Multi-mode intelligent vacuum fresh-keeping machine based on food material identification and identification method

Abstract

The invention discloses a multi-mode intelligent vacuum fresh-keeping machine and method based on food material recognition. The fresh-keeping machine comprises a vacuum cavity assembly, a depth vision collection assembly, a pneumatic execution assembly and a main control processing assembly. The pneumatic execution assembly comprises a frequency conversion pneumatic pump module, an airflow direction switching valve module and an array type differential suction nozzle module. The array type differential suction nozzle module is provided with a plurality of independent air suction units connected with the independent electromagnetic valve module in series. The main control processing assembly is configured to monitor the local deformation rate of the surface of the vacuum packaging bag through the depth vision collection assembly, generate a differential control signal based on the deformation rate distribution difference and independently adjust the on-off state of each independent electromagnetic valve module so as to construct a gradient negative pressure field. The main control processing assembly is further configured to identify object attributes by detecting the virtual pneumatic impedance and monitor the liquid diffusion risk in combination with an optical flow method. According to the invention, active self-adaptive vacuum preservation based on visual-airflow field coupling is realized, and the packaging quality of complex food materials is effectively improved.

Description

Technical Field

[0001] This invention relates to the field of food preservation equipment technology, specifically to a multi-mode intelligent vacuum preservation machine and identification method based on food ingredient recognition. Background Technology

[0002] Vacuum packaging technology is widely used in the food storage and processing industry. Its core principle is to reduce the oxygen content by removing air from the packaging container, thereby inhibiting the growth of aerobic microorganisms and delaying the oxidation and spoilage process of food. At the same time, it can save storage and transportation space by compressing the volume.

[0003] Existing vacuum packaging equipment mainly includes household vacuum machines and commercial vacuum packaging machines. Their hardware structure typically consists of a vacuum chamber, a negative pressure pump unit, and a heat-sealing assembly. In a conventional workflow, the user places the packaging bag containing the object to be preserved into the vacuum chamber or attaches the bag opening to the suction nozzle, and the equipment starts the vacuum pump to perform the evacuation operation. Most existing control logics rely on preset time parameters or a single negative pressure threshold parameter; that is, the equipment continuously evacuates until the sensor detects that the absolute pressure in the air path has reached the set standard, and then automatically triggers the heat-sealing action to complete the sealing.

[0004] However, this passive control mode based on a single air pressure or time parameter has significant limitations when handling food with complex physical properties. Because the device establishes a uniformly distributed negative pressure field inside the packaging bag during the suction process, and cannot sense the real-time deformation or physical texture of the internal objects, the closure of the airflow channel is often uncontrolled. When handling irregularly shaped food, the packaging film tends to prematurely adhere in low-resistance areas, blocking the airflow path and leaving unextracted air pockets in the food's gaps. Conversely, when handling softer food, continuous and indiscriminate strong negative pressure suction can easily cause irreversible over-compression deformation or mechanical damage, making it difficult to balance degassing efficiency with the integrity of the food's physical form. Summary of the Invention

[0005] The first aspect of this invention provides a multi-mode intelligent vacuum preservation machine based on food ingredient recognition.

[0006] The multi-mode intelligent vacuum preservation machine based on food identification mainly includes a vacuum chamber assembly, a depth vision acquisition assembly, a pneumatic actuation assembly, a sensing and detection assembly, and a main control processing assembly. The vacuum chamber assembly defines a sealed operating space for vacuum processing and includes a heat-sealing assembly for sealing vacuum packaging bags. The sensing and detection assembly includes a depth vision acquisition assembly positioned above the vacuum chamber assembly and a high-frequency pressure sensor assembly positioned in the main airflow pipe. The depth vision acquisition assembly is configured to acquire 3D point cloud data of the vacuum packaging bag and its contents, while the high-frequency pressure sensor assembly is configured to detect the absolute pressure value and instantaneous pressure fluctuation rate within the airflow system. The pneumatic actuation assembly includes a variable frequency pneumatic pump module connected to the vacuum chamber assembly via an airflow pipe, an airflow direction switching valve module, and an array of differential suction nozzles located inside the vacuum chamber assembly.

[0007] The array-type differential suction nozzle module comprises at least three independent suction units arranged linearly in the transverse direction. Each independent suction unit has its air path connected in series with an independent solenoid valve module. These independent solenoid valve modules converge and connect to a variable frequency pneumatic pump module. The main control processing unit communicates with the depth vision acquisition unit, the variable frequency pneumatic pump module, the airflow direction switching valve module, and the independent solenoid valve modules. The main control processing unit monitors the local deformation rate of the vacuum packaging bag surface through the depth vision acquisition unit and, based on the differences in the distribution of this local deformation rate, generates differential control signals for the array-type differential suction nozzle module. This allows for independent control of the on / off state of each independent solenoid valve module, thereby adjusting the suction distribution at different transverse positions at the bag opening.

[0008] Regarding the drive structure of the array-type differential suction nozzle module, the module is mechanically coupled with a linear displacement drive unit. The main control processing unit controls the linear displacement drive unit to drive the independent suction units to reciprocate between a working position extending into the opening of the vacuum packaging bag and a clearance position exiting the opening. Based on the physical layout parameters of the independent suction units, the main control processing unit divides the two-dimensional projected outline of the vacuum packaging bag into virtual control sub-regions corresponding to each independent suction unit, thereby establishing a mapping relationship between the visual space and the physical air path space.

[0009] For identifying the physical properties of food ingredients, the main control processing unit controls the variable frequency pneumatic pump module to output short-term negative pressure pulses at the initial stage of the vacuuming process. Based on the ratio of pressure gradient data collected by the high-frequency pressure sensor to the volume change of the object collected by the depth vision acquisition component, the main control processing unit calculates the virtual aerodynamic impedance value. Based on the virtual aerodynamic impedance value, the main control processing unit determines the physical property category of the object inside the vacuum packaging bag and sets the target vacuum level threshold and maximum allowable deformation rate threshold for the vacuuming process accordingly.

[0010] In terms of flow field control logic, the main control processing component calculates the rate deviation of the local deformation rate of each virtual control sub-region relative to the global average deformation rate. When the rate deviation indicates that a certain virtual control sub-region is a fast deformation region, the main control processing component sends a pulse width modulation signal to the corresponding independent solenoid valve module to intermittently or completely close the air path of the independent solenoid valve module; when the rate deviation indicates that a certain virtual control sub-region is a slow deformation region, the main control processing component controls the corresponding independent solenoid valve module to remain in a fully open state, guiding the airflow out of the slow deformation region.

[0011] To control the surface flatness of vacuum packaging bags, the main control processing component calculates the variance of the depth gradient distribution within the region of interest below the heat-sealing component, and uses this variance as a wrinkle complexity index. When the wrinkle complexity index exceeds a preset flatness threshold, the main control processing component controls the airflow direction switching valve module to switch to a positive pressure output path and controls the variable frequency pneumatic pump module to output a short-term positive pressure airflow to the array-type differential suction nozzle module. The positive pressure airflow generated by the airflow path switching causes tension on the surface of the vacuum packaging bag, causing it to bulge outwards and eliminating the physical folds on the surface of the vacuum packaging bag.

[0012] For monitoring liquid diffusion, the main control processing component uses an optical flow algorithm to calculate the two-dimensional velocity vector of the liquid front inside the vacuum packaging bag and filters out valid pixels whose longitudinal component is greater than a preset longitudinal velocity threshold, which is greater than the maximum physical shrinkage rate of the vacuum packaging bag under high negative pressure. Based on the real-time position of the liquid front and the sealing baseline coordinates of the heating sealing component, the main control processing component predicts the remaining safety time. When the remaining safety time is less than the emergency response time threshold, the main control processing component sends a full shutdown command to all independent solenoid valve modules and simultaneously drives the heating sealing component to perform mechanical pressing and heat sealing.

[0013] Regarding the determination of air extraction termination, the main control processing component establishes parallel monitoring channels. The air extraction process is determined to terminate when any of the following conditions are met: the real-time absolute pressure value of the air path detected by the high-frequency air pressure sensor is lower than or equal to the target vacuum threshold; the volumetric strain rate of the object calculated by the depth vision acquisition component is greater than or equal to the maximum allowable deformation rate threshold; or the continuous operation time of the main air extraction stage exceeds the preset safety timeout threshold.

[0014] Regarding sealing control, the main control processing component initiates a hysteresis balancing program after the mechanical pressing mechanism of the heating and sealing assembly closes. During the preset hysteresis balancing time interval, the main control processing component maintains the clamping force of the mechanical pressing mechanism and prohibits power supply to the heating and sealing assembly to eliminate the elastic rebound potential energy of the vacuum packaging bag in the pressing area. After the hysteresis balancing time interval ends, the main control processing component calculates the target thermal energy based on ambient temperature data and the preset material heat transfer coefficient, and generates a three-segment pulse width modulation drive signal. The three-segment pulse width modulation drive signal controls the heating and sealing assembly to sequentially execute the heating impact stage with a 100% output duty cycle, the isothermal fusion stage where the output duty cycle is maintained to balance heat dissipation, and the cooling and solidification stage where the output duty cycle is zero and mechanical pressing is maintained. During the isothermal fusion stage, the main control processing component integrates the output power, and stops heating when the accumulated output heat reaches the target thermal energy.

[0015] A second aspect of the present invention provides an active adaptive vacuum preservation method based on apparent-airflow field coupling.

[0016] The active adaptive vacuum preservation method based on view-airflow field coupling includes the following steps: acquiring initial 3D point cloud data of the vacuum packaging bag using a depth vision acquisition component to construct an initial surface topology model; controlling the variable frequency pneumatic pump module to output short-term negative pressure pulses, calculating the virtual aerodynamic impedance value based on the ratio of pressure change to volume deformation, and determining the physical property category of the objects inside the vacuum packaging bag; during continuous air extraction, monitoring the local deformation rate of different areas on the surface of the vacuum packaging bag in real time, and calculating the deviation between the local deformation rate and the global average deformation rate; generating differential control signals based on the deviation, independently adjusting the opening and closing states of each independent solenoid valve module connected to the back end of the array-type differential suction nozzle module, limiting the flow in the rapid deformation area, and maintaining full conduction in the slow deformation area; monitoring the fold complexity and internal liquid diffusion risk of the vacuum packaging bag in real time, and terminating air extraction when the preset air pressure threshold or deformation rate threshold is met, driving the heating sealing component to complete the sealing.

[0017] This invention provides a multi-mode intelligent vacuum preservation machine and identification method based on food ingredient recognition. It has the following beneficial effects: 1. The depth vision acquisition component of this invention monitors the local deformation rate of the vacuum packaging bag surface in real time. Combined with the physical properties of the object calculated by the aerodynamic impedance calculation module, a dynamic air extraction threshold is set. The differential control signal is used to independently adjust the on / off state of each independent solenoid valve module at the back end of the array-type differential suction nozzle module. During the air extraction process, the flow-limiting modulation is implemented in the fast deformation area while maintaining full conduction in the slow deformation area. This achieves the effect of adaptively adjusting the airflow field distribution according to the shape and texture of the food, effectively eliminating local residual air pockets caused by irregular object shapes, and avoiding mechanical damage to soft food caused by excessive suction.

[0018] 2. This invention uses a visual reconstruction analysis module to calculate the variance of the depth gradient distribution in the bag opening area to identify surface wrinkles. When wrinkles are detected, the airflow direction switching valve module is controlled to output a short-term positive pressure airflow to flatten the bag opening using tension. At the same time, the optical flow algorithm is used to calculate the two-dimensional velocity vector of the liquid front and predict the remaining safe time for it to reach the sealing line. When the risk of liquid diffusion is detected, the air circuit is immediately shut off and an emergency mechanical pressing procedure is triggered. This achieves proactive intervention in the quality of vacuum packaging and protection of the core components of the equipment under complex working conditions, solving the technical problems of liquid aspiration leading to air pump damage and bag opening wrinkles leading to sealing failure.

[0019] 3. This invention eliminates the elastic rebound potential energy of the vacuum packaging bag material by maintaining mechanical pressure without energizing after the heating and sealing assembly is closed through a hysteresis balancing program. Subsequently, the target thermal energy is calculated based on the ambient temperature and the material's thermal melting coefficient. The heating element is controlled by a three-stage nonlinear pulse width modulation signal that includes heating impact, constant temperature fusion, and cooling solidification. The fusion endpoint is determined based on energy integration logic. This achieves consistent sealing quality under different ambient temperatures and continuous working heat accumulation conditions, ensuring the flatness and tensile strength of the sealing weld and avoiding packaging failure caused by false sealing or overheating. Attached Figure Description

[0020] Figure 1 This is a system schematic diagram of the present invention; Figure 2 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0021] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Example: Please see the appendix Figure 1 - Appendix Figure 2 The present invention provides a multi-mode intelligent vacuum preservation machine based on food identification, including a vacuum chamber assembly, a sensing and detection assembly, a pneumatic actuation assembly, and a main control processing assembly.

[0023] The vacuum chamber assembly includes a base unit and a pressure-applying cover unit movably connected above the base unit. When the base unit and the pressure-applying cover unit are closed, they define a sealed operating space for vacuum processing. The top surface of the base unit has a support plane for placing a vacuum packaging bag. At the front end of the vacuum chamber assembly and on one side of the support plane, a laterally extending heat-sealing assembly is provided for heat-sealing the open end of the vacuum packaging bag.

[0024] The sensing and detection component includes a depth vision acquisition component disposed above the vacuum cavity component and a high-frequency pressure sensing component disposed in the main gas pipeline; the depth vision acquisition component is configured to acquire three-dimensional point cloud data of the vacuum packaging bag and the object inside, and the high-frequency pressure sensing component is configured to detect the absolute pressure value and instantaneous fluctuation rate of pressure change inside the gas pipeline system.

[0025] The depth vision acquisition component is located on the top inner side of the pressure-applying cover unit, with its field of view facing downwards, covering the bearing plane area inside the vacuum chamber assembly and the area where the heat-sealing assembly is located. The depth vision acquisition component is configured to acquire 3D point cloud data and 2D image data of the vacuum packaging bag and its internal objects placed on the bearing plane. The depth vision acquisition component uses a structured light camera or a time-of-flight camera, and its optical axis is set perpendicular to the bearing plane or tilted at a preset angle to fully capture the deformation characteristics of the vacuum packaging bag surface.

[0026] The pneumatic actuator includes a variable frequency pneumatic pump module. This module is connected to the vacuum chamber assembly via a main air duct. The variable frequency pneumatic pump module is equipped with a power regulation circuit, capable of changing the pumping or charging rate based on received control signals. An airflow direction switching valve module is installed in the air path between the variable frequency pneumatic pump module and the vacuum chamber assembly. This valve module switches the airflow direction in the air path, allowing the system to transition between negative pressure pumping and positive pressure blowing states.

[0027] An array-type differential suction nozzle module is disposed inside the vacuum chamber assembly and adjacent to the heat-sealing assembly. The array-type differential suction nozzle module includes a linear displacement drive unit and several independent suction units arranged linearly in the transverse direction. The number of independent suction units is at least three, and they are evenly distributed along the length of the heat-sealing assembly. The linear displacement drive unit is mechanically connected to the independent suction units and drives the independent suction units to reciprocate between a working position extending into the vacuum packaging bag opening and a clearance position exiting the vacuum packaging bag opening. Each independent suction unit has an independent air inlet port.

[0028] Each independent suction unit has an independent solenoid valve module connected in series at its rear air path. Each independent solenoid valve module independently controls the on / off state of its corresponding independent suction unit's air path. The outputs of all independent solenoid valve modules converge into the main air pipe, which is then connected to the airflow direction switching valve module and the variable frequency pneumatic pump module. By independently controlling the opening and closing combinations of each independent solenoid valve module, the suction distribution at different lateral positions at the opening of the vacuum packaging bag can be altered.

[0029] The high-frequency air pressure sensing component is located on the common pipeline between the array-type differential nozzle module and the variable frequency pneumatic pump module. It is used to detect the absolute pressure value inside the air circuit system and the instantaneous fluctuation rate of pressure changes.

[0030] The main control processing unit establishes electrical connections with the depth vision acquisition unit, the high-frequency pneumatic sensing unit, the variable frequency pneumatic pump module, the airflow direction switching valve module, the independent solenoid valve module, the linear displacement drive unit, and the heating sealing component. The main control processing unit receives point cloud data from the depth vision acquisition unit and pressure data from the high-frequency pneumatic sensing unit, and sends control commands to the pneumatic actuator and the heating sealing component according to the preset control logic.

[0031] The electrical control architecture is centered on the main control processing component, which interacts with the sensing and detection components and the pneumatic actuator components through electrical interfaces and power drive circuits to achieve signal interaction and energy transmission.

[0032] The main control processing component includes a microprocessor unit, a storage unit, and a multi-channel input / output interface. The main control processing component is also electrically connected to a power drive unit, which serves as the interface between low-voltage logic signals and high-voltage execution components.

[0033] The signal input terminals of the main control processing component are connected to the depth vision acquisition component and the high-frequency barometric pressure sensor component, respectively. The depth vision acquisition component transmits point cloud data frames and RGB image data frames to the main control processing component via a high-speed data transmission interface. The high-frequency barometric pressure sensor component converts the detected analog barometric pressure signal into a digital signal and sends it to the main control processing component at a preset sampling frequency.

[0034] The storage unit stores computer-readable instructions, and the microprocessor unit executes the computer-readable instructions to form logical functional modules, including a visual reconstruction analysis module, an aerodynamic impedance calculation module, and an adaptive flow field control module.

[0035] The visual reconstruction and analysis module receives point cloud data, performs background culling and coordinate transformation on the point cloud data, constructs a three-dimensional geometric model of the vacuum packaging bag and the objects inside the vacuum packaging bag, and calculates the real-time volume variables of the objects inside the vacuum packaging bag.

[0036] The aerodynamic impedance calculation module simultaneously processes pressure gradient data from the high-frequency pneumatic pressure sensing component and volumetric deformation data from the visual reconstruction and analysis module. The module performs a ratio calculation on the pressure gradient data and volumetric deformation data to obtain a virtual aerodynamic impedance value. Based on this virtual aerodynamic impedance value, it determines the physical property category of the object and generates the maximum allowable deformation threshold parameter and the target vacuum level parameter.

[0037] The adaptive flow field control module monitors the differences in local deformation rates across different areas of the vacuum packaging bag surface. When the difference in local deformation rate exceeds a preset deviation threshold, the adaptive flow field control module generates a differential control signal for the array-type differential suction nozzle module. The adaptive flow field control module also includes optical flow detection logic, which analyzes the fluid motion vector in the image within the pneumatic leveling time window and calculates the remaining time for the fluid front to reach the sealing position.

[0038] The control output of the main control processing unit is connected to the pneumatic actuator unit via the power drive unit. The main control processing unit sends a pulse width modulation signal to the power drive unit, which adjusts the drive voltage of the variable frequency pneumatic pump module in response to the pulse width modulation signal, thereby controlling the speed and torque of the variable frequency pneumatic pump module. The main control processing unit sends a level switching signal to the airflow direction switching valve module, controlling the airflow direction switching valve module to switch between the negative pressure suction passage and the positive pressure output passage.

[0039] The main control processing unit establishes a multi-channel control connection with the independent solenoid valve module through the power drive unit. Based on the differential control signal, the main control processing unit independently sends opening and closing commands to each solenoid valve in the independent solenoid valve module to control the on / off state of the corresponding independent intake unit's air path.

[0040] The main control processing unit is electrically connected to the linear displacement drive unit. The linear displacement drive unit is mechanically coupled to the array-type differential suction nozzle module. During the vacuum processing stage, the main control processing unit controls the linear displacement drive unit to send the independent suction unit into the working position; during the sealing preparation stage, the main control processing unit controls the linear displacement drive unit to retract the independent suction unit to the avoidance position.

[0041] The heating control terminal of the main control processing unit is connected to the heating sealing assembly via a power drive unit. After confirming that the independent suction unit has retreated to the avoidance position and the sealing triggering conditions are met, the main control processing unit sends a heating trigger signal to the power drive unit. The power drive unit then turns on the power supply of the heating sealing assembly and maintains the energized state according to preset time parameters to complete the sealing operation of the vacuum packaging bag.

[0042] This invention provides a multi-mode intelligent vacuum preservation machine identification method based on food ingredient recognition, the method comprising the following steps: Step S0: Initialization and 3D Scene Construction. After the system detects that the pressure-applying cover unit and base unit of the vacuum chamber assembly are closed and the vacuum packaging bag is positioned on the bearing plane, the main control processing unit sends an acquisition command to the depth vision acquisition unit. The depth vision acquisition unit scans the vacuum packaging bag on the bearing plane, acquiring the initial 3D point cloud data and RGB image data frames. The microprocessor unit runs the visual reconstruction and analysis module, which, based on the initial 3D point cloud data, constructs the initial surface topology model of the vacuum packaging bag and calculates the initial volume of the object to be preserved inside the vacuum packaging bag.

[0043] Step S0: Transient pneumatic excitation and active impedance detection. With the linear displacement drive unit maintaining the array-type differential suction nozzle module in the working position and the independent suction unit inside the opening of the vacuum packaging bag, the main control processing unit controls the variable frequency pneumatic pump module to output a short-term negative pressure pulse via the power drive unit. During the period when the variable frequency pneumatic pump module outputs the short-term negative pressure pulse, the high-frequency pressure sensing component collects pressure gradient data within the air path, and the depth vision acquisition component simultaneously collects the instantaneous deformation of the object inside the vacuum packaging bag. The microprocessor unit runs the pneumatic impedance calculation module, which determines the physical property category of the object inside the vacuum packaging bag based on the ratio of pressure gradient data to instantaneous deformation, and sets the target vacuum threshold and maximum allowable deformation rate threshold for subsequent pumping processes.

[0044] Step S0: Regional Differential Pumping and Flow Field Reconstruction. The main control processing unit controls the variable frequency pneumatic pump module to enter continuous pumping mode. In continuous pumping mode, the microprocessor unit runs the adaptive flow field control module, which continuously monitors the local deformation rate of different regions on the surface of the vacuum packaging bag. When it is calculated that the deformation rate of a certain region lags behind the average deformation rate and the difference exceeds a preset threshold, the adaptive flow field control module generates a differential control command. Based on the differential control command, the main control processing unit selectively closes the independent solenoid valve module corresponding to the high deformation rate region through the power drive unit, while keeping the independent solenoid valve module corresponding to the low deformation rate region in the open state, guiding the airflow inside the vacuum packaging bag from the deformation lag region to the independent suction unit.

[0045] Step S0: Pneumatic-optical Co-correction and Liquid Detection. During Step S0, the visual reconstruction analysis module evaluates the surface wrinkle complexity of the vacuum packaging bag opening area in real time. When the surface wrinkle complexity exceeds a preset threshold, the main control processing component pauses the pumping action of the variable frequency pneumatic pump module and controls the airflow direction switching valve module to switch to the positive pressure output path. The variable frequency pneumatic pump module outputs positive pressure micropulses to smooth out the wrinkles on the surface of the vacuum packaging bag. Within the time window of vacuum packaging bag smoothing, the depth vision acquisition component acquires images, and the adaptive flow field control module uses an optical flow algorithm to calculate the displacement velocity of the fluid front edge inside the vacuum packaging bag and the remaining time for the fluid front edge to reach the sealing position.

[0046] Step S500: Termination Judgment and Sealing Execution. The main control processing component monitors the system status in real time. When the air pressure value detected by the high-frequency air pressure sensor component reaches the target vacuum threshold, the object compression ratio calculated by the visual reconstruction analysis module reaches the maximum allowable deformation rate threshold, or the fluid remaining time predicted by the adaptive flow field control module is less than the safety threshold, the main control processing component determines that the termination conditions are met.

[0047] When the termination conditions are met, the main control processing unit stops the operation of the variable frequency pneumatic pump module. Subsequently, the main control processing unit sends a retraction command to the linear displacement drive unit, which drives the array-type differential suction nozzle module to retract from the opening of the vacuum packaging bag to a clearance position. The main control processing unit controls the independent solenoid valve module to fully close to perform a pressure balancing operation. After pressure balancing is completed, the main control processing unit activates the heating sealing assembly via the power drive unit to perform a heat-sealing process on the opening of the vacuum packaging bag.

[0048] The visual preprocessing and 3D reconstruction processes are executed by the visual reconstruction analysis module running from the main control processing component.

[0049] The visual reconstruction and analysis module is configured to read the raw data frames output by the depth vision acquisition component via a data interface. The raw data frames contain a two-dimensional RGB image matrix and a depth map matrix aligned with the RGB pixel coordinates. The storage unit stores the camera intrinsic parameter matrix and extrinsic parameter rotation and translation matrix of the depth vision acquisition component relative to the internal bearing plane of the vacuum cavity component. Using the camera intrinsic parameter matrix and extrinsic parameter rotation and translation matrix, the visual reconstruction and analysis module performs coordinate transformation on the effective pixels in the depth map matrix, mapping the effective pixels to a world coordinate system with the bearing plane as the XY reference plane and the vertical upward direction as the Z-axis, generating an initial 3D point cloud set.

[0050] The visual reconstruction and analysis module performs background culling to separate the target object. The storage unit contains a reference depth model of the vacuum cavity assembly in its unloaded state. The visual reconstruction and analysis module compares the spatial coordinates of discrete points in the initial 3D point cloud set with the spatial coordinates of the corresponding positions in the reference depth model. When the difference between the Z-axis height value of a discrete point and the height of the corresponding position in the reference depth model is less than a preset height difference threshold, the visual reconstruction and analysis module determines that the discrete point belongs to background noise or a bearing plane and removes it from the initial 3D point cloud set. After background culling, the remaining point cloud data constitutes a target point cloud set containing only the vacuum packaging bag and its internal objects. .

[0051] To improve the robustness of volumetric calculations, the visual reconstruction and analysis module processes a two-dimensional RGB image matrix. The module runs a semantic segmentation network to perform pixel-level classification of the RGB image matrix, generating a binarized mask to calibrate the vacuum packaging bag region. The module then maps the binarized mask onto the target point cloud set. Outliers located outside the range of the binarization mask are filtered out to obtain valid point cloud data.

[0052] Based on the effective point cloud data, the visual reconstruction and analysis module performs gridded resampling processing to calculate the initial volume of the object inside the vacuum packaging bag. The visual reconstruction analysis module divides the XY bearing plane into sections with a resolution of [resolution value missing]. The data is processed into regular micro-grids, and the effective point cloud data is projected into the corresponding regular micro-grids. The average height of the point cloud within each regular micro-grid is taken as the representative height of the current grid. Initial volume The calculation satisfies the following relationship:

[0053] in, This represents the total number of regular micro-grids that have valid height data after projection. Indices representing regular micromesh elements; and These represent the physical dimensions of the regular micro-element mesh along the X and Y axes in the world coordinate system, respectively; Indicates the first The representative height value of each regular micro-mesh relative to the bearing plane. The visual reconstruction analysis module will calculate the initial volume. It is stored in the storage unit and used as a reference value for subsequent calculation of volume compression ratio.

[0054] The transient aerodynamic excitation and impedance calculation process is executed by the main control processing component running the aerodynamic impedance calculation module in conjunction with the visual reconstruction analysis module.

[0055] In the initial stage of the vacuum processing process, the main control processing component first sends a full-open command to the independent solenoid valve modules to ensure that all independent solenoid valve modules stored in the storage unit are in the conducting state, so that the variable frequency pneumatic pump module forms a connected air path with the internal space of the vacuum packaging bag through the array-type differential suction nozzle module.

[0056] The main control processing unit controls the variable frequency pneumatic pump module to perform pulse excitation operation through the power drive unit. The variable frequency pneumatic pump module outputs a constant power negative pressure airflow within a preset short test time window, causing a step drop in the air pressure value of the air circuit system and inside the vacuum packaging bag.

[0057] Within the test time window, the high-frequency pneumatic pressure sensing component continuously acquires the absolute pressure value inside the pneumatic system at a preset sampling frequency and transmits it to the pneumatic impedance calculation module. The pneumatic impedance calculation module extracts the start time of the test time window. air pressure value With the end of the test time window air pressure value .

[0058] Synchronously with the high-frequency pressure sensing component, the depth vision acquisition component captures a sequence of deformation images of the vacuum packaging bag under negative pressure excitation. The visual reconstruction and analysis module invokes the aforementioned gridded resampling processing logic to calculate the start time of the test time window. Volume of objects inside vacuum packaging bags With the end of the test time window Volume of objects inside vacuum packaging bags .

[0059] The aerodynamic impedance calculation module performs a ratio calculation based on the change in air pressure and the change in object volume to calculate the virtual aerodynamic impedance value that characterizes the object's physical properties. Virtual aerodynamic resistance value The calculation follows the formula below:

[0060] in, and These represent the air pressure values ​​detected by the high-frequency air pressure sensing component at the start and end times, respectively. The reference volume at the initial moment; the denominator part It represents the volumetric strain rate under negative pressure excitation.

[0061] The aerodynamic impedance calculation module internally stores a physical property mapping table, which records the virtual aerodynamic impedance values ​​in different ranges. The correspondence between the control parameter group and the control parameter group. The aerodynamic impedance calculation module calculates the virtual aerodynamic impedance value. Matching with the physical property mapping table to determine the target vacuum threshold for the subsequent main pumping stage. and the maximum allowable deformation rate threshold .

[0062] When the virtual aerodynamic resistance value When the target vacuum level is within the soft region defined by the physical property mapping table, the aerodynamic impedance calculation module generates a lower target vacuum threshold value. and a smaller maximum allowable deformation rate threshold To prevent excessive compression of the contents inside the vacuum-sealed bag during the evacuation process. When the virtual pneumatic resistance value... When the target vacuum threshold is located within the hard region defined by the physical property mapping table, the aerodynamic impedance calculation module generates a higher target vacuum threshold value. This increases the amount of air expelled from inside the vacuum packaging bag. The main control processing unit will determine the target vacuum threshold. and the maximum allowable deformation rate threshold Set as the termination condition for system operation.

[0063] The mapping between the visual space and the air path execution space is established by the adaptive flow field control module and the visual reconstruction analysis module running in conjunction with the main control processing component.

[0064] The adaptive flow field control module establishes a rigid transformation relationship between the visual coordinate system and the physical airflow coordinate system. The storage unit stores the physical layout parameters of the array-type differential suction nozzle module in the world coordinate system. These physical layout parameters define the center point coordinates of each independent suction unit in the load-bearing plane coordinate system. ,in It is the physical number index for the independent intake unit, and all independent intake units are arranged linearly along the X-axis.

[0065] The adaptive flow field control module calls upon the 3D point cloud data of the vacuum packaging bag generated by the visual reconstruction analysis module. The visual reconstruction analysis module extracts the 2D projection contour of the vacuum packaging bag on the bearing plane and determines the minimum boundary coordinates of the vacuum packaging bag along the X-axis. with the maximum boundary coordinates .

[0066] The adaptive flow field control module performs region segmentation on the two-dimensional projected contour of the vacuum packaging bag based on the physical distribution of the independent suction units. The module calculates the arithmetic mean of the coordinates of the center points of two adjacent independent suction units as the virtual segmentation boundary. For index... Independent intake unit, virtual control sub-region corresponding to independent intake unit Coverage in the X-axis direction The following logic must be satisfied:

[0067]

[0068] in, Represents the lateral coordinates of the point cloud on the surface of the vacuum packaging bag; This represents the total number of independent intake units. The adaptive flow field control module traverses each discrete point in the target point cloud set, determining whether the lateral coordinate of the discrete point falls within the virtual control sub-region. The coverage area. If a discrete point falls within the coverage area, the adaptive flow field control module will mark the discrete point as belonging to the first... Each has an independent intake unit.

[0069] The adaptive flow field control module constructs a spatial mapping index table. This spatial mapping index table establishes virtual control sub-regions. Deformation eigenvalues ​​of internal point cloud data and the first The logical connections between individual intake units. The adaptive flow field control module, based on the spatial mapping index table, will target the virtual control sub-regions. The flow field adjustment requirements are transformed into opening and closing control commands for independent solenoid valve modules connected to the independent intake unit air path, realizing point-to-point air path control based on local deformation characteristics.

[0070] The local deformation rate monitoring and feedback process is executed by the adaptive flow field control module running by the main control processing component.

[0071] When the variable frequency pneumatic pump module is in continuous pumping operation, the adaptive flow field control module samples according to a preset time interval. Depth data frames are read from the depth vision acquisition component. The adaptive flow field control module calls the spatial mapping index table stored in the storage unit for each virtual control sub-region. Extracting the virtual control sub-region A subset of local point cloud data.

[0072] The adaptive flow field control module calculates each virtual control sub-region. The real-time average height value of a subset of local point cloud data. The adaptive flow field control module will use the current sampling time... average height value Compared with the previous sampling time average height value Perform difference operations to solve the problem corresponding to the first... Local deformation rate of each independent intake unit Local deformation rate The calculation follows the formula below:

[0073] The adaptive flow field control module calculates the local deformation rate of all virtual control sub-regions. The arithmetic mean is used to obtain the global average deformation rate. The adaptive flow field control module traverses each local deformation rate. Calculate the local deformation rate Relative to the global average deformation rate The adaptive flow field control module compares the rate deviation value with a preset rate deviation threshold.

[0074] When the Local deformation rate of a virtual control sub-region Below the global average deformation rate And when the absolute value of the difference exceeds the rate deviation threshold, the adaptive flow field control module determines the first... There is a pumping lag in one virtual control sub-region, and the first... The first virtual control sub-region is marked as a slow deformation region. Local deformation rate of a virtual control sub-region Higher than the global average deformation rate And when the absolute value of the difference exceeds the rate deviation threshold, the adaptive flow field control module determines the first... The virtual control sub-region has been significantly compressed, and the first... Each virtual control sub-region is marked as a rapid deformation region.

[0075] The adaptive flow field control module generates dynamic differential control commands based on the marking results. The module sends current-limiting modulation signals to the independent solenoid valve modules corresponding to the rapid deformation regions via the power drive unit. Responding to the current-limiting modulation signals, the power drive unit controls the independent solenoid valve modules corresponding to the rapid deformation regions to intermittently close or fully close using pulse width modulation (PWM), increasing the aerodynamic resistance flowing through these regions.

[0076] Meanwhile, the adaptive flow field control module maintains the independent solenoid valve modules corresponding to the slow deformation regions in a fully open state. By changing the suction weights of the array-type differential suction nozzle modules in the lateral distribution, the system guides the airflow inside the vacuum packaging bag from the slow deformation regions to the independent suction units corresponding to the slow deformation regions until the difference in the local deformation rates of all virtual control sub-regions is reduced to a preset equilibrium range.

[0077] The differential nozzle array control law is executed by the adaptive flow field control module running on the main control processing component.

[0078] The adaptive flow field control module is equipped with proportional control logic based on negative pressure feedback. The adaptive flow field control module reads the local deformation rate calculated in the previous control cycle. and global average deformation rate The adaptive flow field control module calculates the first... Rate deviation of individual inhalation units speed deviation Local deformation rate Subtract the global average deformation rate The algebraic difference.

[0079] The adaptive flow field control module is based on the rate deviation. Calculation for the first The target duty cycle of the independent solenoid valve module in each independent intake unit The storage unit stores the proportional gain coefficients used for flow field regulation. Basic conduction duty cycle constant and minimum maintenance duty cycle Target conduction duty cycle The calculation follows the piecewise control function as follows:

[0080] Among them, when the rate deviation When less than or equal to zero, the adaptive flow field control module determines the first... The deformation rate of the region corresponding to each independent intake unit is lagging or at the average level; the adaptive flow field control module will target the duty cycle. Set to 0%, output full-open command to provide maximum pumping flow. When the rate deviation... When the value is greater than zero, the adaptive flow field control module determines that the deformation rate of the corresponding region is too fast, and the adaptive flow field control module adjusts the rate deviation accordingly. The numerical value linearly reduces the target conduction duty cycle Until the target conduction duty cycle Achieve minimum maintenance duty cycle .

[0081] The adaptive flow field control module calculates the target duty cycle for each independent solenoid valve module. The signal is converted into a pulse width modulation (PWM) control signal with a fixed carrier frequency. The adaptive flow field control module sends the PWM control signal to the power drive unit through a digital output interface.

[0082] The power drive unit receives the pulse width modulation (PWM) control signal and amplifies it, driving the electromagnetic coil in the independent solenoid valve module to perform a high-frequency switching action. Within each PWM cycle, the ratio of the time the independent solenoid valve module remains open to the total duration of the PWM cycle is equal to the target duty cycle. The system establishes a gradient negative pressure distribution in the lateral dimension of the array-type differential suction nozzle module that matches the local deformation impedance of the vacuum packaging bag by adjusting the equivalent opening time of different independent solenoid valve modules in the array.

[0083] At the end of each control cycle, the adaptive flow field control module determines the rate deviation of all virtual control sub-regions. Is it within the preset balance dead zone range? If the rate deviation... Once the flow exceeds the equilibrium dead zone, the adaptive flow field control module enters the next sampling cycle for cyclic monitoring and calculation until the flow field reaches a convergent state.

[0084] The wrinkle detection and pneumatic leveling mechanism is executed by the main control processing component running the visual reconstruction analysis module in coordination with the airflow direction switching valve module.

[0085] The visual reconstruction and analysis module defines a region of interest (ROI) in the opening area of ​​the vacuum packaging bag. The ROI covers a predetermined sealing path below the heat-sealing assembly. The module reads depth point cloud data from the depth vision acquisition assembly within the ROI. It then calculates the depth gradient distribution along a cross-section perpendicular to the airflow direction within the ROI. Finally, the module calculates the variance of the depth gradient distribution and uses this variance as a wrinkle complexity index characterizing surface smoothness.

[0086] The visual reconstruction and analysis module compares the calculated wrinkle complexity index with a preset flatness threshold. When the wrinkle complexity index is less than or equal to the flatness threshold, the visual reconstruction and analysis module determines that the region of interest meets the sealing conditions, and the main control processing component maintains the current evacuation operation. When the wrinkle complexity index is greater than the flatness threshold, the visual reconstruction and analysis module determines that the region of interest has surface folding features and generates a flattening request signal.

[0087] In response to the leveling request signal, the main control processing unit executes the pneumatic leveling subroutine. First, it sends a full-conduction command to the independent solenoid valve module covering the area of ​​interest, forcing the independent solenoid valve module to remain normally open and ensuring unobstructed airflow. Then, it sends a reversing command to the airflow direction switching valve module. This valve module switches the internal airflow from a negative pressure suction path to a positive pressure output path, connecting the outlet of the variable frequency pneumatic pump module to the array-type differential suction nozzle module.

[0088] The main control processing unit controls the variable frequency pneumatic pump module to output a short-duration positive pressure airflow for a preset duration via the power drive unit. The short-duration positive pressure airflow is injected into the internal space of the vacuum packaging bag through an array of differential suction nozzle modules. The short-duration positive pressure airflow reverses the pressure difference between the inside and outside of the vacuum packaging bag, forcing the vacuum packaging bag to generate tension on the inner surface of the region of interest and bulge outward, eliminating any existing physical folds.

[0089] After the preset duration of the short-term positive pressure airflow ends, the main control processing component controls the airflow direction switching valve module to reset to the negative pressure extraction path and restarts the extraction operation of the variable frequency pneumatic pump module. The visual reconstruction analysis module re-acquires depth point cloud data of the region of interest and recalculates the wrinkle complexity index. The main control processing component counts the cumulative number of times the pneumatic leveling subroutine is executed. If the recalculated wrinkle complexity index is still higher than the flatness threshold and the cumulative number of times has not reached the preset maximum retry threshold, the main control processing component repeats the pneumatic leveling subroutine. If the cumulative number of times reaches the preset maximum retry threshold, the main control processing component outputs an abnormal alarm signal and terminates system operation.

[0090] The process of monitoring and blocking liquid diffusion risk is executed by the main control processing component running the liquid diffusion monitoring module in conjunction with the visual reconstruction and analysis module.

[0091] The liquid diffusion monitoring module establishes a region of interest (ROI) for liquid monitoring within a two-dimensional RGB image matrix. The physical extent of the ROI is limited to the area between the leading edge of the object inside the vacuum-sealed bag and the physical compression centerline of the heat-sealing assembly. The module extracts the contour boundary coordinates of the object inside the vacuum-sealed bag on the bearing plane and extends these coordinates along the airflow direction to the physical compression centerline of the heat-sealing assembly, thus defining the ROI that contains only potential liquid flow paths.

[0092] During the pumping operation performed by the main control processing component, the liquid diffusion monitoring module receives continuous time-series image frames from the depth vision acquisition component at a fixed frame rate. The liquid diffusion monitoring module selects the current moment... Image frame and the previous moment Image frames are used as input data pairs. The liquid diffusion monitoring module uses a dense optical flow algorithm to calculate the two-dimensional velocity vector of each pixel within the region of interest for liquid monitoring. Two-dimensional velocity vector Decomposed into lateral components perpendicular to the suction direction With the longitudinal component parallel to the pumping direction The vertical component is defined here. The positive direction points to the side where the heating sealing assembly and the array-type differential suction nozzle module are located.

[0093] The liquid diffusion monitoring module performs motion vector filtering to eliminate background motion interference caused by surface shrinkage of the vacuum packaging bag. The liquid diffusion monitoring module sets a longitudinal flow velocity threshold. Longitudinal velocity threshold The value is set to be greater than the maximum physical shrinkage rate of the vacuum packaging bag under high negative pressure. The liquid diffusion monitoring module filters out the longitudinal component. The value is greater than the longitudinal velocity threshold. The effective pixels. The liquid diffusion monitoring module marks the position with the largest vertical coordinate value among all effective pixels as the real-time liquid front position. .

[0094] The liquid diffusion monitoring module is based on the real-time liquid front position. The instantaneous longitudinal propulsion velocity of the liquid flow is calculated from the rate of displacement change between consecutive frames. The liquid diffusion monitoring module reads the pre-stored sealing baseline coordinates of the heating sealing assembly from the storage unit. The liquid diffusion monitoring module predicts the remaining safe time for the liquid to reach the sealing baseline based on the following kinematic formula. :

[0095] in, This represents the longitudinal intercept of the heat sealing assembly in a unified world coordinate system; This represents the vertical coordinate of the liquid's foremost point at the current moment.

[0096] The liquid diffusion monitoring module will calculate the remaining safety time. Compare with the system's preset emergency response time threshold. When the remaining safe time... When the time is less than or equal to the emergency response time threshold, the liquid diffusion monitoring module determines that a liquid intrusion into the gas path event is about to occur and immediately generates a liquid blockage interruption signal.

[0097] In response to the liquid interruption signal, the main control processing unit executes an emergency sealing procedure. It prioritizes sending a full shutdown command to all independent solenoid valve modules, severing the airflow connection between the array-type differential nozzle module and the variable frequency pneumatic pump module, thus physically isolating the pump to prevent liquid from entering. Simultaneously, the main control processing unit drives the mechanical pressing mechanism of the heat-sealing assembly to close the vacuum bag, pressing the film at the bag opening tightly. It also activates the heating element of the heat-sealing assembly to perform a heat-sealing process, completing the sealing of the vacuum bag before the liquid reaches the sealing baseline.

[0098] The termination determination process of the air extraction process is executed by the main control processing component's process termination determination module 45.

[0099] The main control processing component continuously monitors multi-dimensional system status parameters during the main air extraction phase. The process termination determination module 45 establishes parallel monitoring channels to receive real-time absolute pressure values ​​of the air path from the high-frequency pressure sensing component. And receive the real-time calculated volume of the object inside the vacuum-packed bag from the visual reconstruction analysis module. .

[0100] Process termination determination module 45 calls the target vacuum threshold determined in the aerodynamic impedance calculation stage. and the maximum allowable deformation rate threshold The process termination determination module 45 calculates the volumetric strain rate of the object inside the vacuum packaging bag in real time. Volumetric strain rate The calculation formula is as follows:

[0101] in, The initial volume of the object inside the vacuum packaging bag is recorded at the moment when the main control processing component starts the variable frequency pneumatic pump module.

[0102] The process termination determination module 45 determines whether the gas extraction process should be terminated based on the following logical conditions. The process termination determination module 45 will use the real-time absolute pressure value of the gas path. With the target vacuum threshold Perform numerical comparisons. If the real-time absolute pressure value of the gas path... Below or equal to the target vacuum threshold The process termination judgment module 45 generates a gas pressure compliance termination signal.

[0103] Meanwhile, the process termination determination module 45 will calculate the volumetric strain rate in real time. With the maximum allowable deformation rate threshold Perform numerical comparisons. If the volumetric strain rate... Greater than or equal to the maximum allowable deformation rate threshold This indicates that the deformation of the object inside the vacuum packaging bag has reached the preset physical protection limit, and the process termination judgment module 45 generates a deformation limit termination signal.

[0104] In addition, the process termination determination module 45 monitors the continuous running time of the main air extraction stage. If the continuous running time exceeds a preset safety timeout threshold, the process termination determination module 45 generates a timeout forced termination signal.

[0105] The main control processing unit responds in real time to pressure compliance termination signals, deformation limit termination signals, and timeout forced termination signals. When any of these termination signals is triggered, the main control processing unit immediately executes the main air extraction process interruption procedure. The main control processing unit sends a shutdown command to the variable frequency pneumatic pump module to stop the negative pressure output.

[0106] While stopping the negative pressure output, the main control processing unit sends a full shutdown command to all independent solenoid valve modules, cutting off the connection between the internal space of the vacuum packaging bag and the external air path to prevent air backflow. Subsequently, the main control processing unit drives the array-type differential suction nozzle module to perform a retraction action or drives the heating sealing assembly to perform a mechanical pressing action, ensuring that there are no foreign objects interfering with the pressing path of the heating sealing assembly. The main control processing unit controls the mechanical pressing mechanism of the heating sealing assembly to press the bag opening film of the vacuum packaging bag with a preset pressure value, and activates the heating element of the heating sealing assembly to perform a heat-melt sealing operation, completing the vacuum preservation and sealing process.

[0107] The hysteresis balancing and sealing control process is executed by the main control processing component running the sealing compensation module in coordination with the heating sealing component.

[0108] After confirming that the mechanical pressing mechanism of the heating and sealing assembly has completed its closing action, the main control processing component activates the sealing compensation module to execute the hysteresis balancing program. The sealing compensation module starts the hysteresis balancing timer and sets the hysteresis balancing time interval. During the hysteresis equilibrium time interval Inside, the main control processing component maintains the clamping force of the mechanical pressing mechanism of the heat sealing assembly and prohibits power supply to the heating element of the heat sealing assembly. The hysteresis balancing program is used to eliminate the physical wrinkle rebound force of the vacuum packaging bag in the pressing area before heating, so that the upper and lower films of the vacuum packaging bag can achieve flat bonding under mechanical pressure.

[0109] The sealing compensation module during the hysteresis balance time interval After completion, the nonlinear thermal melting control stage begins. The sealing compensation module reads the current ambient temperature data collected by the ambient temperature sensor. And call the preset material heat transfer coefficient in the storage unit. and sealing thickness grade The sealing compensation module calculates the total target Joule heat required to complete the sealing process based on the following energy control formula. :

[0110] in, This is the preset theoretical melting temperature of the material.

[0111] The sealing compensation module generates a three-segment pulse width modulation (PWM) drive signal to control the heating element of the heating sealing assembly. The three-segment PWM drive signal includes a heating impact stage, a isothermal fusion stage, and a cooling curing stage.

[0112] During the temperature rise shock phase, the sealing compensation module outputs a full-power drive signal with a 0% duty cycle. The sealing compensation module is based on the rated power of the heating element. Calculate the duration of the temperature rise The sealing compensation module controls the heating element to heat up rapidly, thereby shortening the time it takes for heat to be conducted to unsealed areas.

[0113] Duration of temperature rise After completion, the system enters the isothermal fusion stage. The sealing compensation module reduces the duty cycle of the drive signal to maintain the duty cycle. Maintain duty cycle The value is set to a power level that can balance the heat dissipation of the heating element to the environment. The sealing compensation module performs time integration calculation on the output power, and when the cumulative output heat energy reaches the total target Joule heat... At that time, the sealing compensation module determines that the fusion is complete.

[0114] The cooling and curing stage then begins. The sealing compensation module stops supplying power to the heating element, reducing its duty cycle to 0%. The main control processing unit continues to maintain the mechanical pressing mechanism of the heating and sealing assembly in a high-pressure closed state. The sealing compensation module starts the curing timer and sets the curing cooling time. During the curing and cooling time Inside, the molten sealing material cools and solidifies under pressure, forming a sealing weld.

[0115] When the curing timer expires, the main control processing unit controls the mechanical pressing mechanism of the heating sealing assembly to perform an opening and releasing action. The main control processing unit sends a reset command to the variable frequency pneumatic pump module and all independent solenoid valve modules, and the system returns to its initial standby state.

[0116] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A multi-mode intelligent vacuum preservation machine based on food ingredient recognition, characterized in that, include: A vacuum chamber assembly that defines a sealed operating space and has a heat-sealing assembly for sealing a vacuum packaging bag inside the vacuum chamber assembly. The sensing and detection component includes a depth vision acquisition component disposed above the vacuum cavity component and a high-frequency pressure sensing component disposed in the main gas pipeline; the depth vision acquisition component is configured to acquire three-dimensional point cloud data of the vacuum packaging bag and the object inside, and the high-frequency pressure sensing component is configured to detect the absolute pressure value and instantaneous fluctuation rate of pressure change inside the gas pipeline system. The pneumatic actuator includes a variable frequency pneumatic pump module connected to the vacuum chamber assembly via an air passage, an airflow direction switching valve module, and an array-type differential suction nozzle module located inside the vacuum chamber assembly. The array-type differential suction nozzle module includes at least three independent suction units arranged linearly in the transverse direction. Each independent suction unit has an air path connected in series with an independent solenoid valve module. The independent solenoid valve modules are connected to the variable frequency pneumatic pump module after they converge. The main control processing component is communicatively connected to the depth vision acquisition component, the variable frequency pneumatic pump module, the airflow direction switching valve module, and the independent solenoid valve module, respectively. The main control processing component is configured to: monitor the local deformation rate of the surface of the vacuum packaging bag through the depth vision acquisition component, and generate differential control signals for the array-type differential suction nozzle module based on the distribution difference of the local deformation rate, and independently control the on / off state of each of the independent solenoid valve modules to adjust the suction force distribution at different lateral positions at the opening of the vacuum packaging bag.

2. The multi-mode intelligent vacuum preservation machine based on food identification according to claim 1, characterized in that, The array-type differential suction nozzle module is also mechanically coupled with a linear displacement drive unit; the main control processing component controls the linear displacement drive unit to drive the independent suction unit to reciprocate between the working position extending into the opening of the vacuum packaging bag and the avoidance position exiting the opening of the vacuum packaging bag. The main control processing component is equipped with an adaptive flow field control module. The adaptive flow field control module divides the two-dimensional projection outline of the vacuum packaging bag into virtual control sub-regions corresponding to the independent air intake units based on the physical layout parameters of the independent air intake units.

3. The multi-mode intelligent vacuum preservation machine based on food identification according to claim 1, characterized in that, The main control processing component runs the pneumatic impedance calculation module, which is configured to control the variable frequency pneumatic pump module to output a short-term negative pressure pulse in the initial stage of the air extraction process, and calculate the virtual pneumatic impedance value based on the ratio of the pressure gradient data collected by the high-frequency air pressure sensing component to the change in object volume collected by the depth vision acquisition component. The aerodynamic impedance calculation module determines the physical property category of the object inside the vacuum packaging bag based on the virtual aerodynamic impedance value, and sets the target vacuum degree threshold and the maximum allowable deformation rate threshold for the air extraction process accordingly.

4. A multi-mode intelligent vacuum preservation machine based on food ingredient recognition according to claim 2, characterized in that, The main control processing component runs an adaptive flow field control module, configured to calculate the rate deviation of the local deformation rate of each virtual control sub-region relative to the global average deformation rate. When the rate deviation indicates that a certain virtual control sub-region is a rapidly deforming region, the main control processing component sends a pulse width modulation signal to the corresponding independent solenoid valve module to intermittently or completely close the gas path. When the rate deviation indicates that a certain virtual control sub-region is a slow deformation region, the main control processing component controls the corresponding independent solenoid valve module to remain in a fully conductive state, guiding the airflow out of the slow deformation region.

5. A multi-mode intelligent vacuum preservation machine based on food ingredient recognition according to claim 1, characterized in that, The main control processing component runs the visual reconstruction analysis module, which is configured to calculate the variance of the depth gradient distribution of the vacuum packaging bag in the region of interest below the heating sealing component, as the wrinkle complexity index. When the wrinkle complexity index is greater than the preset flatness threshold, the main control processing component controls the airflow direction switching valve module to switch to the positive pressure output path, and controls the variable frequency pneumatic pump module to output a short-term positive pressure airflow to the array differential suction nozzle module, forcing the surface of the vacuum packaging bag to generate tension and bulge outward to eliminate the physical fold structure.

6. A multi-mode intelligent vacuum preservation machine based on food ingredient recognition according to claim 1, characterized in that, The main control processing component runs a liquid diffusion monitoring module, which is configured to use an optical flow algorithm to calculate the two-dimensional velocity vector of the liquid front inside the vacuum packaging bag, and filter out effective pixels whose longitudinal component is greater than a preset longitudinal flow velocity threshold, wherein the longitudinal flow velocity threshold is greater than the maximum physical shrinkage rate of the vacuum packaging bag under high negative pressure. The liquid diffusion monitoring module predicts the remaining safe time based on the real-time position of the liquid front and the sealing baseline coordinates of the heating sealing assembly. When the remaining safety time is less than the emergency response time threshold, the main control processing component sends a full shutdown command to all the independent solenoid valve modules and simultaneously drives the heating sealing component to perform mechanical pressing and heat fusion sealing.

7. A multi-mode intelligent vacuum preservation machine based on food ingredient recognition according to claim 3, characterized in that, The main control processing component's process termination determination module (45) is configured to establish parallel monitoring channels; the main control processing component determines that the air extraction process is terminated when any of the following conditions are met: The real-time absolute pressure value of the air path detected by the high-frequency air pressure sensing component is lower than or equal to the target vacuum threshold. The object volume strain rate calculated by the depth vision acquisition component is greater than or equal to the maximum allowable deformation rate threshold. The continuous operation time of the main air extraction stage exceeds the preset safety timeout threshold.

8. A multi-mode intelligent vacuum preservation machine based on food ingredient recognition according to claim 1, characterized in that, The main control processing component operates the sealing compensation module, which is configured to start a hysteresis balancing program after the mechanical pressing mechanism of the heating sealing component closes. During the preset hysteresis balance time interval, the main control processing component maintains the clamping force of the mechanical pressing mechanism and prohibits power supply to the heating sealing component to reduce the physical rebound force of the vacuum packaging bag in the pressing area; after the hysteresis balance time interval ends, the main control processing component controls the heating sealing component to start heating.

9. A multi-mode intelligent vacuum preservation machine based on food ingredient recognition according to claim 8, characterized in that, The sealing compensation module calculates the target thermal energy based on ambient temperature data and the preset material heat melting coefficient, and generates a three-segment pulse width modulation drive signal to control the heating sealing component. The three-segment pulse width modulation drive signal includes: a heating impact stage with an output duty cycle of 0%, a constant temperature fusion stage with the output duty cycle maintained to balance heat dissipation, and a cooling and solidification stage with an output duty cycle of 0% and mechanical pressing maintained; the sealing compensation module integrates the output power in the constant temperature fusion stage, and stops heating when the accumulated output heat reaches the target thermal energy.

10. A multi-mode intelligent vacuum preservation machine identification method based on food ingredient recognition, based on implementing the multi-mode intelligent vacuum preservation machine based on food ingredient recognition as described in any one of claims 1 to 9, characterized in that, The method includes the following steps: The initial 3D point cloud data of the vacuum packaging bag is obtained using a depth vision acquisition component, and an initial surface topology model is constructed. The variable frequency pneumatic pump module outputs a short-term negative pressure pulse, and the virtual pneumatic impedance value is calculated based on the ratio of pressure change to volume deformation to determine the physical property category of the object inside the vacuum packaging bag. During continuous evacuation, the local deformation rate of different areas on the surface of the vacuum packaging bag is monitored in real time, and the deviation between the local deformation rate and the global average deformation rate is calculated. Based on the deviation, a differential control signal is generated to independently adjust the opening and closing state of each independent solenoid valve module connected to the back end of the array-type differential nozzle module, thereby limiting the flow in the rapid deformation area and keeping the slow deformation area fully conductive. The system monitors the complexity of the folds in the vacuum packaging bag and the risk of internal liquid diffusion in real time, and stops pumping when the preset air pressure threshold or deformation rate threshold is met, driving the heating sealing component to complete the sealing.

Images