A potato chip cutting device, a blade detection method thereof, a medium and a system

Abstract

The application provides a potato slice cutting device and a blade detection method, medium and system thereof, and belongs to the technical field of potato slice cutting devices. The device has an L-shaped groove body. A push block is arranged on one side of the groove body without a side wall. The push block is pushed by multiple air cylinders. A limiting block is arranged on the top of the groove body. A blade holder is arranged on the side wall opposite to the push block. Multiple blades are closely and equidistantly arranged on the blade holder. A blade hole matching the cross section of the blade is opened on the top of the blade holder for taking out or putting in the blade. The sizes of adjacent blades on the blade holder are the same. The blade detection method uses ultrasonic detection technology to realize nondestructive detection of the deformation and wear of a single blade. The core idea is that the vibration characteristics of the blade are extracted by comparing the ultrasonic reflection signals of the blade in the non-working state and the working state, and then the deformation degree and the wear degree of the blade are analyzed from the characteristics.

Description

Technical Field

[0001] This invention belongs to the technical field of potato chip cutting devices, specifically, it relates to a potato chip cutting device and its blade detection method, medium and system. Background Technology

[0002] Potatoes are a widely cultivated and consumed agricultural product, playing an important role in our daily diet. Potatoes typically require pre-processing such as washing and slicing before they can be cooked or processed. Slicing the potatoes is a crucial pre-processing step, and its quality directly affects the results of subsequent processing and the quality of the final product.

[0003] Currently, potato slice cutting is typically accomplished using automated equipment. These potato cutting devices generally consist of a feeding mechanism, a cutting mechanism, and a collecting mechanism. Among these, the cutting mechanism is the core component, affecting cutting quality and production efficiency. Common cutting mechanisms usually employ a blade-type design, utilizing high-speed rotating blades to cut the potatoes. This method offers high cutting efficiency, is suitable for mass production, and is widely used in potato processing settings such as factories and restaurants.

[0004] While automated cutting equipment significantly improves the efficiency of potato processing, several problems exist in practical applications. First, blades wear and deform under prolonged high-intensity operation, leading to decreased cutting quality. Users need to regularly inspect the blades and replace worn ones promptly, increasing maintenance workload. Second, the sharpness of the blades poses safety hazards during inspection and maintenance. Accidental contact with the blades can easily cause injury. Furthermore, inspecting individual blades typically requires removing them from the cutting device, increasing complexity and interrupting production. Summary of the Invention

[0005] In view of this, the present invention provides a potato chip cutting device and its blade detection method, medium and system, which can solve the technical problems of uneven cutting thickness in existing potato chip cutting devices and the fact that the detection of a single blade usually requires removing it from the cutting device, which can easily cause personal injury.

[0006] This invention is implemented as follows:

[0007] A first aspect of the present invention provides a potato chip cutting device, wherein an L-shaped groove is provided, a pushing block is provided on one side of the groove without sidewalls, the pushing block is pushed by a plurality of cylinders, a limiting block is provided on the top of the groove, a blade holder is provided on the sidewall of the groove opposite to the pushing block, a plurality of blades are arranged closely and equidistantly on the blade holder, and a blade feeding hole matching the cross-section of the blades is opened on the top of the blade holder for taking out or putting in the blades, and adjacent blades on the blade holder are the same size.

[0008] A second aspect of the present invention provides a method for detecting the blades of a potato slicer, for detecting the blades on the blade holder of the potato slicer or the blades of a potato slicer having multiple blades, comprising the following steps:

[0009] S10. Obtain the ultrasonic detection signals of the knife holder in the non-working state and the working state of the potato slice cutting device, and record them as the first ultrasonic signal and the second ultrasonic signal, respectively.

[0010] S20. Preprocess the first ultrasonic signal and the second ultrasonic signal;

[0011] S30. Extract the ultrasonic signal reflected by the blade from the preprocessed first ultrasonic signal and second ultrasonic signal, and record them as the first reflection signal and the second reflection signal, respectively.

[0012] S40. After aligning the first reflection signal and the second reflection signal, calculate the difference curve between the first reflection signal and the second reflection signal to reflect whether the blade has deformed or shifted.

[0013] S50. Based on the tool holder displacement, the displacement data in the difference curve is deleted by using a moving average filter to obtain the deformation curve;

[0014] S60. Based on the deformation curve, perform Fourier analysis on the deformation curve to extract the deformation characteristic parameters of each blade, which are denoted as the blade deformation curve.

[0015] S70. Based on the blade deformation curve, the wear degree of the corresponding blade is calculated using a pre-established blade deformation-wear relationship model.

[0016] Further, step S10 specifically includes: setting an ultrasonic transmitting device on the outside of the blade holder of the potato slice cutting device, the incident angle between the ultrasonic transmitting device and the plane where the blade holder is located is 10~30 degrees; setting an ultrasonic probe on the outside of the blade holder, the ultrasonic probe cyclically moving along a direction perpendicular to the blade to receive reflected ultrasonic signals; acquiring the ultrasonic signal of the blade holder when the potato slice cutting device is not in operation, as the first ultrasonic signal; acquiring the ultrasonic signal of the blade holder when the potato slice cutting device is in operation, as the second ultrasonic signal.

[0017] Furthermore, step S20 specifically includes: using a low-pass filter to denoise the first ultrasonic signal and the second ultrasonic signal; and normalizing the denoised first ultrasonic signal and the second ultrasonic signal to normalize the signal amplitude to between 0 and 1.

[0018] Furthermore, step S30 specifically includes: extracting the ultrasonic signal generated by the blade reflection from the preprocessed first ultrasonic signal using an amplitude threshold method, as the first reflection signal; and extracting the ultrasonic signal generated by the blade reflection from the preprocessed second ultrasonic signal using an amplitude threshold method, as the second reflection signal.

[0019] Furthermore, step S40 specifically includes: using a dynamic time warping algorithm to time-align the first reflected signal and the second reflected signal; calculating the amplitude difference between the time-aligned first reflected signal and the second reflected signal at corresponding time points to obtain a difference curve.

[0020] Furthermore, step S50 specifically includes: based on the displacement information of the tool holder, using a moving average filtering method with 5-10 sampling points, deleting the displacement-related part of the difference curve to obtain the deformation curve of the blade.

[0021] Step S60 specifically includes: performing Fourier analysis on the blade deformation curve to extract the deformation characteristic parameters of each blade, including at least the main frequency components and amplitude of the deformation, denoted as the blade deformation curve.

[0022] Furthermore, the blade deformation-wear relationship model is a linear regression model or a support vector machine regression model; the training set of the model includes the deformation characteristic parameters of the blade and the actual wear data under different usage durations and conditions; the model training adopts the least squares method or cross-validation method to ensure the generalization performance of the model.

[0023] A third aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores program instructions, which, when executed, are used to perform the aforementioned potato slice cutting device and its blade detection method.

[0024] A fourth aspect of the present invention provides a blade detection system for a potato chip cutting apparatus, wherein the aforementioned computer-readable storage medium is included.

[0025] Compared with the prior art, the beneficial effect of the potato slice cutting device provided by the present invention is that by setting the blades closely and evenly, the technical problem of uneven cutting thickness in the existing potato slice cutting devices is solved.

[0026] The beneficial effects of the present invention in providing a blade detection method, medium, and system for a potato chip cutting device are:

[0027] 1. Fully automated, non-destructive testing without manual blade disassembly. Traditional testing methods typically require removing the blade from the cutting device for individual inspection, which increases complexity and disrupts production. This invention utilizes ultrasonic detection technology to monitor the blade in real-time without interfering with its normal operation, significantly improving testing efficiency.

[0028] 2. It can accurately assess the wear degree of a single blade. Existing indirect monitoring methods, such as inferring blade wear through changes in power consumption or temperature, can only provide rough estimates and are insufficient for quantitatively assessing the actual wear degree of a single blade. This invention utilizes Fourier analysis and machine learning techniques to extract detailed wear information from the vibration characteristics of the blade, providing a more accurate basis for timely replacement.

[0029] 3. Reduces the safety hazards of manual inspection. Traditional manual inspection methods pose certain safety risks, as operators can easily come into contact with sharp blades and be injured. This invention achieves fully automated, contactless inspection, significantly reducing operational risks and creating a safer working environment for workers.

[0030] This invention solves the technical problem that during the operation and maintenance of potato chip cutting devices, the inspection of individual blades usually requires removing them from the cutting device, which can easily cause personal injury. Attached Figure Description

[0031] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention 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.

[0032] Figure 1 This is a schematic diagram of the structure of a potato slice cutting device provided by the present invention;

[0033] Figure 2 A flowchart of the method provided by the present invention;

[0034] The attached diagram lists the components represented by each number as follows:

[0035] 01. Groove; 02. Limiting block; 03. Pushing block; 04. Tool holder; 31. Cylinder; 41. Blade; 42. Tool feed hole. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0037] like Figure 1 The diagram shown is a structural schematic of a potato chip cutting device provided in the first aspect of the present invention. The device has an L-shaped groove 01. A pushing block 03 is provided on the side of the groove 01 without sidewalls. The pushing block 03 is pushed by multiple cylinders 31. A limiting block 02 is provided on the top of the groove 01. A blade holder 04 is provided on the sidewall of the groove 01 opposite to the pushing block 03. Multiple blades 41 are arranged closely and equidistantly on the blade holder 04. A blade feeding hole 42 matching the cross-section of the blade 41 is opened on the top of the blade holder 04 for taking out or putting in the blade 41. The adjacent blades on the blade holder 04 have the same size.

[0038] like Figure 2 The diagram shown is a flowchart of a method for detecting the blades of a potato chip cutting device, provided by a second aspect of the present invention. This method is used to detect the blades on the blade holder of a potato chip cutting device or the blades of a potato chip cutting device having multiple blades, and includes the following steps:

[0039] S10. Obtain the ultrasonic detection signals of the knife holder in the non-working state and the working state of the potato slice cutting device, and record them as the first ultrasonic signal and the second ultrasonic signal, respectively.

[0040] S20. Preprocess the first ultrasonic signal and the second ultrasonic signal;

[0041] S30. Extract the ultrasonic signal reflected by the blade from the preprocessed first ultrasonic signal and second ultrasonic signal, and record them as the first reflection signal and the second reflection signal, respectively.

[0042] S40. After aligning the first reflection signal and the second reflection signal, calculate the difference curve between the first reflection signal and the second reflection signal to reflect whether the blade has deformed or shifted.

[0043] S50. Based on the tool holder displacement, the displacement data in the difference curve is deleted by using a moving average filter to obtain the deformation curve;

[0044] S60. Based on the deformation curve, perform Fourier analysis on the deformation curve to extract the deformation characteristic parameters of each blade, which are denoted as the blade deformation curve.

[0045] S70. Based on the blade deformation curve, the wear degree of the corresponding blade is calculated using a pre-established blade deformation-wear relationship model.

[0046] Chinese utility model patent application number CN202221507650.1 discloses a potato slicer for food processing, which has a blade holder similar to the potato slicer cutting device provided by this invention. The method provided by this invention can also be used to detect the blades on the blade holder of the potato slicer for food processing in this utility model.

[0047] The specific implementation methods of the above steps are described in detail below:

[0048] The specific implementation method of step S10 is as follows:

[0049] First, ultrasonic sensors are used to acquire ultrasonic signals from the tool holder in both its non-working and working states. Specifically, an ultrasonic transmitter is positioned on the outside of the tool holder, with an incident angle of 10-30 degrees to the tool holder plane. Simultaneously, an ultrasonic probe is positioned on the outside of the tool holder, which moves cyclically along a direction perpendicular to the blades to receive reflected ultrasonic signals. This allows the acquisition of ultrasonic signals from the tool holder in its non-working and working states, denoted as the first ultrasonic signal and the second ultrasonic signal, respectively. The ultrasonic transmitter operates at a frequency of 40kHz, with an incident angle of 20 degrees to the tool holder plane; the ultrasonic receiving probe moves cyclically along a direction perpendicular to the blades to receive reflected ultrasonic signals. During data acquisition, the ultrasonic detection parameters were first optimized and adjusted. The distance between the ultrasonic probe and the tool holder was set to 50mm, the probe movement speed to 20mm / s, and the scanning time for each blade to be approximately 1.5s.

[0050] The specific implementation method of step S20 is as follows:

[0051] The acquired first and second ultrasonic signals are preprocessed. First, noise reduction is performed, for example using a low-pass filter, such as a 4th-order Butterworth low-pass filter with a cutoff frequency of 500Hz, to remove high-frequency noise. Then, the signals are normalized, reducing the signal amplitude to between 0 and 1. This eliminates amplitude differences caused by factors such as sensor sensitivity and amplifier gain, preparing the signal for subsequent feature extraction.

[0052] The specific implementation method of step S30 is as follows:

[0053] The ultrasonic signal generated by the blade reflection is extracted from the preprocessed first and second ultrasonic signals. Due to the good reflectivity of the blade surface, the characteristic signal generated by the blade reflection is clearly visible in the ultrasonic signal. These characteristic signals can be extracted using the amplitude thresholding method or the peak detection method, and are denoted as the first reflection signal and the second reflection signal, respectively.

[0054] The specific implementation method of step S40 is as follows:

[0055] After aligning the first and second reflected signals, the difference curve between the two signals is calculated. Here, the Dynamic Time Warping (DTW) algorithm can be used to align the two signals in time, and then the amplitude difference at corresponding time points is calculated to obtain the difference curve. This difference curve reflects the changes in the blade in its non-working and working states, and can be used to determine whether the blade has deformed or shifted.

[0056] The specific implementation method of step S50 is as follows:

[0057] Considering that the tool holder may shift during operation, it is necessary to remove the changes caused by this displacement from the difference curve. A moving average filtering method can be used to remove the displacement-related portion of the difference curve based on the tool holder's displacement information, resulting in a pure tool deformation curve. Here, the moving average window size can be set to 5-10 sampling points.

[0058] The specific implementation method of step S60 is as follows:

[0059] Fourier analysis was performed on the obtained blade deformation curves to extract the deformation characteristic parameters of each blade. Fourier analysis can transform the time-domain signal into the frequency domain, obtaining the amplitude and phase information of each frequency component. The spectral characteristics of these signals can well reflect the deformation of the blade, such as the main frequency components and amplitudes of the blade deformation, and are denoted as the blade deformation curve.

[0060] The specific implementation method of step S70 is as follows:

[0061] Based on the obtained blade deformation curve, the wear degree of the corresponding blade is calculated using a pre-established blade deformation-wear relationship model. This model can be trained using a large amount of experimental data, with the input being the blade's deformation characteristic parameters and the output being the blade's wear degree. The model's training set can include blade deformation characteristics and actual wear degree data under different usage durations and conditions. Commonly used model structures include linear regression models and support vector machine regression models. During training, methods such as least squares and cross-validation can be used to optimize the model parameters to ensure the model's generalization performance.

[0062] In summary, this ultrasonic-based method for monitoring the condition of potato chip cutting blades enables non-destructive and automated detection of individual blade deformation and wear, providing a basis for timely replacement of worn blades, avoiding the tedious process of manual disassembly and inspection, and improving detection efficiency and safety.

[0063] To better understand this method, the specific implementation methods of steps S30-S70 are described in more detail below, using specific formulas:

[0064] Step S30: Extract the blade reflection signal from the preprocessed first and second ultrasonic signals.

[0065] In step S20, the acquired first ultrasonic signal Second ultrasonic signal Denoising and normalization were performed to obtain the preprocessed signal. and Now we need to extract the ultrasonic signal generated by the blade reflection from these two preprocessed signals.

[0066] Because the blade surface has excellent acoustic wave reflection characteristics, a significant blade reflection signal will appear in the original ultrasonic signal. This reflection signal can be extracted using the amplitude thresholding method. Specifically, for the first ultrasonic signal... Define the amplitude threshold as The ultrasonic signal reflected by the blade It can be represented as:

[0067]

[0068] Similarly, for the second ultrasonic signal Define the amplitude threshold as The ultrasonic signal reflected by the blade It can be represented as:

[0069]

[0070] The amplitude threshold here and Adjustments can be made based on the actual situation; for example, one could take... and for and The amplitude average is 50%. Using this method, the ultrasonic signal generated by the blade reflection was successfully extracted from the original ultrasonic signal. and This laid the foundation for subsequent analysis.

[0071] In practice, operators found that when the ultrasonic signal amplitude exceeded 45% of the original signal average, it was essentially the characteristic signal generated by the blade reflection. Therefore, the following method was used to extract these reflected signals:

[0072] for Extracting the satisfying Part of it, as the first reflected signal ;

[0073] for Extracting the satisfying as the second reflected signal .

[0074] Step S40: Calculate the difference curve between the first reflected signal and the second reflected signal

[0075] In step S30, the ultrasonic signal generated by the blade reflection has been extracted from the preprocessed ultrasonic signal and . Now it is necessary to compare the differences between these two signals to detect whether the blade has deformed or displaced

[0076] Since and may have different time axes, it is necessary to align the two signals in time first. The DTW algorithm can be used to achieve this. The DTW algorithm can find the optimal alignment path between two time series to minimize the distance between the two series

[0077] Assume and have time lengths of and respectively. Define the distance matrix as:

[0078]

[0079] where , . Then the optimal alignment path can be solved by dynamic programming, making the sum of distances along the path the smallest

[0080] With the optimal alignment path , the difference curve between the first reflected signal and the second reflected signal can be calculated :

[0081]

[0082] This difference curve reflects the changes of the blade in the non-working state and the working state, and can be used to judge whether the blade has deformed or displaced

[0083] Step S50: Delete the displacement data in the difference curve according to the tool rest displacement information

[0084] In step S40, the difference curve of the blade reflection signal has been calculated . However, due to the possible displacement of the tool rest during the working process, resulting in This includes changes caused by displacement, not just pure cutting tool deformation information. To eliminate the influence of this displacement, it is necessary to adjust the tool holder's displacement information accordingly. Further processing is required.

[0085] Assuming the acquired tool post displacement information can be represented as Therefore, a moving average filtering method can be used to delete... From the displacement-related portion, we obtain the pure curve of the blade deformation. :

[0086]

[0087] in Using the half-length of the moving average window, 5-10 sampling points can be taken. This yields the blade deformation curve that excludes the influence of tool holder displacement. This laid the foundation for subsequent Fourier analysis.

[0088] Step S60: Perform Fourier analysis on the blade deformation curve.

[0089] In step S50, the insert deformation curve excluding the influence of tool holder displacement is obtained. Now we need to perform Fourier analysis on this deformation curve to extract the deformation characteristic parameters of each blade.

[0090] Fourier analysis can transform time-domain signals By converting to the frequency domain, the amplitude and phase information of each frequency component can be obtained. This conversion can be achieved using the Fast Fourier Transform (FFT) algorithm.

[0091]

[0092] in The sampling interval is... The number of sampling points. The frequency is given. Through FFT transformation, we can obtain... Representation in the frequency domain This includes the spectral characteristics of blade deformation.

[0093] To better describe the deformation of the blade, it is possible to extract... Center front The amplitude and phase of each frequency component are used as characteristic parameters of blade deformation, denoted as... .here It can be set according to the actual situation, for example, taking .

[0094] By following the steps described above, the blade deformation curve can be successfully obtained. The spectral characteristic parameters of blade deformation were extracted. This laid the foundation for subsequent wear estimation.

[0095] Step S70: Estimate blade wear based on blade deformation characteristic parameters

[0096] In step S60, the deformation feature parameters of each blade were extracted. Now we need to use these characteristic parameters to estimate the wear degree of the corresponding blade through a pre-established blade deformation-wear relationship model.

[0097] Two common machine learning models can be used here: linear regression and support vector machine regression.

[0098] 1. Linear regression model:

[0099] Assuming blade deformation characteristic parameters With wear There is a linear relationship between them, that is ,in For the weight vector, This is the bias term. The model parameters can be adjusted using the least squares method. and Make an estimate:

[0100]

[0101] in The number of training samples, The training set is used, for example, N=200, meaning 200 training samples are used. By solving the above optimization problem, the parameters of the linear regression model can be obtained, and then the new blade deformation characteristics can be predicted. The corresponding degree of wear.

[0102] 2. Support Vector Machine Regression Model:

[0103] Besides linear regression, support vector machine regression (SVR) models can also be used to fit the blade deformation-wear relationship. SVR models can learn a non-linear mapping relationship, as shown in the formula:

[0104]

[0105] in For kernel function, These are Lagrange multipliers. A radial basis function (RBF) kernel can be used. And optimize kernel function parameters using grid search. and penalty parameters This minimizes the prediction error of the model on the training set. In practice, the kernel function parameters are... Set to 0.1, slack variable Set it to 0.1.

[0106] Using the two regression models described above, the deformation characteristic parameters of the blade can be determined. This predicts the corresponding wear level, providing a basis for timely replacement of worn blades and improving detection efficiency and safety.

[0107] A third aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores program instructions, which, when executed, are used to perform the aforementioned potato slice cutting device and its blade detection method.

[0108] A fourth aspect of the present invention provides a blade detection system for a potato chip cutting apparatus, wherein the aforementioned computer-readable storage medium is included.

[0109] Specifically, the principle of this invention is to use ultrasonic detection technology to achieve non-destructive testing of the deformation and wear of individual blades. The core idea is to extract the vibration characteristics of the blade by comparing the ultrasonic reflection signals of the blade in its non-working and working states; then, using Fourier analysis and machine learning techniques, the degree of deformation and wear of the blade is analyzed from these characteristics, providing a basis for timely replacement.

[0110] Specifically, the method includes the following steps:

[0111] 1. Ultrasonic signal acquisition

[0112] An ultrasonic transmitter and detector are installed on the outside of the tool holder to acquire ultrasonic reflection signals of the blade in both its non-working and working states. Because the blade surface has excellent acoustic reflectivity, these reflection signals can accurately reflect the blade's vibration characteristics.

[0113] 2. Blade reflection signal extraction

[0114] The acquired raw ultrasound signal is preprocessed, including noise reduction and normalization. Then, the characteristic signal generated by the blade reflection is extracted from the preprocessed signal. This can be achieved using an amplitude thresholding method, where the portion of the ultrasound signal with an amplitude exceeding a certain threshold is identified as the blade reflection signal.

[0115] 3. Blade Deformation Curve Analysis

[0116] The blade reflection signals in the non-working and working states are aligned, and the difference curve between the two is calculated. This difference curve reflects the deformation of the blade during operation. Considering that the tool holder itself may also displace, the difference curve needs to be filtered using the tool holder displacement information to remove the component caused by displacement, thus obtaining the pure blade deformation curve.

[0117] 4. Blade Deformation Feature Extraction

[0118] Fourier analysis was performed on the obtained blade deformation curve to extract spectral characteristic parameters describing the blade deformation, such as the main frequency components and amplitude. These parameters can reflect the degree of blade deformation well.

[0119] 5. Blade wear estimation

[0120] Finally, using a pre-established blade deformation-wear relationship model, the wear degree of the corresponding blade is calculated based on the extracted deformation feature parameters. This relationship model can be trained using a large amount of experimental data, and the mapping relationship between blade deformation features and wear degree is established using methods such as linear regression or support vector machine regression.

[0121] By following the steps described above, non-destructive testing and quantitative assessment of the deformation and wear of individual blades can be achieved. This provides a basis for timely replacement of worn blades, greatly improving the service life and safety of potato cutting equipment.

[0122] The key to this method lies in utilizing ultrasonic detection technology to non-destructively acquire the vibration characteristics of the blade, while Fourier analysis and machine learning techniques can effectively extract and analyze these characteristics to obtain the precise wear state of the blade. Compared to some existing indirect monitoring methods, this invention can more accurately assess the actual wear degree of a single blade, providing more valuable information support for maintenance. Furthermore, this method eliminates the need for manual blade disassembly, significantly improving detection efficiency, reducing safety hazards, and better meeting the actual needs of the potato processing industry.

[0123] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for detecting the blades of a potato chip cutting device, used to detect the blades of a potato chip cutting device having multiple blades, characterized in that, Includes the following steps: S10. Obtain the ultrasonic detection signals of the knife holder in the non-working state and the working state of the potato slice cutting device, and record them as the first ultrasonic signal and the second ultrasonic signal, respectively. S20. Preprocess the first ultrasonic signal and the second ultrasonic signal; S30. Extract the ultrasonic signal reflected by the blade from the preprocessed first ultrasonic signal and second ultrasonic signal, and record them as the first reflection signal and the second reflection signal, respectively. S40. After aligning the first reflection signal and the second reflection signal, calculate the difference curve between the first reflection signal and the second reflection signal to reflect whether the blade has deformed or shifted. S50. Based on the tool holder displacement, the displacement data in the difference curve is deleted by using a moving average filter to obtain the deformation curve; S60. Based on the deformation curve, perform Fourier analysis on the deformation curve to extract the deformation characteristic parameters of each blade, which are denoted as the blade deformation curve. S70. Based on the blade deformation curve, the wear degree of the corresponding blade is calculated using a pre-established blade deformation-wear relationship model.

2. The method for detecting the blade of a potato slice cutting device according to claim 1, characterized in that, Step S10 specifically includes: setting an ultrasonic transmitter outside the blade holder of the potato slice cutting device, the incident angle between the ultrasonic transmitter and the plane where the blade holder is located is 10~30 degrees; setting an ultrasonic probe outside the blade holder, the ultrasonic probe cyclically moving along a direction perpendicular to the blade to receive reflected ultrasonic signals; acquiring the ultrasonic signal of the blade holder when the potato slice cutting device is not in operation, as the first ultrasonic signal; acquiring the ultrasonic signal of the blade holder when the potato slice cutting device is in operation, as the second ultrasonic signal.

3. The method for detecting the blade of a potato slice cutting device according to claim 2, characterized in that, Step S20 specifically includes: using a low-pass filter to denoise the first ultrasonic signal and the second ultrasonic signal; and normalizing the denoised first ultrasonic signal and the second ultrasonic signal to normalize the signal amplitude to between 0 and 1.

4. The blade detection method of a potato slice cutting device according to claim 3, characterized in that, Step S30 specifically includes: extracting the ultrasonic signal generated by the blade reflection from the preprocessed first ultrasonic signal using the amplitude threshold method, as the first reflection signal; and extracting the ultrasonic signal generated by the blade reflection from the preprocessed second ultrasonic signal using the amplitude threshold method, as the second reflection signal.

5. The method for detecting the blade of a potato slice cutting device according to claim 4, characterized in that, Step S40 specifically includes: using a dynamic time warping algorithm to time-align the first reflected signal and the second reflected signal; calculating the amplitude difference between the time-aligned first reflected signal and the second reflected signal at corresponding time points to obtain a difference curve.

6. The method for detecting the blade of a potato slice cutting device according to claim 5, characterized in that, Step S50 specifically includes: based on the displacement information of the tool holder, using a moving average filtering method with 5-10 sampling points, deleting the displacement-related part of the difference curve to obtain the deformation curve of the blade. Step S60 specifically includes: performing Fourier analysis on the blade deformation curve to extract the deformation characteristic parameters of each blade, including at least the main frequency components and amplitude of the deformation, denoted as the blade deformation curve.

7. The method for detecting the blade of a potato slice cutting device according to claim 6, characterized in that, The blade deformation-wear relationship model is a linear regression model or a support vector machine regression model; the training set of the model includes the deformation characteristic parameters of the blade and the actual wear data under different usage durations and conditions; the model training adopts the least squares method or cross-validation method to ensure the generalization performance of the model.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores program instructions, which, when executed, perform a blade detection method for a potato slice cutting device according to any one of claims 1-7.

9. A blade detection system for a potato chip cutting device, characterized in that, It includes the computer-readable storage medium of claim 8.

Images