Cigarette lining paper defect online detection method and system based on eddy current

Abstract

The invention discloses a cigarette lining paper defect online detection method and system based on eddy current, and relates to the technical field of cigarette detection.The method comprises the steps that a double-probe transmission type eddy current sensor collects a voltage signal of cigarette packet lining paper; performing feature extraction on the voltage signal to obtain a feature vector; inputting the feature vector into a pre-trained defect identification model, and outputting a defect type and a defect type confidence coefficient; and marking the cigarette according to the defect identification result. According to the scheme, the defects of the cigarette packet lining paper made of the aluminum foil material can be accurately recognized, positioned in real time and accurately removed.

Description

Technical Field

[0001] This invention relates to the field of cigarette inspection technology, and in particular to an online detection method and system for defects in cigarette inner liner paper based on eddy currents. Background Technology

[0002] As the initial packaging layer of cigarettes, the integrity of the inner lining paper directly affects the aroma preservation, moisture protection, and appearance quality of the cigarette product. The inner lining paper is produced from rolled raw materials through more than ten packaging stations, including creasing, cutting, conveying, and folding. During the wrapping process, due to the compact design and limited space of the packaging machine molds, the detection devices cannot cover the entire production process. This results in the inability to effectively control tearing, misalignment, and damage to the inner lining paper caused by cutting and folding at multiple stations, thus negating its functions of moisture protection, mildew prevention, and preventing loss of tobacco aroma.

[0003] Currently, the industry commonly uses inductive proximity switches for inner lining paper defect detection. While this method has some effectiveness, it has the following significant shortcomings: Shielding problem: With product upgrades, many cigarette brands use aluminum foil coating on the outer hard pack packaging. High-frequency electromagnetic fields cannot penetrate this aluminum foil layer, causing the built-in inductive sensor to fail and be unable to detect the inner lining paper; Detection blind zone: Due to the limited internal space of the packaging machine, the sensor cannot cover the entire area of ​​the cigarette pack's inner lining paper, resulting in missed detection of defects such as folds, scratches, and localized damage. At the same time, traditional switch sensors can only detect the "presence" of metal and cannot quantify the shape and size of defects, making them insensitive to minor defects. Summary of the Invention

[0004] In view of the above, the present invention aims to provide an online detection method and system for defects in cigarette inner liner paper based on eddy currents, so as to solve the aforementioned technical problems.

[0005] The technical solution adopted in this invention is as follows:

[0006] This invention provides an online detection method for defects in cigarette inner liner paper based on eddy currents, including:

[0007] Voltage signals of the inner lining paper of cigarette packs are collected using a dual-probe transmission eddy current sensor.

[0008] Feature extraction is performed on the voltage signal to obtain a feature vector;

[0009] Input the feature vector into the pre-trained defect identification model, and output the defect type and the defect type confidence score;

[0010] Cigarettes are marked based on the defect identification results.

[0011] Optionally, feature extraction is performed on the voltage signal to obtain a feature vector, including:

[0012] The voltage signal is preprocessed to obtain a standard voltage signal;

[0013] The standard voltage signal is uniformly sampled within a preset frequency band to obtain N frequency points;

[0014] The amplitudes of the N frequency points are extracted to obtain an N-dimensional feature vector with time-domain statistical characteristics.

[0015] Optionally, the feature vector is input into a pre-trained defect recognition model, which outputs the defect type and confidence level, including:

[0016] The defect identification model extracts local frequency domain feature values ​​from the feature vector;

[0017] The defect type is obtained by mapping the local frequency domain feature values.

[0018] Calculate the normalized probability of the defect type to obtain the defect type confidence level.

[0019] Optionally, when the confidence level of any defect type is greater than a preset threshold, it is determined that the cigarette inner liner paper has a defect of that type.

[0020] Optionally, the defect types include missing, broken, and folded.

[0021] Optionally, the online detection method for defects in cigarette inner liner paper based on eddy currents further includes:

[0022] When a marked defective cigarette pack arrives at the rejection station, the rejection mechanism removes the marked defective cigarette pack from the conveyor channel.

[0023] Optionally, the defect identification model training process includes:

[0024] Collect detection signals from historical normal cigarette packs and historical defective cigarette packs;

[0025] The detection signals of historical normal cigarette packs and historical defective cigarette packs were subjected to fast Fourier transform and then frequency domain feature extraction was performed to obtain a labeled training dataset.

[0026] The training dataset is input into a one-dimensional convolutional neural network for model training, and the parameters are optimized using the cross-entropy loss function to obtain a defect recognition model.

[0027] The present invention also provides an online detection system for defects in cigarette inner liner paper based on eddy currents, which performs the method described above, the system comprising:

[0028] Cigarette pack conveying channel;

[0029] An eddy current detection module is installed on the cigarette pack conveying channel;

[0030] A rejection mechanism is installed on the cigarette pack conveying channel;

[0031] The eddy current detection module and the rejection mechanism are both electrically connected to the programmable logic controller.

[0032] Optionally, the eddy current detection module uses a dual-probe transmission-type eddy current sensor, symmetrically arranged on both sides of the cigarette pack conveying channel, to emit an alternating magnetic field that penetrates the outer aluminum foil of the cigarette pack and receive the magnetic field signal modulated by the inner lining paper.

[0033] Optionally, the online detection system for defects in cigarette inner liner paper based on eddy currents further includes: a rejection collection box fixed on one side of the cigarette pack conveying channel.

[0034] The above-described solution of the present invention has at least the following beneficial effects:

[0035] The above-described solution of the present invention acquires the voltage signal of the inner lining paper of cigarette packs using a dual-probe transmission eddy current sensor; extracts features from the voltage signal to obtain a feature vector; inputs the feature vector into a pre-trained defect recognition model, outputting the defect type and defect type confidence level; and marks the cigarettes according to the defect recognition results. This method enables accurate identification, real-time location, and precise removal of defects in the inner lining paper of aluminum foil cigarette packs. Attached Figure Description

[0036] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described below with reference to the accompanying drawings, wherein:

[0037] Figure 1 A flowchart of an online defect detection method for cigarette inner liner paper based on eddy currents provided in an embodiment of the present invention.

[0038] Figure 2 A schematic diagram of an online defect detection system for cigarette inner liner paper based on eddy currents provided in an embodiment of the present invention.

[0039] Figure 3 This is a partial schematic diagram of an online defect detection system for cigarette inner liner paper based on eddy currents, provided in an embodiment of the present invention.

[0040] Figure 4 This is a schematic diagram of the eddy current detection module provided in an embodiment of the present invention.

[0041] Figure 5 This is a schematic diagram showing the arrangement of various sensors according to an embodiment of the present invention. Detailed Implementation

[0042] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0043] This invention proposes an embodiment of an online detection method for defects in cigarette inner liner paper based on eddy currents. Specifically, as follows: Figure 1 As shown, it includes:

[0044] Step 11: Acquire the voltage signal of the inner lining paper of the cigarette pack using a dual-probe transmission eddy current sensor.

[0045] In this embodiment, the eddy current detection module uses a dual-probe transmission-type eddy current sensor excited by low frequency (preferably 500Hz-10kHz). The sensor is symmetrically arranged on both sides of the cigarette pack conveying channel 1. The transmitting coil generates an alternating magnetic field that penetrates the cigarette pack, and the receiving coil senses the magnetic field signal modulated by the inner lining paper. The low-frequency excitation effectively increases the eddy current penetration depth and overcomes the shielding effect of the outer aluminum foil.

[0046] Step 12: Extract features from the voltage signal to obtain a feature vector;

[0047] In this embodiment, the acquired raw voltage signal is preprocessed, including: wavelet denoising using the Db4 wavelet basis with 5-level decomposition to filter out high-frequency interference and obtain a standard voltage signal;

[0048] A fast Fourier transform is performed on the denoised standard voltage signal to extract the amplitude corresponding to N frequency points uniformly sampled within the 0–20Hz frequency band (especially at 2Hz), forming an N-dimensional feature vector. At the same time, time-domain statistical features (such as mean, variance, and peak value) are added as supplements to form a comprehensive feature vector, that is, an N-dimensional feature vector with time-domain statistical features is obtained.

[0049] Step 13: Input the feature vector into the pre-trained defect recognition model and output the defect type and defect type confidence score;

[0050] In this embodiment, the pre-selected model is first trained to obtain a defect recognition model. The specific defect recognition model training process includes:

[0051] Detection signals from historical normal cigarette packs and historical defective cigarette packs are collected. These signals are then subjected to Fast Fourier Transform (FFT) for frequency domain feature extraction, resulting in a labeled training dataset. The dataset is then divided into training, validation, and test sets (e.g., 7:2:1). The labeled training dataset is input into a one-dimensional convolutional neural network (1D-CNN) or a long short-term memory network (LSTM) for training. Parameter optimization is performed using the cross-entropy loss function, with Adam selected as the optimizer. The resulting defect recognition model for the target is then obtained.

[0052] Early stopping is employed during training to prevent overfitting, and a validation set is used to monitor model performance. When the production line changes cigarette brand or packaging materials, new data can be collected to fine-tune the model and improve its adaptability.

[0053] Furthermore, the N-dimensional feature vector with time-domain statistical features obtained in step 12 is input into the defect identification model to identify defects in the cigarette inner liner paper.

[0054] When a defect recognition model is trained using a one-dimensional convolutional neural network, the input layer of the defect recognition model receives an N-dimensional feature vector. The N-dimensional feature vector is output to the convolutional layer; where B is the batch size and 1 is the number of channels.

[0055] The convolutional layer uses ReLU activation to extract local frequency domain patterns, obtaining the energy changes in the 0–20Hz frequency band, and outputs them to the pooling layer. There are two convolutional layers here; the first convolutional layer performs a convolution operation on the input feature vector X.

[0056] ;

[0057] The second convolutional layer performs a convolution operation on the output of the first convolutional layer:

[0058] ;

[0059] in, , , , For convolution kernel, , K represents the number of channels, and K represents the kernel size. , , , For bias.

[0060] Pooling layers affect the output of convolutional layers. Max pooling and dimensionality reduction compression are performed to retain important information while reducing data volume, and key features are output to the fully connected layer.

[0061] P is the step size;

[0062] Output shape: .

[0063] The fully connected layer maps key features to corresponding defect categories, including "missing", "damaged", and "folded", and outputs the corresponding defect categories to the output layer.

[0064] Pooled features Flattened into a vector:

[0065] ;

[0066] Fully connected transformation:

[0067] weight matrix bias Where D is the number of hidden units:

[0068] Output shape .

[0069] The output layer uses a Softmax activation function to normalize defect categories into a probability distribution, outputting the confidence scores for defect categories such as "missing," "damaged," and "folded."

[0070] weight matrix bias C represents the number of categories, such as "missing", "damaged", and "folded" (3 categories).

[0071] The confidence level p is obtained by normalizing the probability distribution using Softmax:

[0072] ;

[0073] ;

[0074] Output shape: .

[0075] When training a defect identification model using a Long Short-Term Memory (LSTM) network, the feature vector is treated as a time series, input into an LSTM unit to capture temporal dependencies, and then connected to a fully connected layer and an output layer for classification and output.

[0076] The defect identification model sets a threshold (e.g., 0.9). When the confidence level of a certain type of defect exceeds the threshold, the cigarette is determined to have that type of defect. The output includes: defect type, confidence level, timestamp, and cigarette pack serial number.

[0077] Step 14: Mark the cigarettes based on the defect identification results.

[0078] In this embodiment, the defect identification results are input into the system's programmable logic controller (PLC). The PLC receives the defect identification results, marks defective cigarette packs, and the marked packs move along the cigarette pack conveyor channel, being relayed sequentially by a series of downstream photoelectric sensors spaced less than the length of the pack. When a marked pack reaches the final rejection station, the PLC immediately triggers the rejection mechanism to remove the defective pack from the production line. This mechanism eliminates the need for real-time calculation of the pack's position and speed, greatly simplifying the control logic and improving reliability and response speed.

[0079] Embodiments of the present invention can also aggregate and store all detection data such as time, serial number, defect type, confidence level, and image snapshots in real time, dynamically display production speed, instantaneous pass rate, and defect distribution map, generate quality statistical reports by shift, day, and month, automatically perform defect Pareto analysis, locate the root cause of major problems, predict potential quality fluctuations based on historical data trends, and provide early warnings when key indicators (such as the number of consecutive defects) exceed limits. It also provides a standard data interface to seamlessly upload quality data to the enterprise MES system, realizing closed-loop quality management.

[0080] like Figures 2 to 4 As shown, embodiments of the present invention also provide an online detection system for defects in cigarette inner liner paper based on eddy currents, which performs the method as described in the above embodiments. The system includes:

[0081] Cigarette pack conveying channel 1;

[0082] Eddy current detection module 2 is installed on the cigarette pack conveying channel 1;

[0083] The rejection mechanism 3 is installed on the cigarette pack conveying channel 1;

[0084] Both the eddy current detection module 2 and the rejection mechanism 3 are electrically connected to the programmable logic controller.

[0085] In this embodiment, the cigarette pack conveying channel 1 is the conveying channel for cigarette packs in the packaging machine. The eddy current detection module 2 mainly consists of a mounting housing, an accelerating roller assembly, a dual-probe transmission-type eddy current sensor 21, and an embedded processor. The dual-probe transmission-type eddy current sensor 21 is symmetrically arranged on both sides of the cigarette pack conveying channel 1, and is used to emit an alternating magnetic field that penetrates the outer aluminum foil of the cigarette pack 5 and receive the voltage signal modulated by the inner lining paper. The eddy current detection module 2 is mounted on the cigarette pack conveying channel 1 via a mounting plate 11. It should be noted that the cigarette pack conveying channel 1 in this embodiment is divided into two sections, located at the inlet and outlet ends of the eddy current detection module 2, respectively. The accelerating roller assembly is located at the inlet end of the eddy current detection module 2. The cigarette pack 5 is conveyed to the eddy current detection module 2 through the cigarette pack conveying channel 1. Before entering the detection area, the cigarette pack 5 passes through the acceleration roller assembly, which clamps the cigarette pack 5 and accelerates it through the detection area and conveys it to the cigarette pack conveying channel 1 at the rear exit. This allows the dual-probe transmission eddy current sensor 21 to quickly acquire and detect the voltage signal of the inner lining paper of the cigarette pack, improving detection efficiency. The eddy current detection module 2 is also equipped with guide bars 23 between the dual-probe transmission eddy current sensors 21 to guide the cigarette pack 5 during acceleration and maintain an orderly queue. The embedded processor is equipped with a trained defect recognition model, which obtains the defect recognition result.

[0086] The accelerating roller assembly includes a motor 221, a transmission gear set 222 fixedly connected to the rotating shaft of the motor, and a speed-increasing wheel 223 coaxially connected to the transmission gear set 222. An accelerating circular belt 224 is fixedly installed on the speed-increasing wheel 223. The accelerating roller assembly clamps the cigarette pack 5 through the accelerating circular belt 224. The motor 221 drives the transmission gear set 222 to rotate, which in turn drives the speed-increasing wheel 223 to move, thereby driving the accelerating circular belt 224 to move, so that the cigarette pack 5 can quickly pass through the detection area.

[0087] The rejection mechanism 3 uses a pneumatic nozzle, and its movement is controlled by correlating with the front-end detection results. A rejection collection box 4 is fixedly installed on one side of the end of the cigarette pack conveying channel 1. Both the eddy current detection module 2 and the rejection mechanism 3 are electrically connected to a programmable logic controller (PLC), which is located in the electrical control cabinet 6.

[0088] The eddy current detection module 2 transmits the detected signal to the PLC. The PLC receives the defect identification result and marks the defective cigarette packs. The marked cigarette packs move along the cigarette pack conveyor channel and are relayed by a series of photoelectric sensors with a spacing smaller than the length of the cigarette pack downstream. When the marked cigarette packs move to the end rejection station, the PLC immediately triggers the rejection mechanism to remove the defective cigarette packs from the production line to the rejection collection box 4.

[0089] like Figure 5As shown, in this embodiment, when the online detection system for cigarette inner liner paper defects based on eddy current is used, the cigarette pack 5 moves forward along the conveyor channel. When it enters the eddy current detection module 2, the dual-probe transmission eddy current sensor 21 detects defects on both sides of the cigarette pack 5. When the detection is qualified, the cigarette pack 5 flows to the next process through the conveyor line. When the marking and identification sensor 8 detects that it has an attached mark, that is, the cigarette pack is unqualified, the PLC controls the rejection mechanism 3 at the rear end according to the detection result. The rejection mechanism 3 removes the cigarette pack 5 from the cigarette pack conveyor channel 1 by pneumatic means through the air nozzle of the rejection mechanism 3 and enters the rejection collection box 4.

[0090] The online defect detection system for cigarette inner liner paper based on eddy current in this embodiment is also equipped with a human-machine interface screen 7. The human-machine interface screen 7 is mainly used for setting up the eddy current detection module 2, controlling the rejection and triggering of sensors, and querying the detection structure.

[0091] The present invention relates to an online detection method and system for cigarette inner lining paper defects based on eddy currents. This method utilizes eddy current signal acquisition and combines it with three-dimensional simulation technology to construct a simulation model of cigarette inner lining paper defects based on the eddy current principle. By performing wavelet analysis denoising preprocessing on the original detection signal of aluminum foil paper, and by using fast Fourier transform to extract and analyze spectral features, the method achieves accurate identification of various defects in aluminum foil paper. An online defect detection device adapted to the conveyor line is designed, integrating defect identification, real-time positioning, and precise rejection functions. Furthermore, a human-computer interaction and system control platform is built to achieve functions such as defect quantity statistics and data visualization.

[0092] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0093] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0094] In the embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0095] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0096] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0097] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0098] Furthermore, it should be noted that in the apparatus and method of the present invention, it is obvious that the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered equivalent solutions of the present invention. Moreover, the steps performing the above series of processes can naturally be executed in the order described, but are not necessarily required to be executed in chronological order; some steps can be executed in parallel or independently of each other. Those skilled in the art will understand that all or any step or component of the method and apparatus of the present invention can be implemented in any computing device (including processors, storage media, etc.) or network of computing devices, in hardware, firmware, software, or a combination thereof. This is something that those skilled in the art can achieve by using their basic programming skills after reading the description of the present invention.

[0099] Therefore, the object of the present invention can also be achieved by running a program or a set of programs on any computing device. The computing device can be a known general-purpose device. Therefore, the object of the present invention can also be achieved simply by providing a program product containing program code implementing the method or apparatus. That is, such a program product also constitutes the present invention, and the storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any known storage medium or any storage medium developed in the future. It should also be noted that in the apparatus and method of the present invention, it is obvious that the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered equivalent to the present invention. Furthermore, the steps performing the above series of processes can naturally be performed in the order described, but are not necessarily required to be performed in chronological order. Some steps can be performed in parallel or independently of each other.

[0100] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for online detection of defects in cigarette inner liner paper based on eddy currents, characterized in that, include: Voltage signals of the inner lining paper of cigarette packs are collected using a dual-probe transmission eddy current sensor. Feature extraction is performed on the voltage signal to obtain a feature vector; Input the feature vector into the pre-trained defect identification model, and output the defect type and the defect type confidence score; Cigarettes are marked based on the defect identification results.

2. The online detection method for cigarette inner liner paper defects based on eddy currents according to claim 1, characterized in that, Feature extraction is performed on the voltage signal to obtain a feature vector, including: The voltage signal is preprocessed to obtain a standard voltage signal; The standard voltage signal is uniformly sampled within a preset frequency band to obtain N frequency points; The amplitudes of the N frequency points are extracted to obtain an N-dimensional feature vector with time-domain statistical characteristics.

3. The online detection method for cigarette inner liner paper defects based on eddy currents according to claim 1, characterized in that, Input the feature vector into the pre-trained defect recognition model, and output the defect type and confidence level, including: The defect identification model extracts local frequency domain feature values ​​from the feature vector; The defect type is obtained by mapping the local frequency domain feature values. Calculate the normalized probability of the defect type to obtain the defect type confidence level.

4. The online detection method for cigarette inner liner paper defects based on eddy currents according to claim 1, characterized in that, When the confidence level of any defect type is greater than a preset threshold, it is determined that the cigarette inner liner paper has a defect of that type.

5. The online detection method for cigarette inner liner paper defects based on eddy currents according to claim 1, characterized in that, The defect types include missing, broken, and folded.

6. The online detection method for cigarette inner liner paper defects based on eddy currents according to claim 1, characterized in that, Also includes: When a marked defective cigarette pack arrives at the rejection station, the rejection mechanism removes the marked defective cigarette pack from the conveyor channel.

7. The online detection method for cigarette inner liner paper defects based on eddy currents according to claim 1, characterized in that, The training process of the defect identification model includes: Collect detection signals from historical normal cigarette packs and historical defective cigarette packs; The detection signals of historical normal cigarette packs and historical defective cigarette packs were subjected to fast Fourier transform and then frequency domain feature extraction was performed to obtain a labeled training dataset. The training dataset is input into a one-dimensional convolutional neural network for model training, and the parameters are optimized using the cross-entropy loss function to obtain a defect recognition model.

8. An online detection system for defects in cigarette inner liner paper based on eddy currents, characterized in that, The system for performing the method as described in any one of claims 1 to 7 includes: Tobacco pack conveying channel (1); An eddy current detection module (2) is installed on the cigarette pack conveying channel (1); The rejection mechanism (3) is installed on the cigarette pack conveying channel (1); The eddy current detection module (2) and the rejection mechanism (3) are both electrically connected to the programmable logic controller.

9. The online detection system for cigarette inner liner paper defects based on eddy currents according to claim 8, characterized in that, The eddy current detection module (2) uses a dual-probe transmission eddy current sensor, which is symmetrically arranged on both sides of the cigarette pack conveying channel (1) to emit an alternating magnetic field that penetrates the outer aluminum foil of the cigarette pack and receive the magnetic field signal modulated by the inner lining paper.

10. The online detection system for cigarette inner liner paper defects based on eddy currents according to claim 8, characterized in that, Also includes: A rejection collection box (4) is fixed on one side of the cigarette pack conveying channel (1).

Images