Plate automatic detection method based on air coupling sensor

Abstract

The invention discloses a plate automatic detection method based on an air coupling sensor, and the method comprises the following steps: S1, plate positioning and conveying: placing a to-be-detected plate on a conveying mechanism, positioning the plate through a positioning sensor (such as a laser positioning sensor), and guaranteeing that the plate is stably conveyed along a preset path in a detection process; according to the automatic panel detection method based on the air coupling sensor, environmental interference is eliminated through environmental parameter acquisition compensation, comprehensive panel detection is realized by adopting a multi-channel air coupling sensor array, and the defect identification and positioning precision is improved by combining wavelet transform filtering preprocessing and a deep learning defect identification model; detection of physical parameters such as plate defects, thickness and density can be synchronously completed, the detection process is high in automation degree, manual intervention is not needed, the detection accuracy, stability and efficiency are effectively improved, the equipment investment cost is reduced, and the device is suitable for large-scale online detection and assists in improving the plate production quality control level.

Description

Technical Field

[0001] This invention relates to the field of sheet metal inspection technology, and in particular to an automatic sheet metal inspection method based on an air-coupled sensor. Background Technology

[0002] In the production and processing of sheet materials, quality inspection is a crucial step in ensuring product performance and safety. Traditional sheet material inspection methods mainly include manual visual inspection, ultrasonic contact inspection, and radiographic inspection. Manual visual inspection is inefficient, highly subjective, and cannot accurately identify internal defects in the sheet material; ultrasonic contact inspection requires applying a coupling agent between the sensor and the sheet surface, which is cumbersome, and the uniformity of the coupling agent application can affect the detection accuracy. Furthermore, the coupling agent may contaminate the sheet surface, making it unsuitable for sheet materials requiring high surface cleanliness; radiographic inspection poses radiation safety hazards, is harmful to the health of operators, and is costly, making large-scale online inspection difficult.

[0003] Non-contact ultrasonic sensors (air-coupled sensors) do not require direct contact with the object being measured or the use of a coupling agent, offering advantages such as high detection efficiency, no pollution, and convenient operation, making them promising for applications in the field of sheet metal inspection. However, current sheet metal inspection methods based on air-coupled sensors still have some shortcomings. For example, detection accuracy is easily affected by environmental factors (such as temperature, humidity, and airflow), high requirements are placed on the flatness of the sheet metal surface, and there is a lack of adaptive detection algorithms for different types of sheet metal, making it difficult to meet diverse inspection needs. Therefore, we propose an automatic sheet metal inspection method based on air-coupled sensors. Summary of the Invention

[0004] The main objective of this invention is to provide an automatic detection method for sheet metal based on an air-coupled sensor, which can effectively solve the problems in the background art.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: An automatic inspection method for sheet metal based on an air-coupled sensor includes the following steps: S1. Sheet Positioning and Conveying: The sheet to be inspected is placed on the conveying mechanism. Positioning sensors (such as laser positioning sensors) are used to locate the sheet, ensuring stable conveying along a preset path during the inspection process. The conveying mechanism is driven by a servo motor, and the conveying speed can be adjusted according to the size of the sheet and the inspection requirements.

[0006] S2. Environmental Parameter Acquisition and Compensation: Environmental sensors are installed in the detection area to collect parameters such as ambient temperature, humidity, and airflow velocity in real time. Based on a pre-set environmental compensation model, environmental compensation is applied to the detection signal of the air-coupled sensor to eliminate the influence of environmental factors on detection accuracy. The environmental compensation model is trained using a large amount of experimental data, establishing a mapping relationship between environmental parameters and the correction amount of the detection signal.

[0007] S3. Arrangement and Signal Acquisition of the Air-Coupled Sensor Array: A multi-channel air-coupled sensor array is used. The sensor array is mounted on the testing frame, with the transmitter and receiver of the sensor located on the top and bottom sides of the material, respectively, maintaining a preset distance (generally 5-20mm) from the surface of the material. The operating frequency (e.g., 200kHz-2MHz) and sampling frequency of the sensor are set according to the type of material and testing requirements. The sensor array is activated to transmit ultrasonic signals to the material and to receive the reflected and transmitted signals after propagation through the material.

[0008] S4. Signal Preprocessing: The acquired raw signal is preprocessed, including filtering, amplification, denoising, and signal normalization. Wavelet transform filtering is used to remove high-frequency noise and low-frequency interference from the signal. Automatic gain control circuitry amplifies the signal to ensure the amplitude is within a suitable range. Then, signal normalization is performed to eliminate the impact of amplitude differences on subsequent analysis.

[0009] S5. Feature Extraction and Defect Identification: Feature extraction is performed on the preprocessed signal. Extracted feature parameters include peak values, trough values, propagation time, frequency spectrum characteristics, and energy entropy. The extracted feature parameters are input into a trained defect identification model. This model is constructed using deep learning algorithms (such as convolutional neural networks and recurrent neural networks) and trained on a large amount of sample data labeled with defect types and locations. The model outputs whether the board material has defects, the type of defect (e.g., voids, cracks, delamination), and its location coordinates.

[0010] S6. Physical Parameter Detection: The thickness of the board is calculated based on the propagation time of the ultrasonic signal and the sound velocity characteristics of the board. The sound velocity characteristics are obtained by pre-calibrating a standard board, establishing a model relating the board thickness to the ultrasonic propagation time. Simultaneously, the density of the board is estimated based on the energy attenuation characteristics of the signal, combined with the correspondence between the board density and the sound attenuation coefficient.

[0011] S7. Output and Feedback of Inspection Results: The defect inspection results (defect type, location, size) and physical parameters (thickness, density) are displayed in real time on the screen, and an inspection report is generated. For boards found to have defects, the control system controls the conveyor mechanism to sort them to the non-conforming product area; for qualified boards, they are conveyed to the next process. Simultaneously, the inspection data is uploaded to the database for storage and management, facilitating subsequent quality traceability and data analysis.

[0012] Compared with the prior art, the present invention has the following beneficial effects: By employing environmental parameter acquisition and compensation steps, the influence of environmental factors such as temperature, humidity, and airflow on detection accuracy can be effectively eliminated, thereby improving the accuracy and stability of detection results. Using a multi-channel air-coupled sensor array for signal acquisition expands the detection range, improves detection efficiency, enables comprehensive detection of the board material, and avoids missed detections. By employing signal preprocessing methods such as wavelet transform filtering and deep learning defect recognition models, signal features can be accurately extracted, enabling rapid identification and localization of different types of defects with high accuracy and strong anti-interference capability. It can simultaneously detect defects in sheet materials and measure physical parameters (thickness, density), offering comprehensive functionality to meet diverse testing needs and reduce the investment cost of testing equipment. The entire testing process is highly automated, requires no manual intervention, is easy to operate, and has high testing efficiency. It is suitable for large-scale online testing and helps improve the quality control level of board production. Attached Figure Description

[0013] Figure 1 This is a flowchart of an automatic detection method for sheet metal based on an air coupling sensor according to the present invention. Detailed Implementation

[0014] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments. Example

[0015] An automatic detection method for wood panels based on a hollow-coupled sensor is disclosed. The object to be detected is a wood panel with dimensions of 2440mm × 1220mm × 18mm. The detection method of this invention is used for detection, and the specific steps are as follows: Step 1, Board Positioning and Conveying: Place the wood boards on the belt conveyor mechanism, and use laser positioning sensors to position the boards. Set the conveying speed to 0.5m / s.

[0016] Step 2: Environmental parameter acquisition and compensation: The environmental sensor collects the temperature of the detection area in real time as 25℃, humidity as 60%, and airflow speed as 0.3m / s, and corrects the detection signal according to the environmental compensation model.

[0017] Step 3: Arrangement and signal acquisition of the air-coupled sensor array: A 32-channel air-coupled sensor array is used, with the working frequency set to 500kHz and the sampling frequency to 10MHz. The distance between the sensor and the surface of the board is 10mm. The transmitting end emits ultrasonic signals, and the receiving end acquires reflected and transmitted signals.

[0018] Step 4, Signal Preprocessing: Wavelet transform filtering (db4 wavelet basis, decomposition level 3) is used to remove noise, automatic gain control circuit amplifies the signal to 0-5V, and then normalization is performed.

[0019] Step 5: Feature Extraction and Defect Identification: Extract the signal's peak value, propagation time, frequency spectrum features, and energy entropy, and input them into the trained convolutional neural network defect identification model. The model accurately identifies a 5mm diameter hole on the board with coordinates (1500mm, 800mm).

[0020] Step 6: Physical parameter detection: Based on the ultrasonic propagation time and the pre-calibrated sound velocity of wood (5000m / s), the thickness of the board is calculated to be 17.8mm, with an error within ±0.2mm; the density of the board is estimated to be 0.6g / cm³ based on the signal energy attenuation characteristics.

[0021] Step 7: Output and feedback of test results: The display screen shows the defect type as void, location coordinates (1500mm, 800mm), size 5mm, thickness 17.8mm, and density 0.6g / cm³. A test report is generated and uploaded to the database. The control system sorts the board into the non-conforming product area. Example

[0022] An automatic inspection method for sheet metal based on an air-coupled sensor is disclosed. The object to be inspected is a sheet metal (aluminum alloy) with dimensions of 1000mm × 500mm × 5mm. The inspection steps are as follows: Step 1, Plate Positioning and Conveying: The metal plate is placed on the roller conveyor mechanism, positioned by a laser positioning sensor, and the conveying speed is set to 1m / s.

[0023] Step 2: Environmental parameter acquisition and compensation: Ambient temperature 23℃, humidity 55%, airflow velocity 0.2m / s. The detection signal is corrected according to the environmental compensation model.

[0024] Step 3: Arrangement and signal acquisition of the air-coupled sensor array: A 16-channel air-coupled sensor array is used, with an operating frequency of 1MHz and a sampling frequency of 20MHz. The distance between the sensor and the surface of the board is 8mm. Ultrasonic reflected and transmitted signals are collected.

[0025] Step 4, Signal Preprocessing: Wavelet transform filtering (db6 wavelet basis, decomposition level 4) for noise reduction, automatic gain amplification and normalization.

[0026] Step 5: Feature Extraction and Defect Recognition: Extract feature parameters and input them into the recurrent neural network defect recognition model to identify a crack on the board with a length of 20mm and a width of 0.3mm, with location coordinates (600mm, 300mm).

[0027] Step 6: Physical parameter testing: Based on the sound velocity of aluminum alloy (6300m / s), the thickness is calculated to be 4.98mm, with an error of ±0.02mm; the estimated density is 2.7g / cm³.

[0028] Step 7, Output and Feedback of Test Results: Display the test results, generate a report and upload it to the database. Qualified boards are then transported to the next process.

[0029] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. An automatic detection method for sheet metal based on a hollow coupling sensor, characterized in that, Includes the following steps: S1. Plate positioning and conveying: The plate to be tested is placed on the conveying mechanism, the plate is positioned by the positioning sensor, and the speed of the conveying mechanism is adjusted to make the plate be conveyed along the preset path. S2. Environmental parameter acquisition and compensation: Environmental sensors are set up in the detection area to collect temperature, humidity and airflow speed, and the detection signal of the air coupler sensor is compensated according to the environmental compensation model. S3. Arrangement and signal acquisition of air-coupled sensor array: A multi-channel air-coupled sensor array is adopted, with the transmitter and receiver located on both sides of the plate. The working frequency and sampling frequency are set to acquire ultrasonic reflected signals and transmitted signals. S4. Signal preprocessing: Filtering, amplifying, denoising, and normalizing the acquired raw signal; S5. Feature Extraction and Defect Recognition: Extract signal feature parameters, input them into the trained defect recognition model, and output defect information; S6. Physical parameter detection: Calculate the thickness of the plate based on the ultrasonic propagation time and sound velocity characteristics, and estimate the density based on the signal energy attenuation characteristics; S7. Test Result Output and Feedback: Display test results, generate reports and upload them to the database, and sort out unqualified boards.

2. The automatic detection method for sheet metal based on an air-coupled sensor according to claim 1, characterized in that: The positioning sensor in S1 is a laser positioning sensor, and the conveying mechanism is driven by a servo motor.

3. The automatic detection method for sheet metal based on an air-coupled sensor according to claim 1, characterized in that: The environmental compensation model in S2 is established through training with experimental data, and a mapping relationship is established between environmental parameters and the correction amount of the detection signal.

4. The automatic detection method for sheet metal based on an air-coupled sensor according to claim 1, characterized in that: The S3 hollow coupling sensor is 5-20mm away from the surface of the plate, and its operating frequency is 200kHz-2MHz.

5. The automatic detection method for sheet metal based on an air-coupled sensor according to claim 1, characterized in that: In S4, wavelet transform filtering is used to remove noise, and the signal is amplified by automatic gain control circuit.

6. The automatic detection method for sheet metal based on an air-coupled sensor according to claim 1, characterized in that: The feature parameters extracted in S5 include peak value, valley value, propagation time, frequency spectrum features, and energy entropy. The defect identification model is constructed using a deep learning algorithm.

7. The automatic detection method for sheet metal based on an air-coupled sensor according to claim 1, characterized in that: The sound velocity characteristics of the plate in S6 are obtained by calibration using a standard plate.

Images