Aluminum-foil paper defect detection method based on wavelet analysis and Fourier transform

Abstract

The invention discloses an aluminum-foil paper defect detection method based on wavelet analysis and Fourier transform, and the main design conception of the method is as follows: an eddy current type sensor is adopted to collect more detailed defect information of aluminum-foil paper inside the cigarette case through the surface of the cigarette case; wavelet analysis and fast Fourier transform are introduced into detection signal analysis of the cigarette case aluminum-foil paper, the wavelet analysis is used for removing noise existing in the signals, then the specific frequency and amplitude of the signals are analyzed through the fast Fourier transform, existence of defects of the cigarette case aluminum-foil paper is detected, features are extracted from finer frequency domain and time domain information, and the detection accuracy is improved. And the method is particularly suitable for detecting complicated surfaces such as aluminum-foil paper and various types of defects. The influence of illumination and light rays is avoided, and detection is performed from the perspective of complete aluminum-foil paper, so that the efficiency, comprehensiveness and accuracy of cigarette case aluminum-foil paper defect detection are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of cigarette manufacturing, and in particular, to a method for detecting defects in aluminum foil paper based on wavelet analysis and Fourier transform. Background Art

[0002] During the cigarette production process, the aluminum foil paper of the cigarette case, as an inner lining material, plays an important role in preventing the loss of cigarette aroma, moisture loss, and mildew. However, due to the compact design of the packaging box, the aluminum foil paper is prone to defects such as missing tear tongues, offset, and breakage during cutting and folding at multiple workstations. These defects directly affect the appearance and quality of cigarettes, and further affect the market competitiveness of products and consumers' purchasing decisions. Therefore, ensuring the integrity of the aluminum foil paper is the key to maintaining the quality of cigarettes, and developing an efficient and accurate method for detecting defects in the aluminum foil paper of cigarette cases has important practical significance.

[0003] Currently, the methods for detecting defects in the aluminum foil paper of cigarette cases include visual inspection, laser scanning inspection, etc. Among them, visual inspection mainly uses a high-resolution camera and image processing technology to scan the surface of the aluminum foil paper, and identify and classify defects in real time, which has high accuracy. However, it depends on the images collected by the camera, and low light or strong light reflection may affect the identification of defects, resulting in a decrease in the detection accuracy of defects; while laser scanning inspection can identify minute defects on the surface of the aluminum foil, and has the advantages of high precision and high sensitivity, but it is mainly used for detecting surface defects, and may not be able to effectively identify defects inside the aluminum foil paper. In addition, the equipment costs of these detection methods are relatively high, the maintenance difficulty is relatively large, and there are certain limitations, and they cannot well complete the task of detecting defects in the aluminum foil paper of cigarette cases. Summary of the Invention

[0004] In view of the above, the present invention aims to provide a method for detecting defects in aluminum foil paper based on wavelet analysis and Fourier transform to solve the aforementioned technical problems.

[0005] The technical solution adopted by the present invention is as follows:

[0006] The present invention provides a method for detecting defects in aluminum foil paper based on wavelet analysis and Fourier transform, which includes:

[0007] Using an eddy current sensor to collect the aluminum foil paper measurement signal through the surface of the cigarette case;

[0008] Using wavelet analysis to remove the noise in the aluminum foil paper measurement signal to obtain an aluminum foil paper measurement signal containing only useful information;

[0009] Performing spectral analysis on the denoised aluminum foil paper measurement signal by using fast Fourier transform to obtain a corresponding spectrogram;

[0010] According to the spectrogram, determine the target frequency at which the maximum amplitude of the denoised signal of the cigarette pack aluminum foil appears, and determine whether there are defects in the cigarette pack aluminum foil based on the target frequency.

[0011] In at least one possible implementation, the acquisition of the aluminum foil measurement signal includes:

[0012] After setting a predetermined distance between the eddy current sensor and the surface of the cigarette pack, the eddy current sensor probe detects the cigarette pack to be measured placed on a rotating disk.

[0013] The single and double sheet identifier receives the aluminum foil signal sensed by the eddy current sensor probe to obtain the original waveform, and the original waveform includes a normal complete waveform and a measured waveform.

[0014] In at least one possible implementation, the acquisition of the aluminum foil measurement signal further includes: controlling the disk to rotate at different preset speeds through a speed control motor to simulate the movement state of the cigarette packs on an actual production line.

[0015] In at least one possible implementation, the determination of whether there are defects in the cigarette pack aluminum foil based on the target frequency includes: based on the aluminum foil measurement signals at different rotation speeds, respectively comparing the first maximum amplitude at the target frequency in the spectrogram of the denoised signal with the second maximum amplitude at the target frequency in the spectrogram of the complete signal; if the first maximum amplitude is less than the second maximum amplitude, it is determined that there are defects in the aluminum foil of the cigarette pack; wherein, the target frequency is the frequency in the frequency band where the aluminum foil defect is located obtained in advance.

[0016] In at least one possible implementation, after selecting the wavelet basis function and setting the number of decomposition levels, the aluminum foil measurement signal is decomposed into detail coefficients and approximation coefficients of different frequency components.

[0017] As the number of decomposition levels increases, the detail coefficients are gradually decomposed into high-frequency information of smaller scales.

[0018] Based on the high-frequency information of smaller scales, obtain the changes and noises in the aluminum foil measurement signal.

[0019] Set the noises in the detail coefficients that are not correlated and have small absolute values to zero and perform signal reconstruction processing to obtain the denoised aluminum foil measurement signal.

[0020] In at least one possible implementation, the obtaining of the corresponding spectrogram includes: using a high-pass filter to remove the part with a frequency lower than 1 Hz and retain the remaining frequency part.

[0021] Compared with the prior art, the main design concept of the present invention lies in using an eddy current sensor to collect more detailed defect information of the internal aluminum foil through the surface of the cigarette case, and extracting features from more detailed frequency domain and time domain information through wavelet analysis and fast Fourier transform, which is especially applicable to the detection of complex surfaces and various types of defects of aluminum foil. During specific implementation, after denoising the wavelet basis function and decomposition level of the measured signal of the cigarette case aluminum foil, spectrum analysis is performed on the denoised signal through fast Fourier transform, and then compared with the complete signal of normal aluminum foil to determine whether defects exist. The present invention avoids the influence of light and illumination, and conducts detection from the perspective of complete aluminum foil, thereby significantly improving the efficiency, comprehensiveness and accuracy of defect detection of cigarette case aluminum foil. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] To make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described below in conjunction with the accompanying drawings, where:

[0023] Figure 1 is a schematic diagram of the aluminum foil defect detection method based on wavelet analysis and Fourier transform provided by an embodiment of the present invention;

[0024] Figure 2 is a waveform diagram of six aluminum foil measurement signals at three different speeds measured through a test bench provided by an embodiment of the present invention;

[0025] Figure 3 is a wavelet decomposition result diagram of one of the detection signals in the Matlab wavelet toolbox provided by an embodiment of the present invention;

[0026] Figure 4 is a waveform diagram of the denoised signals after wavelet denoising of six signals provided by an embodiment of the present invention;

[0027] Figure 5 is a filtered frequency spectrum diagram generated by fast Fourier transform of six signals provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0028] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions from beginning to end. The embodiments described below by referring to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be construed as a limitation to the present invention.

[0029] An embodiment of the present invention proposes a method for detecting aluminum foil defects based on wavelet analysis and Fourier transform. Specifically, as Figure 1 shown, which includes:

[0030] Step S1: Use an eddy current sensor to collect the aluminum foil measurement signal through the surface of the cigarette case;

[0031] In the embodiment of the present invention, the acquisition means of the cigarette case aluminum foil paper signal adopts eddy current type. Specifically, the eddy current sensor forms an interaction with the measured object through the probe to generate an electromagnetic field, so as to sense the change of the object. Basic process: Keep a preset distance between the eddy current sensor and the surface of the cigarette case. When the alternating magnetic field generated by the probe of the eddy current sensor approaches the aluminum foil paper, the free electrons therein will be affected by the alternating magnetic field and generate eddy currents (i.e., electric eddy currents). The signal changes caused by the eddy currents are received by the device receiver and signal processing circuit connected by electricity, and are processed such as amplification, filtering, and conversion through the electronic circuit in the device, and finally the measurement signal of the aluminum foil paper is output.

[0032] To expand, a test bench for collecting aluminum foil paper signals can be built in advance, which mainly consists of a power supply, a 6YT01 speed control motor for controlling the fixed disc, an Arduino UNO development board, a single and double sheet identifier, and an eddy current sensor probe. Among them, the single and double sheet identifier HJG.SP-812 receives the aluminum foil paper signal sensed by the eddy current sensor probe. The settings of this instrument are as follows: the rated voltage of the instrument is 24V, the analog output voltage is 0~10V, the host response output time is 5ms, the sampling frequency is 2000Hz, and the signal length is 1000. Then, the rotation of the fixed disc is adopted to simulate the movement of the cigarette case on the actual production line. The diameter of the disc can be selected as 500mm, and the reduction ratio of the 6YT01 speed control motor used is 3. Figure 2 Shown are the original waveforms of three complete signals and three defect signals collected at three different speeds.

[0033] Step S2: Use wavelet analysis to remove the noise in the aluminum foil measurement signal to obtain the aluminum foil measurement signal containing only useful information;

[0034] First of all, for the basic process of wavelet noise removal: Select a suitable wavelet basis function and decomposition level, perform multi-scale decomposition on the measured aluminum foil measurement signal, and reconstruct and denoise the signal according to the decomposition result. Specifically, in the first layer, the aluminum foil measurement signal is decomposed into a low-frequency part and a high-frequency part. In the second layer, the low-frequency part obtained in the first layer is decomposed into a low-frequency part and a high-frequency part again, and the subsequent decomposition is carried out in the same way. Subsequently, by setting the irrelevant parts in the high-frequency signal obtained in the above process to zero, the noise signal is filtered out and the signal is reconstructed to obtain the denoised signal.

[0035] In the present invention, wavelet analysis is used to detect defects in cigarette pack aluminum foil paper. First, a suitable wavelet basis function and decomposition level need to be selected to perform wavelet analysis on the above waveform data obtained from the simulation experiment. Here, the Daubechies wavelet is preferably used, which has good time-frequency localization performance. Compared with other wavelet bases, it can capture the detailed changes of the signal more accurately. Secondly, it has good decomposition ability for the details and high-frequency components of the signal and is suitable for most defect detection applications.

[0036] Continuing from the previous text, for the aluminum foil paper measurement signal, one-dimensional discrete wavelet analysis is selected. After importing the aluminum foil paper measurement signal, the db4 wavelet is selected as the wavelet basis function, and the decomposition level is set to 5 layers. The signal is decomposed into detail coefficients and approximation coefficients of different frequency components. As the decomposition level increases, the detail coefficients are gradually decomposed into high-frequency information of smaller scales, and more subtle changes and noises in the aluminum foil paper measurement signal are captured based on the high-frequency information of smaller scales. Subsequently, the high-frequency noise in the detail coefficients with little correlation and relatively small absolute values (which can be judged based on a preset absolute value standard and threshold) is set to zero and the signal is reconstructed to obtain the aluminum foil paper measurement signal with noise removed.

[0037] To expand, in actual operation, the wavelet toolbox of tools such as Matlab can be used. Select the db4 wavelet in it, with a decomposition level of 5 layers, and perform wavelet analysis on the measurement signal. Select the denoising process of one of the six signals in the previous example, as Figure 3 shown. s is the original waveform of the aluminum foil paper measurement signal; d1 - d5 are the detail coefficients of each level, representing the detail information of the signal at different scales; a5 is the final approximation coefficient, which contains the information of the signal in the low-frequency part. d1 - d5 represent the five-layer decomposition of the original signal. As the decomposition level increases, the detail coefficients are gradually decomposed into high-frequency information of smaller scales, thereby capturing more subtle changes and noises of the signal. The waveforms of d1 - d5 have little correlation and relatively small absolute values, so it is considered that the noise of the aluminum foil paper signal is concentrated here. Set all the parameters in the signals of d1 - d5 to zero, and reconstruct the signal to obtain the denoised signal, as Figure 4 shown. Compared with the Figure 2 original signal shown schematically, the high-frequency spike noise of the denoised signal is significantly reduced. At this time, the waveform basically only contains useful signals, the noise is removed, and it is easier to judge whether there are defects in the aluminum foil paper signal.

[0038] Step S3: Use the fast Fourier transform to perform spectral analysis on the denoised aluminum foil paper measurement signal to obtain the corresponding spectrogram;

[0039] The present invention uses the fast Fourier transform to perform spectral analysis on the above-mentioned aluminum foil paper measurement signal after wavelet analysis denoising, so as to detect the integrity of the aluminum foil paper. Specifically, the fast Fourier transform is used to perform spectral analysis on the denoised signal. In order to prevent the amplitude of the spectrogram from concentrating at 0 Hz and affecting the observation of the spectrogram results, it is necessary to remove the DC component in the signal. The method for removing the DC component is to use a high-pass filter to remove the part with a frequency lower than 1 Hz and retain the remaining frequency part, making the signal more stable and facilitating frequency-domain analysis. As Figure 5 shown, it is the spectrogram generated after filtering the six signals in the previous example.

[0040] Step S4: According to the spectrogram, determine the target frequency at which the maximum amplitude appears in the denoised signal of the cigarette case aluminum foil paper, and determine whether there are defects in the cigarette case aluminum foil paper.

[0041] Since the aluminum foil paper defect signal has lost some of the original signal components, its spectrogram usually shows fewer frequency components and energy, especially at the missing part of the frequency, the amplitude decreases. By comparing the specific values of the amplitudes at specific frequencies in the spectrograms of different signals and the complete signal, it is possible to determine whether there are defects in the cigarette case aluminum foil paper. Specifically, from Figure 5 the schematic spectrogram, it can be seen that the effective signal of the cigarette case aluminum foil paper is concentrated in the 0-20 Hz frequency band, and the maximum amplitude appears at 2 Hz. Moreover, for different measurement signals, the maximum amplitude appears at 2 Hz. For the aluminum foil paper detection signals at different rotational speeds, by comparing the maximum amplitude at 2 Hz in the denoised signal spectrogram with the maximum amplitude of the complete signal, it is possible to determine whether there are defects. If the maximum amplitude is less than the maximum amplitude of the complete signal, it is determined that there are defects in the cigarette case aluminum foil paper: for the aluminum foil paper measurement signal at a speed of 120, when the maximum amplitude at 2 Hz in its denoised spectrogram is less than 99.5 at the same target frequency of the complete signal, it can be determined that there are defects, otherwise there are no defects; similarly, for the aluminum foil paper detection signal at a speed of 150, when the maximum amplitude in its denoised spectrogram is less than 100.3, it can be determined that there are defects, otherwise there are no defects; similarly, for the aluminum foil paper detection signal at a speed of 300, when the maximum amplitude in its denoised spectrogram is less than 97.7, it can be determined that there are defects, otherwise there are no defects.

[0042] In summary, the main design concept of the present invention is to use an eddy current sensor to collect more detailed defect information of the internal aluminum foil through the surface of the cigarette case, and extract features from the more detailed frequency domain and time domain information through wavelet analysis and fast Fourier transform, which is particularly applicable to the detection of complex surfaces and various types of defects of aluminum foil. During specific implementation, after denoising the wavelet basis function and decomposition level of the measured signal of the cigarette case aluminum foil, spectral analysis is performed on the denoised signal through fast Fourier transform, and then compared with the complete signal of normal aluminum foil to determine whether defects exist. The present invention avoids the influence of light and illumination, and conducts detection from the perspective of complete aluminum foil, thereby significantly improving the efficiency, comprehensiveness, and accuracy of defect detection of cigarette case aluminum foil.

[0043] In the embodiments of the present invention, if terms expressing directions are mentioned, they are based on the relative concepts of the embodiments. In addition, "at least one" means one or more, and "a plurality" means two or more. "And / or" describes the association relationship of associated objects and indicates that three relationships can exist. For example, A and / or B can represent the situation where A exists alone, A and B exist simultaneously, or B exists alone. Where A and B can be singular or plural. The character " / " generally represents an "or" relationship between the preceding and following associated objects. "At least one of the following" and its similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, and c can represent: a, b, c, a and b, a and c, b and c, or a, b, and c, where a, b, and c can be single or multiple.

[0044] The structure, features, and effects of the present invention have been described in detail based on the embodiments shown in the drawings above. However, the above are only the preferred embodiments of the present invention. It should be noted that for the technical features involved in the above embodiments and their preferred methods, those skilled in the art can reasonably combine and match them into various equivalent solutions without departing from or changing the design concept and technical effects of the present invention; therefore, the scope of implementation of the present invention is not limited by the drawings shown. Any changes made in accordance with the concept of the present invention or equivalent embodiments modified to equivalent changes that still do not exceed the spirit covered by the description and the drawings should be within the protection scope of the present invention.

Claims

1. A method for detecting aluminum foil defects based on wavelet analysis and Fourier transform, characterized in that: include: The eddy current sensor is used to collect the aluminum foil measurement signal through the surface of the cigarette box; Using wavelet analysis to remove noise in the aluminum foil measurement signal to obtain an aluminum foil measurement signal containing only useful information; The denoised aluminum foil measurement signal is analyzed by fast Fourier transform to obtain the corresponding spectrum diagram; According to the frequency spectrum, the target frequency at which the denoised signal of the cigarette box aluminum foil paper has the maximum amplitude is determined, and based on the target frequency, it is determined whether the cigarette box aluminum foil paper has defects.

2. The aluminum foil defect detection method based on wavelet analysis and Fourier transform according to claim 1 is characterized in that: The collecting of the aluminum foil measurement signal comprises: After setting a predetermined distance between the eddy current sensor and the surface of the cigarette box, the eddy current sensor probe detects the cigarette box placed on the rotating disk; The single-sheet and double-sheet identification instrument receives the aluminum foil signal sensed by the eddy current sensor probe to obtain the original waveform, which includes a normal complete waveform and a measured waveform.

3. The aluminum foil defect detection method based on wavelet analysis and Fourier transform according to claim 2 is characterized in that: The collecting of the aluminum foil measurement signal also includes: controlling the disc to rotate at different preset speeds by a speed regulating motor to simulate the motion state of the cigarette box on the actual assembly line.

4. The aluminum foil defect detection method based on wavelet analysis and Fourier transform according to claim 3 is characterized in that: The method of determining whether the aluminum foil of a cigarette box is defective based on a target frequency includes: based on the measurement signals of the aluminum foil at different rotation speeds, comparing the first maximum amplitude at the target frequency in the denoised signal spectrum diagram with the second maximum amplitude at the target frequency in the spectrum diagram of the complete signal; if the first maximum amplitude is smaller than the second maximum amplitude, determining that the aluminum foil of the cigarette box is defective; wherein the target frequency is a pre-obtained frequency in the frequency band where the aluminum foil defect is located.

5. The aluminum foil defect detection method based on wavelet analysis and Fourier transform according to claim 1 is characterized in that: After selecting the wavelet basis function and setting the number of decomposition layers, the aluminum foil measurement signal is decomposed into detail coefficients and approximate coefficients of different frequency components; As the number of decomposition layers increases, the detail coefficients are gradually decomposed into smaller-scale high-frequency information; Based on the high-frequency information at a smaller scale, the variation and noise in the aluminum foil measurement signal are obtained; The noise in the detail coefficients that have no correlation and a small absolute value is set to zero and the signal is reconstructed to obtain the denoised aluminum foil measurement signal.

6. The aluminum foil defect detection method based on wavelet analysis and Fourier transform according to any one of claims 1 to 5, characterized in that: The obtaining of the corresponding frequency spectrum includes: using a high-pass filter to remove the part with a frequency lower than 1 Hz, and retaining the remaining frequency parts.

Images