy threshold of the test set is set. When the test accuracy is greater than the threshold, the current model parameters are saved to obtain a clean image classification model. The data used in the test set is part of the training set. Each image has a corresponding label, and the label is the level of the image. When the classification level output by the image classification model is consistent with the manually calibrated image label, it is considered to be a correct output.

[0108] Figure 3 The second flow chart of the cleaning control method for a mixer according to an embodiment of the present invention is schematically shown. Figure 3 In this example, m represents the cumulative number of agglomeration cleaning during this cleaning process, and n represents the cumulative number of residual cleaning during this cleaning process. The process of intelligent cleaning based on image classification can be found in Figure 3 . The present invention involves multiple cleanings and dynamic setting of cleaning time. The dynamic setting of cleaning time is reflected in that, except for the first cleaning, each subsequent cleaning is performed by the model firstly collecting images, outputting the classification level of the image, and then automatically setting the corresponding cleaning time. For example, when the image classification model outputs residual cleaning, the cleaning time is set to T2; if the model outputs agglomeration cleaning, the cleaning time is set to T3, where T2 and T3 are fixed times. In this way, the purpose of saving water, delaying shaft sticking, intelligent cleaning and safe operation is achieved. For details, please refer to the following steps:

[0109] (1) After the unloading is completed, the cleaning of the mixing tank begins. The time of the first cleaning does not need to be determined by the model, but is determined by the total cleaning time of the previous mixing tank;

[0110] (2) After the first cleaning is completed, the butterfly valve of the material viewing device is opened, the camera collects image data, and extracts a single frame image;

[0111] (3) Setting ROI for a single-frame image to eliminate the interference between the feed port and the inner wall of the upper cover of the mixing main unit and reduce the calculation amount of the model;

[0112] (4) According to the size of the model input, the preprocessed image is normalized and input into the image classification model to output the classification level of the image;

[0113] (5) Determine whether cleaning is complete based on the classification level output by the model. If further cleaning is required, reset the cleaning time based on the classification level output after the cleaning is completed.

[0114] Because various operating conditions within the same mixing tank often share similarities, to improve cleaning efficiency, the first cleaning time is determined based on the total time of the previous cleaning. This ensures that the current first cleaning time is close to the total time of the previous cleaning, reducing the number of judgments made by the image classification model and resulting in more efficient cleaning. Furthermore, the current first cleaning time can be set to be shorter than the total time of the previous cleaning, thus avoiding water waste and increasing wastewater volume. After the first cleaning time, an image of the mixing tank is captured to determine the degree of cleaning completion. The next cleaning time is automatically set based on this degree of completion. This method, based on deep learning image classification, replaces manual cleaning completion assessments, enabling intelligent cleaning. It also improves cleaning efficiency, conserves water, and is safer and more environmentally friendly than traditional cleaning methods. If the image classification model detects that the image is classified as complete, the tank cleaning process is terminated. This results in a more thorough and comprehensive tank rinse, minimizing residual concrete adhering to the agitator shaft or the inner wall of the tank, delaying shaft seizure and paving the way for shaft seizure detection.

[0115] An embodiment of the present invention provides a processor configured to execute any one of the cleaning control methods for a blender in the above embodiments.

[0116] The blender includes a blending tank. Specifically, the processor can be configured to:

[0117] Determine the first cleaning time according to the total time of the previous cleaning;

[0118] After cleaning the stirring tank for a first cleaning time, acquiring a first image of the stirring tank;

[0119] Inputting the first image into the trained image classification model to obtain a classification level of the first image, where the classification level corresponds to the classification cleaning time;

[0120] Reclean the mixing tank according to the classification cleaning time corresponding to the classification level.

[0121] In an embodiment of the present invention, the processor is configured to:

[0122] Classification levels include at least one of the following:

[0123] Complete cleaning, residue cleaning and agglomeration cleaning;

[0124] The classified cleaning time corresponding to completed cleaning is the first time, the classified cleaning time corresponding to residual cleaning is the second time, and the classified cleaning time corresponding to agglomeration cleaning is the third time. The first time includes zero, the second time is greater than the first time, and the third time is greater than the second time.

[0125] In this embodiment of the present invention, the processor is further configured to:

[0126] After cleaning the mixing tank according to the classified cleaning time, reacquiring a second image of the mixing tank;

[0127] Inputting the second image into the trained image classification model to obtain a classification level of the second image;

[0128] When the classification level of the second image is cleaning completed, the cleaning of the mixing tank is completed;

[0129] If the classification level of the second image is residual cleaning, re-cleaning the mixing tank according to the second time;

[0130] When the classification level of the second image is agglomeration cleaning, the mixing tank is re-cleaned according to the third time.

[0131] In an embodiment of the present invention, the processor is configured to:

[0132] Determining the first cleaning time based on the total time of the previous cleaning includes:

[0133] Determine the number of residual washes and the number of agglomeration washes from the previous wash process;

[0134] When the number of agglomeration cleaning is not zero, the first cleaning time is obtained according to formula (1):

[0135]

[0136] in, Indicates that the quotient of T and T3 is rounded up, T0 represents the first cleaning time, T3 represents the third time, and T represents the total time of the last cleaning;

[0137] When the number of agglomeration cleaning is zero and the number of residual cleaning is not zero, the first cleaning time is obtained according to formula (2):

[0138]

[0139] in, It means that the quotient of T and T2 is rounded up, T2 represents the second time, and the first cleaning time is less than the total time of the previous cleaning.

[0140] In an embodiment of the present invention, the processor is configured to:

[0141] The image classification model is built in the following ways:

[0142] Acquire a test image of the mixing tank;

[0143] Determine the image label corresponding to the test image;

[0144] Preprocess the test image;

[0145] According to the pre-processed test image and the pre-trained image classification model, the classification level corresponding to the test image is obtained;

[0146] According to the classification level and image label corresponding to the test image, the pre-trained image classification model is adjusted to obtain the trained image classification model.

[0147] In this embodiment of the present invention, the processor is further configured to:

[0148] When the classification level corresponding to the test image is consistent with the image label, the test image is determined to be the correct output image;

[0149] When the number of correct output images is greater than a first preset threshold and the ratio of the number of correct output images to the total number of test images is greater than a second preset threshold, the training of the image classification model is completed.

[0150] In an embodiment of the present invention, the processor is configured to:

[0151] Inputting the first image into the trained image classification model includes:

[0152] Setting ROI for the first image;

[0153] The preprocessed first image is input into the trained image classification model.

[0154] An embodiment of the present invention provides a stirring tank, comprising:

[0155] An image acquisition device, used for acquiring images of the mixing tank;

[0156] A light source for supplementing light to the mixing tank; and

[0157] The processor mentioned above.

[0158] An embodiment of the present invention provides a blender including the above-mentioned blending tank.

[0159] An embodiment of the present invention provides a machine-readable storage medium storing instructions for enabling a machine to execute the above-mentioned cleaning control method for a blender.

[0160] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0161] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0162] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0163] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1A step that specifies a function in one or more boxes.

[0164] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0165] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0166] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0167] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0168] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A cleaning control method for a mixer, characterized in that: The mixer includes a mixing tank, and the cleaning control method includes: Determine the first cleaning time according to the total time of the previous cleaning; After cleaning the stirring tank for the first cleaning time, acquiring a first image of the stirring tank; Inputting the first image into a trained image classification model to obtain a classification level of the first image, wherein the classification level corresponds to a classification cleaning time; The mixing tank is re-cleaned according to the classification cleaning time corresponding to the classification level, wherein the classification level includes at least one of the following: Complete cleaning, residue cleaning and agglomeration cleaning; The classification cleaning time corresponding to the completion cleaning is a first time, the classification cleaning time corresponding to the residual cleaning is a second time, and the classification cleaning time corresponding to the agglomeration cleaning is a third time. The first time includes zero, the second time is greater than the first time, and the third time is greater than the second time. Determining the first cleaning time according to the total time of the last cleaning includes: Determining the number of residual cleanings and the number of agglomeration cleanings during the last cleaning process; When the number of times of cleaning the agglomerates is not zero, the first cleaning time is obtained according to formula (1): T0=(⌈T / T3⌉-1)*T3 formula (1) Wherein, ⌈T / T3⌉ represents rounding up the quotient of T and T3, T0 represents the first cleaning time, T3 represents the third time, and T represents the total time of the previous cleaning; When the number of agglomeration cleaning is zero and the number of residual cleaning is not zero, the first cleaning time is obtained according to formula (2): T0=(⌈T / T2⌉-1)*T2 formula (2) Here, ⌈T / T2⌉ represents rounding up the quotient of T and T2, T2 represents the second time, and the first cleaning time is less than the total time of the previous cleaning.

2. The cleaning control method according to claim 1, characterized in that: Also includes: After cleaning the mixing tank according to the classified cleaning time, reacquiring a second image of the mixing tank; inputting the second image into the trained image classification model to obtain a classification level of the second image; When the classification level of the second image is cleaning completed, the cleaning of the mixing tank is completed; If the classification level of the second image is residual cleaning, re-cleaning the mixing tank according to the second time; When the classification level of the second image is agglomeration cleaning, the mixing tank is re-cleaned according to the third time.

3. The cleaning control method according to claim 1, characterized in that: The image classification model is established in the following manner: Acquiring a test image of the stirring tank; Determining an image label corresponding to the test image; Preprocessing the test image; Obtaining a classification level corresponding to the test image based on the preprocessed test image and the pre-trained image classification model; The pre-trained image classification model is adjusted according to the classification level corresponding to the test image and the image label to obtain the post-trained image classification model.

4. The cleaning control method according to claim 3, characterized in that: Also includes: When the classification level corresponding to the test image is consistent with the image label, determining that the test image is a correct output image; When the number of the correct output images is greater than a first preset threshold and the ratio of the number of the correct output images to the total number of the test images is greater than a second preset threshold, the training of the image classification model is completed.

5. The cleaning control method according to claim 1, characterized in that: Inputting the first image into the trained image classification model includes: Setting a region of interest for the first image; The preprocessed first image is input into the trained image classification model.

6. A processor, characterized in that: The device is configured to execute the cleaning control method for a blender according to any one of claims 1 to 5.

7. A stirring tank, characterized in that: include: An image acquisition device, used for acquiring images of the mixing tank; A light source, used for supplementing light to the mixing tank; as well as The processor according to claim 6.

8. A mixer, characterized in that: Comprising the stirring tank according to claim 7.

Images