A nondestructive bionic detection device and method for mutton freshness

Abstract

The application provides a mutton freshness nondestructive biomimetic detection device and method, relates to the flexible sensing technical field, and the device comprises a flexible sensor array, a transverse transmission rod, a longitudinal transmission rod sleeve, a sticky probe and the like components. The sensor is connected through a guide rail and a fixing device, accurate detection of the freshness of mutton is realized. The device is flexible and accurate, a variety of sensors such as a pressure sensor, a mechanical vision sensor and the like work cooperatively, and the state of mutton can be comprehensively detected. The operation is simple, and a nondestructive detection method ensures product quality and food safety.

Description

Technical Field

[0001] This invention relates to the field of flexible sensing technology, and in particular to a non-destructive biomimetic detection device and method for the freshness of mutton. Background Technology

[0002] Meat is susceptible to spoilage and deterioration during storage, transportation, and processing due to the action of enzymes and microorganisms, leading to a decline in quality. Lamb, in particular, is geographically distant from its supply and demand and heavily reliant on imports, making its freshness a concern for both consumers and regulators. Research and literature review indicate that meat surface characteristics are closely related to its freshness. However, existing machine vision inspection methods for assessing lamb surface characteristics have limited applicability. Therefore, improving the recognition accuracy of lamb surface characteristics to achieve more precise freshness prediction is crucial. Currently, domestic research in this area is not in-depth. The emergence of multimodal data fusion technology has provided new methods for extracting surface features. Summary of the Invention

[0003] In order to overcome the shortcomings of the existing technology, the purpose of this invention is to provide a non-destructive biomimetic detection device and method for mutton freshness.

[0004] To achieve the above objectives, the present invention provides the following solution:

[0005] A non-destructive biomimetic detection device for mutton freshness includes: a flexible sensor array, a transverse transmission rod, a longitudinal transmission rod sleeve, a longitudinal transmission rod tip, a transverse-to-vertical axis converter, a viscous detector guide rail fixing device, a pressure detector guide rail fixing device, a viscous detector integrated device, a viscous probe sleeve, a flexible strain sensor, a pressure sensor probe, a flexible pressure sensor, a machine vision sensor, a stage, a housing, a door, a control panel, and a host computer.

[0006] The adhesive probe sleeve is connected to the adhesive detector integrated device via an axial sliding pair; the adhesive detector integrated device is fixed below the adhesive detector guide rail fixing device by threads; the transverse transmission rod passes through the entire adhesive detector guide rail fixing device via two bearings.

[0007] The pressure sensor probe is connected to the lower part of the pressure detector guide rail fixing device via an axial sliding joint; the transverse transmission rod passes through the pressure detector guide rail fixing device via a bearing; the transverse and longitudinal axis converter is connected to one end of the transverse transmission rod; the longitudinal transmission rod is sleeved on the shaft hole below the transverse and longitudinal axis converter; the tip of the longitudinal transmission rod is threaded to the longitudinal transmission rod sleeve; the tip of the longitudinal transmission rod is fixed to a preset position on the housing; the platform is connected to the housing with screws; the machine vision sensor is fixed to the corresponding position on the housing; the flexible sensor array is fixed to the platform by embedding; the door is connected to the housing via a door hinge; the control panel is mounted on the housing; the flexible strain sensor is pasted to the working surface below the adhesive probe sleeve; the flexible pressure sensor is pasted to the working surface below the pressure sensor probe; the host computer is connected to the flexible pressure sensor, the machine vision sensor, and the flexible strain sensor respectively; the flexible sensor array is used to acquire pressure change signals reflecting changes in pressure values.

[0008] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0009] This invention provides a non-destructive biomimetic detection device and method for mutton freshness. The device includes: a flexible sensor array, a transverse transmission rod, a longitudinal transmission rod sleeve, a longitudinal transmission rod tip, a transverse-to-vertical axis converter, a viscous detector guide rail fixing device, a pressure detector guide rail fixing device, a viscous detector integrated device, a viscous probe sleeve, a flexible strain sensor, a pressure sensor probe, a flexible pressure sensor, a machine vision sensor, a stage, a housing, a door, a control panel, and a host computer. The viscous probe sleeve is connected to the viscous detector integrated device via an axial sliding joint. The viscous detector integrated device is threaded to the lower part of the viscous detector guide rail fixing device. The transverse transmission rod passes through the entire viscous detector guide rail fixing device through two bearings. The pressure sensor probe is connected to the lower part of the pressure detector guide rail fixing device through an axial sliding joint. The transverse transmission rod passes through the pressure detector guide rail fixing device through the bearings. The device includes a guide rail fixing device; the horizontal and vertical axis converter is connected to one end of the horizontal transmission rod; the vertical transmission rod is sleeved on the shaft hole below the horizontal and vertical axis converter; the tip of the vertical transmission rod is threaded to the vertical transmission rod sleeve; the tip of the vertical transmission rod is fixed to a preset position on the box body; the platform is connected to the box body with screws; the machine vision sensor is fixed to the corresponding position on the box body; the flexible sensor array is fixed to the platform by embedding; the door is connected to the box body by a door hinge; the control panel is installed on the box body; the flexible strain sensor is pasted to the working surface below the adhesive probe sleeve by adhesive; the flexible pressure sensor is pasted to the working surface below the pressure sensor probe by adhesive; the host computer is connected to the flexible pressure sensor, the machine vision sensor, and the flexible strain sensor respectively; the flexible sensor array is used to acquire pressure change signals reflecting pressure value changes. This invention uses a variety of sensors and detectors, including flexible sensors, pressure sensors, and machine vision sensors, to detect the freshness of mutton with high precision. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a schematic diagram of the non-destructive release detection device for mutton freshness based on multimodal data fusion and flexible sensors provided in an embodiment of the present invention.

[0012] Figure 2 This is a schematic diagram of the box structure provided in an embodiment of the present invention;

[0013] Figure 3 This is a schematic diagram of the detection method provided in an embodiment of the present invention;

[0014] Figure 4 This is a schematic diagram of the process for locating the sample center using a flexible sensor array, provided in an embodiment of the present invention.

[0015] Figure 5 This is a schematic diagram illustrating the response process of a flexible sensor array after receiving an external signal, as provided in an embodiment of the present invention.

[0016] Figure 6 This is a schematic diagram of the process for determining the number of viscosity sensor probes provided in an embodiment of the present invention;

[0017] Figure 7 This is a schematic diagram of the positioning and recognition process of the machine vision device provided in an embodiment of the present invention;

[0018] Figure 8 This is a schematic diagram of the flesh color scoring process provided in an embodiment of the present invention;

[0019] Figure 9 This is a schematic diagram of the network model flow provided in an embodiment of the present invention;

[0020] Figure 10 This is a schematic diagram of the overall detection process provided in an embodiment of the present invention;

[0021] Figure 11 This is a schematic diagram of the information flow provided for an embodiment of the present invention.

[0022] Explanation of reference numerals in the attached figures:

[0023] 1. Flexible sensor array; 2. Lateral transmission rod; 3. Longitudinal transmission rod sleeve; 4. Longitudinal transmission rod tip; 5. Horizontal-to-vertical axis converter; 6. Viscous detector guide rail fixing device; 7. Pressure detector guide rail fixing device; 8. Viscous detector integrated device; 9. Viscous probe sleeve; 10. Flexible strain sensor; 11. Pressure sensor probe; 12. Flexible pressure sensor; 13. Machine vision sensor; 14. Stage; 15. Housing; 16. Door; 17. Control panel; 18. Host computer. Detailed Implementation

[0024] The technical solutions of 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.

[0025] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0026] Figure 1 This is a schematic diagram of the non-destructive release detection device for mutton freshness based on multimodal data fusion and flexible sensors provided in an embodiment of the present invention. Figure 1 and Figure 2As shown, this device includes a flexible sensor array 1, a transverse transmission rod 2, a longitudinal transmission rod sleeve 3, a longitudinal transmission rod tip 4, a transverse-to-vertical axis converter 5, a viscous detector guide rail fixing device 6, a pressure detector guide rail fixing device 7, a viscous detector integrated device 8, a viscous probe sleeve 9, a flexible strain sensor 10, a pressure sensor probe 11, a flexible pressure sensor 12, a machine vision sensor 13, a stage 14, a housing 15, a door 16, a control panel 17, and a host computer 18; the flexible sensor array 1 includes 1600 microsensors evenly distributed in a 400mm×400mm transmission... The system comprises a sensor array, a sensor protective layer, and a signal output terminal; the micro-sensor is a pressure sensor, which reacts to changes in pressure above it; the sensor protective layer is made of a thin, flexible material; the flexible strain sensor includes metal electrodes, a flexible base film, and a signal output terminal; the flexible pressure sensor is manufactured using the same method as the flexible strain sensor, but its electrode shape and material differ; the housing includes a lighting system and fixed positions for each device; the adhesive probe sleeve and the adhesive detector integrated device are connected using an axial sliding joint; the adhesive detector integrated device is threaded onto the adhesive detector fixing device. Below; the transverse transmission rod passes through the upper part of the adhesive detector fixing device via two bearings; the pressure sensor probe is connected below the pressure detector fixing device via an axial sliding pair; the transverse transmission rod passes through the pressure detector fixing device via the bearings; the transverse and longitudinal axis converter is connected to one end of the transverse transmission rod; the longitudinal transmission rod sleeve is connected via a shaft hole below the transverse and longitudinal axis converter; the tip of the longitudinal transmission rod is connected to the longitudinal transmission rod sleeve via a thread, the thread length is relatively long, and the length can be changed by twisting the thread; the tip of the longitudinal transmission rod is fixed at the corresponding position of the box; the stage is connected to the box using screws; the machine vision sensor is fixed at the corresponding position of the box; the flexible sensor array is fixed to the stage by embedding, its geometric center coincides with the geometric center of the stage; the door is connected to the box via a door hinge; the control panel is installed at the corresponding position of the box; the flexible strain sensor is pasted on the working surface below the adhesive sensor probe by adhesive; the flexible pressure sensor is pasted on the working surface below the pressure sensor probe by adhesive.

[0027] Specifically, the function of the metal electrode is to detect the stickiness of the meat by objectively relating its own strain to the stickiness of the meat; the function of the second signal output terminal is to output the signal generated by the metal electrode.

[0028] Specifically, the built-in LED light in the housing 15 is turned on for illumination, and the flexible sensor array 1 is used to establish the XY coordinates of the device. When the unknown meat piece is placed on the flexible sensor array, the meat piece will exert pressure on the micro-sensors below, causing a change in the electrical signal of the micro-sensors in that area, thereby detecting the specific placement position of the meat. After obtaining the coordinates of the unknown meat piece, the longitudinal transmission rod sleeve 3, the longitudinal transmission rod tip 4, and the transverse transmission rod 2 move the adhesive detector integrated device 7 to the geometric center of the meat piece. The adhesive probe sleeve 6 moves downward and stops when it contacts the surface of the meat, and then moves in the opposite direction. When the adhesive probe sleeve moves upward, the flexible strain... As the deformation rate of the sensor increases, the electrical signal it transmits changes. When the strain of the flexible strain sensor reaches its maximum value and detaches from the meat surface, the adhesive detector fixing device moves to the edge of the box, the pressure detector fixing device moves to the geometric center of the meat, and the pressure sensor probe 11 descends rapidly. When it contacts the meat surface, it slows down. When the pressure signal transmitted by the flexible pressure sensor 12 reaches a threshold, the pressure sensor probe 11 lifts up. At this time, the mechanical vision sensor 13 detects the pit recovery time and grades the meat color. All signals are output to the host computer 18 for classification through the signal output terminal of the detection part.

[0029] Optionally, the interior of the box 15 is white or a bright color other than red, which makes the meat color more vivid, contrasting with the environment and facilitating machine vision recognition. There are fixing ports for longitudinal transmission rods on both sides of the upper part of the box interior for securing the aforementioned transmission device. The bottom of the box interior features a boss design, reducing material consumption while increasing space for the flexible sensor array circuitry. Recesses are provided at the midpoints of the four sides of the stage, also for this purpose. A camera probe fixing device is located in the middle of the right side of the box, while lighting device fixing devices are located on the left and rear sides.

[0030] When the viscous sensor probe is pressed down, the surface of the flexible strain sensor comes into contact with the meat surface. When the probe is pulled up, the strain of the flexible strain sensor increases, reaching its maximum when it detaches from the meat surface. The viscous characteristic index of the meat is estimated by the maximum strain of the flexible strain sensor when the probe is pulled up. The pressure sensor is pressed down on the meat surface by the same amount, and the recovery time of the indentation is judged by the vision sensor, thereby estimating the elastic characteristic index of the meat. The mechanical vision sensor consists of an image capture device and an image processing device. The mechanical vision sensor consists of an image capture device and a signal transmission device. Its working principle is that image information is acquired at regular intervals through an image capture device such as a camera. This information is in RGB three-channel mode and is transmitted to the image processing device through the signal transmission device. The image processing device is a neural network developed based on PyTorch. Its working principle is: input an image, process it through the neural network, and determine its meat color characteristics. The sensor array module consists of a pressure flexible sensor array and a protective film. Its working principle is: establish an XY coordinate system with a vertex of the flexible sensor array as the origin. When meat is placed on the tray, it sends a pressure signal to the flexible sensor array. The position of the meat is determined by the coordinates of the pressure signal in the XY coordinate system, and the height of the meat is predicted by the magnitude of the pressure signal per unit area.

[0031] The lighting device is fixed in place by a lighting device consisting of 6 LED tubes, which are placed on the upper, middle and lower parts of the left rear side of the fixed box. The power management module of the working part is connected to all the above working modules.

[0032] The adhesive probe sleeve has a large radius of curvature on its working surface, designed to mimic the arc of a human finger and reduce probe damage to the flesh. The flexible strain sensor is fabricated from PDMS, PI, PEN, or PET materials using laser direct writing or inkjet printing technology, with graphene or metal electrodes. The sensor is connected to the sensor probe via adhesive bonding at both ends. The pressure sensor probe is also fabricated from PDMS, PI, PEN, or PET materials using laser direct writing or inkjet printing technology, with graphene or metal electrodes.

[0033] The pressure sensor probe has a small radius of curvature on its working surface, designed to induce larger localized deformation in the meat, thus better determining the meat's resilience coefficient. The sensor is connected to the sensor probe via adhesive bonding at both ends.

[0034] The flexible sensor array consists of 400×400 pressure sensors on the operating platform. A planar XY Cartesian coordinate system is used to define the coordinates. By default, the leftmost pressure sensor on the side of the operating platform closest to the door is designated as (0,0), with the positive X and Y axes pointing to the right and inwards towards the operating platform, respectively. Each pressure sensor converts the pressure received into a signal, transmits it, and stores it in the system as data for subsequent processing. The coordinates of sensors with significantly increased pressure values ​​are aggregated to define the spatial location range for detecting the meat sample. The system repeatedly measures the pressure three times, and the coincident coordinates of each set are taken as the position coordinates of the meat sample. This coordinate information is stored in the microcontroller for subsequent sensor measurement control. Simultaneously, the average pressure value from these coordinates is processed, and the height h of the meat is predicted based on the magnitude of the pressure signal per unit area. s It is also stored in the system.

[0035]

[0036] Among them, P s Pressure per unit volume; A o ρ is the area of ​​the pressure sensor unit. meat The density of the meat block being measured (set based on the average density of meat block samples of this variety).

[0037] Preferably, the positioning of each sensor is determined by a top-mounted retainer. This retainer, driven by electricity, moves along the transverse, longitudinal, and converter transmission rods according to the operating program. Each moving unit corresponds to the smallest unit size of the pressure sensor array. The movement control plane formed by the transmission rods constitutes a planar XY Cartesian coordinate system. Preferably, during system initialization, the (0,0) coordinate of this coordinate system is defined as the sensor retainer moving directly above the (0,0) pressure sensor. The calibration method is that the moving pressure sensor probe generates a pressure signal abrupt change simultaneously with only the (0,0) pressure flexible sensor. Each moving unit is defined as representing a coordinate value change, and the position P of the sensor probe is accurately represented by coordinates. sensor (x k ,y k The input is stored in the microcontroller, which facilitates the running of the program to control the movement of the sensor probe for measurement.

[0038] Preferably, a neural network model is used to fit the set of coordinate points reflecting the location of the detected meat sample to a regular shape such as a rectangle or circle. Since the selected meat sample is generally close to a regular shape, this method facilitates the selection of sampling points, and the sampling points are representative. Then, the geometric center of the fitted shape is taken as the center point P. cFor a rectangle (x0, y0), the sampling points are the four equal divisions of its diagonal; for a circle, the sampling points are the four equal divisions of its diameter parallel to the X and Y axes. A total of five sampling points are used, and their coordinates are stored in the system for the detection process.

[0039] Preferably, a meat viscosity test is performed. After the elasticity test is completed, the elasticity test probe automatically rises and stops, and the transmission device stops for 5 seconds. During this period, the operator can stop the test, restart the test, or change the sequence. If the operator does not change the process, the system automatically drives the stick of the viscosity sensor probe to directly above the sampling center point of the meat sample, controlling the movement process to be consistent with the process in the elasticity test. Then, the probe is controlled to descend and approach the meat sample, with the initial descent speed set at 10-20 mm / s. Preferably, because the extension part of the viscosity test probe is relatively long and the control integrated device needs to adjust the distance between each probe, the descent speed is slowed down and the reserved height is increased. When the predicted test height position is reached, the integrated device controls the rotation of the probe and the extension and retraction of the connector, so that the four probes are matched with the nearest sampling point, placing them directly above the sample point. Preferably, in this experiment, this embodiment selects sample points other than the center point, so that simultaneous measurement can be achieved, reflecting the overall viscosity level of the meat sample. After adjustment, the probe continues to descend at a certain speed, while activating each probe.

[0040] After the probe contacts the sample surface, it continues to descend 3mm, pauses for 0.5 seconds, and then rises at a speed of 2-3mm / s. Information on the probe force and rising height is continuously recorded and stored in the system for evaluating the sample viscosity. Once the recorded probe force is zero, the probe accelerates upwards. After rising a certain height, it drives all four probes to return to their default positions. After returning to their default positions, the connecting rod quickly rises back to the initial position, ready for the next experiment.

[0041] Preferably, the transmission device between the viscous sensor probe and the integrated viscous detector is a hydraulic transmission, which has the advantages of stepless speed regulation, stability and small size.

[0042] Preferably, the flexible strain sensor uses PDMS thin film material as the flexible thin film material and graphene material as the electrode material.

[0043] Preferably, the flexible pressure sensor uses PDMS thin film material as the flexible film material and metal electrodes such as gold, silver, and copper as the electrode material (depending on the requirements).

[0044] Preferably, the pressure sensor probe and the pressure detector guide rail fixing device adopt a hydraulic transmission device, which has the advantages of stepless speed regulation, stability, small size and easy control.

[0045] Preferably, the image processing device employs a neural network computing method developed based on the PyTorch platform; its advantage is that it can quickly identify and analyze image features and classify them.

[0046] Preferably, to achieve the control of changes in the longitudinal and lateral distances of the integrated viscous detector device, as well as the axial movement of the viscous probe and pressure probe, this embodiment employs microcontroller control and electric drive. This embodiment first conducts overall planning, clearly defining the objective as the movement and extension of the drive device, specifying the requirements for the device's extension range, speed, and accuracy, and selecting appropriate microcontrollers and related components based on these requirements, and designing the hardware and software functions accordingly.

[0047] Preferably, this embodiment selects an Arduino microcontroller, which is easy to operate and ensures sufficient input / output pins and computing power to control the movement of the device. A position sensor (such as an encoder or displacement sensor) is selected to monitor the telescopic movement of the device and obtain the required data. For driving, this embodiment selects a stepper motor as the driver, which can perform precise positioning at fixed step angles, suitable for applications requiring high-precision position control. Stepper motors also eliminate the need for position feedback devices, saving cost and complexity. Then, a suitable circuit is designed to connect the microcontroller, sensor, and driver, and appropriate level conversion circuitry and power supply are used to ensure the circuit functions properly.

[0048] Preferably, in order to achieve electric drive control, the power supply voltage and current required by the microcontroller and electronic equipment are first connected to the device through a suitable power supply interface (such as a battery, socket, or power module).

[0049] Design circuits to drive electronic components based on the equipment's requirements. Use appropriate circuit elements (such as transistors, capacitors, resistors, integrated circuits, etc.) to control current and voltage to meet the equipment's operating requirements.

[0050] Preferably, in order to achieve precise positioning, it is necessary to precisely control the motor or other peripherals. In this embodiment, components such as microcontrollers, drivers and sensors are used to achieve fine adjustment and monitoring of power.

[0051] Preferably, to ensure safety, components such as voltage regulators, overload protection, and power filtering are added to the circuit, and appropriate power management circuits are designed and implemented to protect the equipment from power problems such as overcurrent and overvoltage.

[0052] Finally, testing and debugging are conducted. Ensure the electrical connection is established to enable the power drive, allowing the device to move and extend as expected, meeting the required accuracy and speed.

[0053] like Figure 3As shown, the present invention also provides a method for detecting the freshness of mutton, the method comprising:

[0054] Step 201: Use a viscosity detector to detect the viscosity of mutton.

[0055] Step 202: Use a pressure detector to test the elasticity of the mutton.

[0056] Step 203: Use a machine vision sensor to detect the color of the mutton.

[0057] Step 204: Use the signal output terminals of the three sensors to send the detection signal to the host computer for processing.

[0058] Step 205: Display the freshness grade of the mutton on the monitor of the control panel based on the calculation results. The freshness grade and the score in the host computer can be displayed simultaneously.

[0059] In practical applications, the center coordinates of the sample are determined by a flexible sensor array, and the center point iterative search method is used to approximately obtain the center of the meat sample required for the detection experiment. For example... Figure 4 As shown, the specific steps are as follows:

[0060] Step 1: Initialization: Since there is only one center point to be found, the initial state randomly selects a data point as the center point.

[0061] Step 2: Assign data points: For each data point, calculate its distance from the center point and assign it to the set represented by that center point.

[0062] Step 3: Update the centroid: Calculate the average of all data points in the set and use this average as the new centroid.

[0063] Step 4: Repeat steps 2 and 3 until the center point no longer changes or the predetermined number of iterations is reached.

[0064] Step 5: Output the result: The center point P of the set is finally obtained. c (x0,y0).

[0065] Step 6: Based on the obtained center point, other sampling points can be obtained using the following formula:

[0066]

[0067] x max / y max The maximum value of the x / y coordinates in the sample meat coordinate point set;

[0068] x min / y min The minimum value of the x / y coordinates in the sample meat coordinate point set;

[0069] The coordinate values ​​are combined to obtain four sampling points. A total of five sampling points are identified, and their coordinates are stored in the system to prepare for the testing process. Upon receiving the operator's instruction to conduct a meat elasticity test, the transmission device begins operation. First, the probe's movement trajectory is determined:

[0070]

[0071] The electrically driven elastic detection probe's holder moves along the X-axis horizontal bar to directly above the center point of the sample. The distance and direction of movement are determined by the difference in x-coordinate between the holder and the center point. Similarly, the holder moves along the Y-axis horizontal bar to the same position, again determined by the difference in y-coordinate. The X-axis horizontal bar containing the holder moves along the Y-axis vertical bar, aligning the probe with the center point of the sample. Again, the distance and direction of movement are determined by the difference in y-coordinate between the holder and the center point. The probe is then lowered to approach the sample. The initial descent speed is set to 10–20 mm / s, decreasing to 3–5 mm / s upon reaching the predicted experimental height. The pressure sensor on the probe is then activated to detect and transmit pressure information, and the system begins recording pressure data and the probe's descent distance. Predicted experimental height h p It can be calculated according to the following formula:

[0072] h p =h s +h b

[0073] Among them, h s h is the height of the meat at the sampling point. b The fault tolerance height is initialized to 10mm, h b The maximum possible error value for the height of the meat is determined by the fact that the movement is based on the push rod in actual operation. Therefore, it is only necessary to control the probe to move downwards by a distance h. m That's all.

[0074] h m =H-(h) p +h o )

[0075] Where H is the distance between the push rod and the top surface of the operating table, h o To restore altitude.

[0076] After the probe contacts the sampling point on the meat, if the pressure suddenly changes, the probe continues to descend for 2 seconds to form a noticeable indentation on the meat surface. If the pressure remains at a similar level after the sudden change, the test is considered effective. The probe then begins to rise, and the descent height is recorded. Simultaneously, the mechanical vision sensor observes the sampling point and records the time t required for the indentation formed by the elastic probe to return to its initial state. r The raw data is stored in the system for subsequent numerical calculations reflecting elasticity. After rising to the predicted experimental height for this test, the probe is moved directly above the next sampling point. The testing order can be specified separately; the default order is counter-clockwise, starting from the point above the center point (left). The experimental data acquisition process is consistent. The collected data is stored in the system.

[0077] After each measurement, the sensors return to their original positions, ensuring that the initial height of each sensor is not affected by differences in the thickness, stickiness, elasticity, or other properties of the meat, thus facilitating the next measurement.

[0078] In practical applications, once the detector reaches the target area, the microcontroller controls the detector probe to descend, and the sensor on the probe outputs a signal. When the signal reaches a predetermined standard, the detector probe rises. The specific steps are as follows: Figure 5 As shown.

[0079] In practical applications, the shape and area of ​​the meat block to be tested for freshness are determined by the sensor array. Based on the signal given by the sensor array, the flexible strain sensor determines the number of flexible strain sensor probes to be used. The distance between the four probes and the center of the flexible strain sensor is changed by stretching and extending the sensor array to ensure that the viscosity measurement is accurate and effective.

[0080] Specifically, based on the circumscribed circle area of ​​the four probes of the flexible strain sensor at their minimum expandability, the maximum inscribed circle area of ​​the meat block shape obtained by the sensor array is divided into four ranges, from smallest to largest, formed by a first inscribed circle area threshold, a second inscribed circle area threshold, and a third inscribed circle area threshold. The sensor array obtains the shape and size of the meat block, determines the relationship between the maximum inscribed circle area of ​​the shape and the three inscribed circle area thresholds, and obtains a judgment result. If the judgment result indicates that the inscribed circle area of ​​the meat block is less than the first inscribed circle area threshold, the meat block is subjected to four viscosity measurements at different locations using only one of the four probes of the flexible strain sensor. If the judgment result indicates that the inscribed circle area of ​​the meat block is greater than the first inscribed circle area threshold and less than the second inscribed circle area threshold, the maximum inscribed circle area is divided into four ranges, from smallest to largest, formed by a first inscribed circle area threshold, a second inscribed circle area threshold, and a third inscribed circle area threshold. For the circular area threshold, the meat block is subjected to two viscosity measurements at different locations using two of the four probes of the flexible strain sensor. The distance between the four probes of the flexible strain sensor is the average of the maximum and minimum elongation. If the judgment result indicates that the area of ​​the inscribed circle of the meat block is greater than the second inscribed circle area threshold but less than the third inscribed circle area threshold, the meat block is subjected to two viscosity measurements at different locations using three of the four probes of the flexible strain sensor. The distance between the four probes of the flexible strain sensor is the average of the maximum and minimum elongation. If the judgment result indicates that the area of ​​the inscribed circle of the meat block is greater than the third inscribed circle area threshold, the relative distance of the four probes can be changed according to the formula to perform a single viscosity measurement on the meat block, ensuring the effectiveness of the four-point sampling viscosity measurement.

[0081] The flexible strain sensor changes the relative positions of the four probes based on the obtained meat block information and determines the selected points in the four-point sampling method. The mechanical transmission device then moves the flexible strain sensor to the measurement position. The formula for the relative center distance of the four probes of the strain sensor is:

[0082]

[0083] Threshold for the area of ​​the first inscribed circle:

[0084] Second inscribed circle area threshold:

[0085] Threshold for the area of ​​the third inscribed circle: S3 = πl min 2 .

[0086] Among them, l min Let L be the minimum relative center of the four probes of the flexible strain sensor, δ be the perimeter of the meat piece to be measured, δ be the radius of the largest inscribed circle of the meat piece to be measured, S be the area of ​​the meat piece to be measured, S1 be the first area threshold, S2 be the second area threshold, S3 be the third area threshold, and S...内 Let be the area of ​​the largest inscribed circle of the meat piece to be measured. For example... Figure 6 As shown. Exemplary, the minimum stretchability occurs when the elongation is 0 (no stretch).

[0087] like Figure 7 As shown, in practical applications, once the meat piece to be detected is placed on the platform, the machine vision sensor begins to operate. The system controls the sensor to rotate regularly and capture images of the platform. First, the captured image is converted to grayscale. The RGB channel values ​​of each pixel in the color image are linearly combined according to certain weights to obtain the corresponding grayscale value. Since the human eye is most sensitive to green and least sensitive to blue, to more closely approximate the observation effect of the human eye, the weighting of the RGB components is slightly different. The calculation formula is as follows:

[0088] V gray (x, y) = 0.299·V red (x, y) + 0.578·V green (x, y) + 0.114·V blue (x, y)

[0089] After grayscale processing, the gradient of each pixel is calculated. In this embodiment, an edge detection operator is used, and the following convolution kernels are applied to image A in the horizontal and vertical directions respectively:

[0090]

[0091] The gradients of the image in the horizontal and vertical directions are calculated using L. x L y express.

[0092] Then use the intermediate matrix Q

[0093]

[0094] Where p(x,y) is a window function used for weighted summation of pixels, and q(x,y) is an introduced correction function.

[0095] The next step is to calculate the edge corner response function A. p :

[0096] A p = detQ - k(traceQ) 2

[0097] detQ=β1β2

[0098] traceQ=β1+β2

[0099] After calculating the edge corner response values ​​of all pixels, non-maximum suppression is performed to retain local maxima and eliminate redundant corners. Then, a threshold is set based on the corner response values, retaining only pixels with values ​​greater than this threshold. Pixels that pass the threshold filtering are marked as corners. The coordinates of these corners in the image are then obtained.

[0100] The final step in data processing is to calculate the average of the coordinates of all detected corner points to estimate the center point of the target object. This average is then compared with the center coordinates of the image captured by the camera to calculate the offset of the target object in the image.

[0101] Next, based on the offset of the target object in the image, signals are mapped to control the direction and angle of the camera. This allows control of the camera's direction and angle to keep the target object centered, meaning its average coordinates coincide with the image center. After adjusting the angle and height, the focus is adjusted, and the proportion of the meat in the image is adjusted until it reaches approximately 70%.

[0102] After obtaining an accurate image of the meat block, this embodiment will perform meat color detection. First, this embodiment converts the pixel values ​​in the RGB color space to the HSV (hue, saturation, brightness) color space.

[0103] First, the pixel values ​​(R, G, B) in the RGB color space are normalized to the range [0, 1], resulting in new values ​​(r, g, b).

[0104]

[0105] The transformation calculates the (h,s,v) values ​​of each pixel in the HSV space. The transformation calculation formula is as follows:

[0106] v = max{r, g, b}

[0107]

[0108]

[0109] The next step is to determine the freshness of the test meat by comparing the color difference between the test meat and fresh meat.

[0110] First, a representative HSV value needs to be determined for each piece of meat to represent the overall color of that piece. In this embodiment, the center point is used to iteratively find a single color representative value to determine the color of the meat piece. The steps are as follows:

[0111] 1) Data processing: Calculate the average H, S, and V values ​​of the color data points in the meat block, and use the (h,s,v) point as the initial representative value.

[0112] 2) For each data point, calculate its distance from the initial color representative value, and then assign the data point to the cluster containing the nearest color representative value.

[0113] The distance calculation formula is used here:

[0114]

[0115] (h i ,s i ,v i ) represents the HSV value at each point, (h o ,s o ,v o The value of HSV at the center point is )

[0116] 3) Update color representation value: For each cluster, calculate the average value of all data points in the cluster and use this average value as the new color representation value.

[0117] 4) Repeated iteration: Repeat steps 2 and 3 until the color representation value no longer changes on the order of 10^-3.

[0118] 5) Obtain the (h,s,v) color values ​​that represent the overall color of the meat block.

[0119] Secondly, the color of a standard fresh meat chunk is worth noting.

[0120] First, obtain 50 to 100 pieces of freshly slaughtered meat, confirmed by professional inspectors to be of excellent freshness. Perform the aforementioned clustering algorithm on the image of each meat piece to determine a representative color value. Then, using these representative values ​​as the dataset, repeat the aforementioned iterative algorithm. The final color representative values ​​obtained are the color (h, s, v) values ​​of the standard fresh meat piece as defined in this embodiment.

[0121] Finally, a colorimetric system is established.

[0122] A batch of freshly slaughtered meat chunks was obtained and stored under certain conditions. Each chunk was subjected to an experiment at regular intervals (e.g., 12 hours) to obtain (h, s, v) values. The distance data between the (h, s, v) values ​​and those of a standard fresh meat chunk was calculated until the meat chunks were completely rotten. A distance value for each experiment was obtained by averaging the values. The meat was divided into 6 levels, or 6 intervals, ranging from excellent freshness to complete rottenness, with corresponding scores of 5, 4, 3, 2, 1, and 0.

[0123] Due to variations in storage conditions and the condition of the meat itself, a new colorimetric system needs to be established for different specific situations. For example... Figure 8 As shown.

[0124] During the experimental testing, the score M is calculated based on the distance value obtained.

[0125]

[0126] Where d i : To detect the distance between the meat block and the standard fresh meat block, d0: is the distance between the grade 0 meat block and the standard fresh meat block, d5: is the distance between the grade 5 meat block and the standard fresh meat block.

[0127] If a whitish tint is detected, meaning there are many similar RGB values, the weight of the color score in the total score calculation is reduced. The weight W of the color score is calculated as follows:

[0128] W = 0.83·(1-w white )

[0129] Where w red This step detects the proportion of the whitish portion of the overall color. Exemplarily, this step is parallel to the previous step, which is responsible for locating the meat piece; this step detects the color of the meat piece.

[0130] In practical applications, the obtained flesh-colored parameters are RGB three-channel image information, the obtained elastic parameters are the pit recovery time information, and the obtained viscosity parameters are the electrical signal information transmitted by the flexible strain sensor on the viscosity detector probe. The processing of the flesh-colored parameters has been explained in detail above. Figure 7 As shown, the elasticity parameter is processed by comparing it with the index of the standard sample, ranging from excellent freshness to complete spoilage, divided into 6 levels (6 intervals) with corresponding scores of 5, 4, 3, 2, 1, and 0. The viscosity parameter is first stabilized by a voltage follower device to ensure the output voltage of the flexible strain sensor is amplified by an operational amplifier circuit, and then compared with the annual parameter of the standard sample, ranging from excellent freshness to complete spoilage, also divided into 6 levels (6 intervals) with corresponding scores of 5, 4, 3, 2, 1, and 0. Multimodal data fusion is used to input the elasticity, viscosity, and flesh color scores into the config module. The config module mainly uses code to combine multimodal data, allowing them to be input into the subsequent network. The backend network of the config module is the MyModel model, and its specific network structure is as follows. Figure 9As shown, the results are produced through convolution, pooling, linear and nonlinear operations. Specifically, the meat color information is first processed by convolution and pooling to obtain its features. The features are then input into the detect module to analyze its color. If the detect module detects white, it is fed back to the liner layer, so that liner(83) is changed to liner(20). This operation can reduce the weight of meat color in the model when there is a lot of fat on the surface of the mutton. Then, the feature values ​​of viscosity, elasticity and meat color are fused into the same linear function through the cat command. After two convolution operations, a max pooling operation is performed to fuse the data before convolution pooling with the data after convolution pooling in a certain proportion to complete the features. After this step is repeated twice, liner(6) is processed to obtain a list of length six. The 0-5 in the list represent the probability of predicting the freshness of the meat to be of that level. By comparing with the actual freshness of the meat, the data is fed back to the convolution kernel of the convolution step to modify the data in the convolution kernel. After at least 300 iterations, the prediction result reaches stability.

[0131] Specifically, the Conv2d module and maxpool layer are used to extract features. The middle liner layer is used for linearization, which allows image and non-image information to be transformed into tensors of the same dimension for better fusion. The Totensor layer transforms the input information into tensor form. The cat layer combines the linearized tensors along the last dimension. The last liner layer derives the probability of the classification label, and its output is a list of length 6, where each element represents the probability value of the label.

[0132] Convolution formula:

[0133]

[0134] (Where ε is the brightness value of the flesh-colored information at that matrix point, and τ is the value of the convolution kernel at that matrix point)

[0135] In practical applications, the detection process diagram is as follows: Figure 10 As shown.

[0136] In practical applications, information flow diagrams are as follows: Figure 11As shown, input 1 is the power-on button input signal. During the power-on process, this signal enables the power supply to start, outputting to the LED, flexible pressure sensor array, and vision sensor, putting them into working state. Output 1 is the light signal that illuminates the LED. After receiving the power-on signal and input 2 (the pressure signal generated by the placement of the meat), the flexible sensor array transmits the signals to the threshold analysis system and the sampling point positioning system, respectively. The sampling point positioning system transmits the information to the viscosity and elasticity detectors. The threshold analysis system processes the signal and transmits it to the viscosity detector. After these two inputs, the viscosity and elasticity detectors collect the viscosity and elasticity information of the meat and output it as signals to the config module. The config module merges these signals and outputs them, combining them with the information collected by the vision sensor. This data is then input into the MyModel model for analysis. The model outputs a list of length 6, representing the probability information of predicting freshness levels 0, 1, 2, 3, 4, and 5, respectively. This information is processed to obtain the freshness level, which is output as output 2. The freshness detection process is now complete, and the user obtains the freshness level of the meat.

[0137] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0138] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A non-destructive biomimetic method for detecting the freshness of mutton, characterized in that, This invention relates to a non-destructive biomimetic detection device for mutton freshness. The device includes: a flexible sensor array, a transverse transmission rod, a longitudinal transmission rod sleeve, a longitudinal transmission rod tip, a transverse-to-vertical axis converter, a viscous detector guide rail fixing device, a pressure detector guide rail fixing device, a viscous detector integrated device, a viscous probe sleeve, a flexible strain sensor, a pressure sensor probe, a flexible pressure sensor, a machine vision sensor, a stage, a housing, a door, a control panel, and a host computer. The viscous probe sleeve is connected to the viscous detector integrated device via an axial sliding joint. The viscous detector integrated device is threaded to the lower part of the viscous detector guide rail fixing device. The transverse transmission rod passes through the entire viscous detector guide rail fixing device via two bearings. The pressure sensor probe is connected to the lower part of the pressure detector guide rail fixing device via an axial sliding joint. The transverse transmission rod passes through the pressure detector guide rail fixing device via bearings. The system includes: a track fixing device; a transverse and longitudinal axis converter connected to one end of the transverse transmission rod; a longitudinal transmission rod sleeved on the shaft hole below the transverse and longitudinal axis converter; a tip of the longitudinal transmission rod connected to the longitudinal transmission rod sleeve via threads; a tip of the longitudinal transmission rod fixed at a preset position on the housing; a platform connected to the housing using screws; a machine vision sensor fixed at a corresponding position on the housing; a flexible sensor array fixed to the platform by embedding; a door connected to the housing via a door hinge; a control panel mounted on the housing; a flexible strain sensor adhered to the working surface below the adhesive probe sleeve; a flexible pressure sensor adhered to the working surface below the pressure sensor probe; a host computer connected to the flexible pressure sensor, the machine vision sensor, and the flexible strain sensor; and a flexible sensor array used to acquire pressure change signals reflecting changes in pressure values. The method includes: The target meat piece is placed on a flexible sensor array to generate a pressure change signal; Based on the host computer, the coordinates of the target meat block are determined according to the pressure change signal, and the geometric center and shape and area information of the target meat block are determined according to the coordinates of the target meat block. The transverse transmission rod, the longitudinal transmission rod sleeve, and the tip of the longitudinal transmission rod are controlled to move, thereby moving the integrated adhesive detector device to the geometric center of the target meat block. The adhesive probe sleeve is controlled to move downward so that the flexible strain sensor stops when it contacts the surface of the target meat block. After stopping, the adhesive probe sleeve is controlled to move upward. When the viscous probe sleeve moves upward, the deformation rate of the flexible strain sensor increases, which changes the electrical signal transmitted by the flexible strain sensor. When the strain of the flexible strain sensor reaches its maximum value and detaches from the meat surface, the viscous detector guide rail fixing device is controlled to move to the edge of the box. The pressure detector fixing device is moved to the geometric center of the target meat block and the pressure sensor probe is lowered. When the pressure sensor probe contacts the surface of the target meat block, the speed is reduced. When the pressure signal transmitted by the flexible pressure sensor reaches a preset threshold, the pressure sensor probe is raised. During the upward lifting of the pressure sensor probe, the recovery time of the pit of the target meat block and the color of the target meat block are detected by the mechanical vision sensor and the host computer. The viscosity index is graded based on the signal from the flexible strain sensor to obtain a viscosity grading result. The color is graded based on the color of the target meat block to obtain a color grading result. The elasticity index is graded based on the recovery time of the pits in the target meat block to obtain an elasticity grading result. Based on multimodal data fusion, the freshness level of the target meat piece is determined according to the viscosity grading result, the color grading result, and the elasticity grading result. Before controlling the downward movement of the viscous probe sleeve, the following steps are also included: The maximum inscribed circle area of ​​the target meat piece is determined based on the shape and area information. Based on the circumcircle area when the four probes of the flexible strain sensor are at their smallest expandable size, the maximum inscribed circle area is divided to obtain the first inscribed circle area threshold, the second inscribed circle area threshold, and the third inscribed circle area threshold, with values ​​increasing from small to large. Determine the relationship between the maximum inscribed circle area and the first inscribed circle area threshold, the second inscribed circle area threshold, and the third inscribed circle area threshold to obtain the determination result; If the determination result indicates that the maximum inscribed circle area is less than the first inscribed circle area threshold, then one of the four probes of the flexible strain sensor is used to perform four adhesion measurements on the target meat block at different locations; if the determination result indicates that the maximum inscribed circle area is greater than the first inscribed circle area threshold and less than the second inscribed circle area threshold, then two of the four probes of the flexible strain sensor are used to perform two adhesion measurements on the target meat block at different locations; if the determination result indicates that the maximum inscribed circle area is greater than the second inscribed circle area threshold and less than the third inscribed circle area threshold, then three of the four probes of the flexible strain sensor are used to perform two adhesion measurements on the target meat block at different locations; if the determination result indicates that the maximum inscribed circle area is greater than the third inscribed circle area threshold, then one adhesion measurement is performed on the target meat block using all four probes of the flexible strain sensor. The formula for the relative center distance of the four probes of the flexible strain sensor is: The formula for the threshold area of ​​the first inscribed circle is: The formula for the second inscribed circle area threshold is: The formula for the threshold area of ​​the third inscribed circle is: The formula for determining the number of probes is: n = ;in l The distance from each probe to the center This represents the distance between the probe and the center of the flexible strain sensor when it is at its minimum stretchability. To detect the perimeter of the target meat piece, The maximum inscribed circle radius of the target meat piece is determined. The area of ​​the target meat piece is detected. The threshold value for the area of ​​the first inscribed circle. The threshold value for the area of ​​the second inscribed circle. The threshold value for the area of ​​the third inscribed circle. Let n be the area of ​​the largest inscribed circle of the target meat piece, and n be the number of the flexible strain sensor probes.

2. The non-destructive biomimetic detection method for mutton freshness according to claim 1, characterized in that, The geometric center of the flexible sensor array is the same as the geometric center of the stage.

3. The non-destructive biomimetic detection method for mutton freshness according to claim 1, characterized in that, The flexible sensor array includes multiple microsensors, a sensor protective layer, and a first signal output terminal; the microsensors are pressure sensors; the sensor protective layer is made of flexible material; the sensor protective layer is disposed on the shell of the microsensors; the microsensors are used to acquire the pressure change signal; the first signal output terminal is connected to both the microsensors and the host computer.

4. The non-destructive biomimetic detection method for mutton freshness according to claim 1, characterized in that, The flexible strain sensor includes a metal electrode, a flexible base film, and a second signal output terminal; The metal electrode is attached to the flexible base film using 3D printing technology, and the second signal output terminal is welded to both sides of the metal electrode by welding. The second signal output terminal is connected to the host computer via wires.

5. The non-destructive biomimetic detection method for mutton freshness according to claim 1, characterized in that, The enclosure contains built-in LED lights; the LED lights are used to illuminate the interior space of the enclosure.

6. The non-destructive biomimetic detection method for mutton freshness according to claim 1, characterized in that, Based on the host computer, the coordinates of the target meat piece are determined according to the pressure change signal, including: Randomly select a data point from the set of coordinate points of the target meat block as the center point; For each data point, select one data point as the center point, calculate the distance from the other data points to the center point, and assign the data points to the set represented by the nearest center point based on the calculated distances; Calculate the average of the x and y coordinates of all data points in the set represented by each center point, and use the average as the new center point; Repeat the steps "For each data point, select one data point as the center point, calculate the distance of the other data points to the center point, and assign the data points to the set represented by the nearest center point according to the calculated distance" and "Calculate the average of the x and y coordinates of all data points in the set represented by each center point, and use the average as the new center point". After multiple iterations, the iteration stops when the four center points converge to the same coordinate or the specified number of iterations is reached. The center point of each set converges to the true center point of the data point set to obtain the final center point. The coordinates of the other sampling points are calculated based on the final center point; the formula for calculating the coordinates of the other sampling points is as follows: ； ； in, and These represent the maximum values ​​of the x and y coordinates of the coordinate points in the target meat block set, respectively. and These are the minimum x and y coordinates of the coordinate points in the target meat block set, respectively. x i For the x-coordinate values ​​of other sampling points, y i For the ordinate values ​​of other sampling points, x 0 represents the x-coordinate of the final center point. y 0 represents the final ordinate value of the center point; The coordinates of each other sampling point and the final center point are stored to determine the coordinates of the target meat piece.

7. The non-destructive biomimetic detection method for mutton freshness according to claim 1, characterized in that, The method for detecting the color of the target meat piece includes: The RGB channel values ​​of each pixel in a color image captured by a machine vision sensor are linearly combined according to preset weights to obtain the corresponding grayscale value; the formula for calculating the grayscale value is: Finally, the image after grayscale processing is obtained; where, For the pixel's red channel value, For the green channel value of the pixel, For the green channel value of the pixel, The pixel grayscale value; Using an edge detection operator, convolution kernels are applied to the grayscale image in both the horizontal and vertical directions to calculate the gradient of the grayscale image in the horizontal direction. and gradient in the vertical direction The formula for the horizontal convolution kernel is: The formula for the vertical convolution kernel is: ;in, The convolution kernel is in the horizontal direction. The convolution kernel is in the vertical direction. A This is the image after grayscale processing; Based on the gradient in the horizontal direction and gradient in the vertical direction Calculate the intermediate matrix The intermediate matrix The calculation formula is: ;in, It is a window function used for weighted summation of pixels. It is the introduced correction function; According to the intermediate matrix Determine the response function of the edge corner The edge corner response function The calculation formula is: Where det is a matrix Q The determinant, k For matrix Q traces; According to the edge corner response function Calculate the edge corner response values ​​of all pixels and perform non-maximum suppression to retain local maxima in order to eliminate redundant corners. Set a filtering threshold based on the edge corner response values ​​to retain pixels that are greater than the filtering threshold. Mark the pixels that pass the threshold filtering as corners and obtain the coordinates of the corners in the image. The average coordinates of all detected corner points are calculated to estimate the center point of the target object, and compared with the center coordinates of the color image captured by the machine vision sensor to calculate the offset of the detected meat piece in the image. ; ; ; x c To detect the X-axis coordinate of the center of the meat block. y c To detect the Y-axis coordinate of the center of the meat piece, denoted as the deviation distance between the visual center and the center of the detected meat block, where n is the number of corner points; The data type characteristics of the offset are converted into signal characteristics, and the direction and angle of the machine vision sensor are controlled according to the converted signal characteristics to keep the target meat piece in the center position. After the angle and height of the machine vision sensor are adjusted, the focus is adjusted, and the proportion of the target meat piece in the color image is adjusted. Once the preset proportion is reached, the focus is stopped to obtain an accurate image. Convert the pixel values ​​in the RGB color space of the accurate image to the HSV color space; The pixel values ​​in the RGB color space of the accurate image are normalized to obtain new pixel values; the normalization formula is: Where R, G, and B are the pixel values ​​of the points corresponding to the RGB channels in the three channels of the color image, and r, g, and b are the normalized values ​​of the corresponding R, G, and B, respectively. The HSV space conversion calculation is performed based on the new pixel value to obtain the pixel value in HSV space; the formula for the conversion calculation is: ; ; ; h , s and v These are the coordinate values ​​in the corresponding HSV space; Perform iterative calculations of color values; the iterative calculations include: Calculate the H, S, and V values ​​of the pixels in the detected meat block, and use them as follows: The point is used as the coordinates in the HSV space; initially, four coordinates are randomly selected as color representative values. For each pixel, calculate its distance to the color representative value, and assign the data point to the cluster containing the nearest color representative value according to the distance calculation formula; the distance calculation formula is: ; The HSV values ​​at each point, The value of the color's HSV. For each cluster, calculate the average value of all data points in the cluster, and use the average value as the new color representation value; Repeat the steps "For each pixel, calculate its distance to the color data points and assign the data points to the cluster containing the nearest color representative value according to the distance calculation formula" and "For each cluster, calculate the average value of all data points in the cluster and use the average value as the new color representative value". As the number of iterations increases, the four color representative values ​​gradually converge to a single point. Stop when the color representative value stops changing on the order of 10^-3, thus obtaining the overall color situation of the detected meat block. Color value; Obtain 50 to 100 pieces of slaughtered meat that meet a preset freshness requirement. For each meat piece meeting the preset requirement, perform the iterative calculation to determine a color representative value for that meat piece. Then, using this color representative value as the dataset, perform the iterative calculation again. Finally, determine the obtained color representative value as the color of the standard fresh meat piece. value; Obtain freshly slaughtered meat chunks and store them under preset conditions. Each freshly slaughtered meat chunk is tested once every preset time interval. Value, calculated based on the color of a standard fresh meat block. The distance data of the value is recorded until the meat is completely rotten; A distance value for each experiment was obtained by averaging, ranging from excellent freshness to complete spoilage, and divided into 6 levels, or 6 intervals, with corresponding scores of 5, 4, 3, 2, 1, and 0. Calculate the score after obtaining the distance value. The method is as follows: ； in To measure the distance between the meat chunk and a standard fresh meat chunk, The distance between a grade 0 meat block and a standard fresh meat block. The distance between a grade 5 cut of meat and a standard fresh cut of meat; If a whitish tint is detected, meaning there are many similar RGB values, then the weight of the color score in the total score calculation will be reduced. The calculation method is as follows: ； in, This is to detect the proportion of the whitish portion of the color relative to the overall color.

8. The non-destructive biomimetic detection method for mutton freshness according to claim 1, characterized in that, Based on a multimodal data fusion approach, the freshness level of the target meat piece is determined according to the viscosity grading result, the color grading result, and the elasticity grading result, including: The viscosity grading result, the color grading result, and the elasticity grading result are input into the config module to obtain fusion information; The fused information is input into the MyModel model for processing to obtain the freshness level of the target detected meat piece.

Images