Method and system for detecting defects in bamboo board

Abstract

The application discloses a kind of bamboo board defect detection method and system, it is related to plate detection field, the method includes obtaining the target image of bamboo board to be detected, according to target image, based on image defect detection model, the visual defect range and visual confidence of bamboo board to be detected are determined;Obtain the multiple ultrasonic echo signals of bamboo board to be detected in multiple detection regions, according to multiple ultrasonic echo signals, the multiple ultrasonic confidence corresponding to multiple detection regions of bamboo board to be detected in multiple detection regions is determined, multiple detection regions and multiple ultrasonic echo signals one-to-one correspond;According to visual defect range, visual confidence, multiple detection regions and multiple ultrasonic confidence, the defect confidence of bamboo board to be detected in multiple detection regions is determined;According to the defect confidence of multiple detection regions, the defect detection result of bamboo board to be detected is determined.The application can improve the defect detection efficiency and accuracy of bamboo board.

Description

Technical Field

[0001] This application relates to the field of board material testing, and in particular to a method and system for detecting defects in bamboo boards. Background Technology

[0002] Bamboo, as a green and environmentally friendly recycled material, is widely used in flooring, furniture, and construction. During the processing of bamboo into boards, it is very easy to produce surface defects (such as bamboo green / bamboo yellow residue, surface cracks, insect holes, mildew spots, glue spots, etc.) or internal defects (internal cavities (bubbles), insufficient bonding strength (delamination / separation), internal micro-cracks, internal cracks, and impurities, etc.).

[0003] Currently, defect detection in bamboo panels mainly relies on manual experience, resulting in low efficiency and accuracy. Summary of the Invention

[0004] The purpose of this application is to provide a method and system for detecting defects in bamboo panels, which can improve the efficiency and accuracy of defect detection in bamboo panels.

[0005] To achieve the above objectives, this application provides the following solution: In a first aspect, this application provides a method for detecting defects in bamboo-based panels, including: Acquire the target image of the bamboo board to be inspected, and based on the target image, determine the visual defect range and visual confidence level of the bamboo board to be inspected using the image defect detection model. Multiple ultrasonic echo signals of the bamboo board to be tested in multiple detection areas are acquired. Based on the multiple ultrasonic echo signals, multiple ultrasonic confidence levels of the bamboo board to be tested in multiple detection areas are determined. The multiple detection areas correspond one-to-one with the multiple ultrasonic echo signals. Based on the visual defect range, visual confidence level, multiple detection areas, and multiple ultrasonic confidence levels, the defect confidence level of the bamboo board to be tested in multiple detection areas is determined; Based on the defect confidence levels of multiple detection areas, the defect detection results of the bamboo board to be inspected are determined.

[0006] Optionally, based on multiple ultrasonic echo signals, multiple ultrasonic confidence levels are determined for the bamboo board to be tested in multiple detection areas, including: Based on multiple ultrasonic echo signals, multiple instantaneous amplitude envelopes corresponding to the multiple ultrasonic echo signals are determined, and the multiple instantaneous amplitude envelopes correspond one-to-one with multiple detection areas; Based on multiple instantaneous amplitude envelopes, the plate structure type of multiple detection areas is determined; Based on the plate structure type and instantaneous amplitude envelope corresponding to the multiple detection areas, multiple ultrasonic confidence levels are determined for the multiple detection areas.

[0007] Optionally, based on multiple instantaneous amplitude envelopes, the plate structure type of multiple detection areas is determined, including: For each instantaneous amplitude envelope, based on the characteristic parameters corresponding to the instantaneous amplitude envelope and the preset evaluation rules, the plate structure type of the detection area corresponding to the instantaneous amplitude envelope is determined. The characteristic parameters corresponding to the instantaneous amplitude envelope include peak height, pulse width and rise time. The preset evaluation rules include: when the rise time is less than the product of the sound velocity difference coefficient and the preset reference rise time, and the pulse width is less than or equal to the product of the impedance mismatch coefficient and the preset reference pulse width, the board structure type of the detection area is determined to be bamboo-dominant; when the pulse width is greater than the product of the impedance mismatch coefficient and the preset reference pulse width, and the rise time is greater than or equal to the product of the sound velocity difference coefficient and the preset reference rise time, the board structure type of the detection area is determined to be adhesive-dominant; when the rise time is less than the product of the sound velocity difference coefficient and the preset reference rise time, and the pulse width is greater than the product of the impedance mismatch coefficient and the preset reference pulse width, the board structure type of the detection area is determined to be a bamboo-joint-adhesive-layer co-dominant type; when the rise time is greater than or equal to the product of the sound velocity difference coefficient and the preset reference rise time, and the pulse width is less than or equal to the product of the impedance mismatch coefficient and the preset reference pulse width, the board structure type of the detection area is determined to be bamboo-flesh-dominant.

[0008] Optionally, based on the plate structure type and multiple instantaneous amplitude envelopes corresponding to the multiple detection areas, multiple ultrasonic confidence levels corresponding to the multiple detection areas are determined, including: For each detection area, when the board structure type corresponding to the detection area is bamboo flesh-dominant or bamboo joint-dominant, the energy severity is determined based on the peak height corresponding to the detection area and the preset benchmark peak height. The energy severity is used to indicate the proportion of the peak height corresponding to the detection area relative to the energy tolerance limit. When the board structure type corresponding to the detection area is the adhesive layer dominant type, the pulse width severity is determined according to the pulse width corresponding to the detection area and the preset reference pulse width. The pulse width severity is used to indicate the proportion of the pulse width corresponding to the detection area relative to the pulse width tolerance limit. When the board structure type corresponding to the detection area is the bamboo joint adhesive layer type, the energy severity and pulse width severity are determined according to the peak height, preset reference peak height, pulse width and preset reference pulse width corresponding to the detection area. Based on the energy severity and / or pulse width severity corresponding to each detection region, multiple ultrasound confidence levels are determined for the multiple detection regions.

[0009] Optionally, before determining the board structure type of the detection area corresponding to the instantaneous amplitude envelope based on the characteristic parameters corresponding to the instantaneous amplitude envelope and a preset evaluation rule, the bamboo board defect detection method further includes: Based on the peak height, pulse width, and rise time corresponding to the instantaneous amplitude envelope, as well as the preset reference rise time, preset reference pulse width, and preset reference peak height, determine the peak height fluctuation value, pulse width fluctuation value, and rise time fluctuation value. Based on the fluctuation values ​​of peak height, pulse width, and rise time, as well as the corresponding fluctuation safety thresholds for peak height, pulse width, and rise time, update the preset reference rise time, preset reference pulse width, and preset reference peak height.

[0010] Optionally, based on the visual defect range, visual confidence level, multiple detection areas, and multiple ultrasonic confidence levels, the defect confidence level of the bamboo board to be inspected in multiple detection areas is determined, including: Based on the visual defect range and multiple detection areas, a first area and a second area are determined. The first area is the detection area with the largest overlap with the visual defect range, and the other detection areas are the first area. The defect confidence levels for the first region and the second region are determined based on visual and ultrasonic confidence levels.

[0011] Optionally, the defect confidence levels of the first region and the second region are determined based on visual confidence and ultrasonic confidence, including: Modal consistency is determined based on visual confidence and ultrasonic confidence, and modal consistency is used to indicate the consistency between visual and ultrasonic detection. Based on modal consistency, determine the weights for visual detection and ultrasonic detection; Based on the visual inspection weight, ultrasonic inspection weight, visual confidence level, and ultrasonic confidence level, the defect confidence level of the first region and the defect confidence level of the second region are determined.

[0012] Optionally, the image defect detection model is a YOLOv8 object detection optimization model, which includes multiple anisotropic texture suppression modules; The anisotropic texture suppression module is used to extract horizontal crack features and vertical texture features from the feature map output by the backbone network. Based on the extracted horizontal crack features and vertical texture features, the directional difference weights are determined. The feature map output by the backbone network is updated according to the directional difference weights, and the updated feature map output by the backbone network is used as the input of the neck network.

[0013] Optionally, the target loss function applied during the training of the image defect detection model is as follows: in, In the formula, Represents the target loss function. Indicates the loss of a complete intersection and union. This indicates a form perception penalty. Represents the balance coefficient. , These represent the width and height of the actual bounding box, respectively. , These represent the width and height of the prediction box, respectively. Indicates the focusing coefficient. This represents the intersection-union ratio (IoU) between the predicted bounding box and the ground truth bounding box.

[0014] Secondly, this application provides a defect detection system for bamboo-based panels, comprising: The system includes a frame, a conveying mechanism, an image acquisition module, an ultrasonic flaw detection module, a photoelectric sensor, and a central control module; the central control module is electrically connected to the image acquisition module, the ultrasonic flaw detection module, and the photoelectric sensor. The conveying mechanism is mounted on the frame and is used to transport the bamboo boards to be tested; The image acquisition module includes an industrial camera and an auxiliary light source, used to acquire target images of the bamboo board to be inspected; The ultrasonic flaw detection module includes an ultrasonic probe, which is used to acquire multiple ultrasonic echo signals in multiple detection areas of the bamboo board to be tested; The photoelectric sensor is installed at the left end of the frame and is used to send detection signals to the central control module; The central control module is used to control the image acquisition module to acquire the target image based on the detection signal, and to control the ultrasonic flaw detection module to acquire multiple ultrasonic echo signals. The central control module is also used to determine the visual defect range and visual confidence level of the bamboo board to be inspected based on the target image and the image defect detection model; to determine multiple ultrasonic confidence levels of the bamboo board to be inspected in multiple detection areas based on multiple ultrasonic echo signals, with each detection area corresponding to a specific ultrasonic echo signal; to determine the defect confidence level of the bamboo board to be inspected in multiple detection areas based on the visual defect range, visual confidence level, multiple detection areas, and multiple ultrasonic confidence levels; and to determine the defect detection result of the bamboo board to be inspected based on the defect confidence levels of multiple detection areas.

[0015] Thirdly, this application provides a defect detection device for bamboo boards, comprising: The first processing module is used to acquire the target image of the bamboo board to be inspected, and based on the target image and the image defect detection model, determine the visual defect range and visual confidence level of the bamboo board to be inspected. The second processing module is used to acquire multiple ultrasonic echo signals of the bamboo board to be tested in multiple detection areas, and to determine multiple ultrasonic confidence levels of the bamboo board to be tested in multiple detection areas based on the multiple ultrasonic echo signals. The multiple detection areas correspond one-to-one with the multiple ultrasonic echo signals. The fusion module is used to determine the defect confidence of the bamboo board to be tested in multiple detection areas based on the visual defect range, visual confidence, multiple detection areas, and multiple ultrasonic confidence. The determination module is used to determine the defect detection results of the bamboo board to be inspected based on the defect confidence levels of multiple detection areas.

[0016] Fourthly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the bamboo board defect detection method described above.

[0017] Fifthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the bamboo board defect detection method described above.

[0018] Sixthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the bamboo board defect detection method described above.

[0019] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application provides a method for detecting defects in bamboo panels. It involves image defect detection of the target image of the bamboo panel to be inspected. Based on a pre-trained image defect detection model that outputs a prediction box reflecting the location of surface defects and its confidence level, the method quickly and accurately obtains the visual defect range (i.e., the range contained in the prediction box) and visual confidence level (i.e., the confidence level of the prediction box) reflecting the surface defects of the bamboo panel. Then, it performs ultrasonic detection on the bamboo panel across multiple detection areas. Based on multiple ultrasonic echo signals from these areas, and utilizing the different echo characteristics of normal and defective parts of the bamboo panel, it quickly and accurately obtains multiple ultrasonic confidence levels reflecting the internal defects of the bamboo panel. Based on the visual defect range, visual confidence level, multiple detection areas, and multiple ultrasonic confidence levels, it determines the defect confidence level corresponding to each detection area, which integrates surface and internal defects and reflects the overall defect situation of the bamboo panel within the detection area. Finally, based on the defect confidence levels of multiple detection areas, it quickly and accurately determines the defect detection result of the bamboo panel, improving the efficiency and accuracy of bamboo panel defect detection. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 An application environment diagram for the bamboo board defect detection method provided in the embodiments of this application; Figure 2 A schematic flowchart of a method for detecting defects in bamboo boards provided in an embodiment of this application; Figure 3 This is another flowchart illustrating the method for detecting defects in bamboo boards provided in this application embodiment; Figure 4 This is a schematic diagram of the structure of the YOLOv8 target detection optimization model provided in the embodiments of this application; Figure 5 This is a schematic diagram of a bamboo board defect detection system provided in an embodiment of this application; Figure 6 The electrical control principle block diagram of the bamboo board defect detection system provided in the embodiments of this application; Figure 7 A schematic diagram of the detection logic flow of the bamboo board defect detection system provided in the embodiments of this application; Figure 8 A schematic diagram of the functional modules of the bamboo board defect detection device provided in the embodiments of this application; Figure 9 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0024] The bamboo board defect detection method provided in this application embodiment can be applied to, for example... Figure 1In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be set up independently, integrated into server 104, or placed in the cloud or on another server. Terminal 102 can send the target image and ultrasonic echo signal of the bamboo board to be inspected to server 104. After receiving the data, server 104 determines the visual defect range and visual confidence level based on the target image using an image defect detection model, determines the ultrasonic confidence level based on the ultrasonic echo signal, determines the defect confidence level based on the visual and ultrasonic confidence levels, and determines whether the bamboo board to be inspected has a defect based on the defect confidence level. Server 104 can feed back the defect detection results to terminal 102. Furthermore, in some embodiments, the bamboo board defect detection method can also be implemented independently by server 104 or terminal 102. For example, terminal 102 can directly perform defect detection on the acquired target image and ultrasonic echo signal, or server 104 can acquire the target image and ultrasonic echo signal from the data storage system and perform defect detection on the target image and ultrasonic echo signal.

[0025] The terminal 102 can be, but is not limited to, various desktop computers, laptops, smartphones, and tablets. The server 104 can be implemented using a standalone server or a server cluster consisting of multiple servers, or it can be a cloud server.

[0026] In one exemplary embodiment, such as Figure 2 As shown, a method for detecting defects in bamboo boards is provided. This method is executed by computer equipment, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the method is applied to... Figure 1 Taking server 104 as an example, the following steps S201 to S204 can be included.

[0027] S201. Obtain the target image of the bamboo board to be inspected. Based on the target image and the image defect detection model, determine the visual defect range and visual confidence level of the bamboo board to be inspected.

[0028] For example, an image of the bamboo board to be inspected can be acquired using an industrial camera with a global exposure CMOS structure and used as the target image.

[0029] When acquiring the target image, a shadowless dome light source can be placed directly above the horizontally placed bamboo board, and a low-angle strip light source can be placed at an angle of 30° to 45° below the side of the bamboo board. This incident angle range constitutes the preferred optical conditions for dark-field imaging of the bamboo surface. Within this angle range, the side-incident light can effectively reduce specular reflection glare caused by high incident angles, while suppressing surface texture shadow interference caused by grazing illumination at too small an incident angle, thereby significantly improving the signal-to-noise ratio of defect areas relative to background texture. The above two light sources can highlight features such as cracks, color differences, and stains on the surface of the bamboo board.

[0030] In addition, the brightness of the light source can be adjusted in real time through automatic exposure control to suppress surface reflection and texture interference. The automatic exposure control method can calculate the average gray value of the region of interest (ROI) in real time (the determination of the average gray value of the ROI is existing technology and will not be elaborated here), and then calculate the absolute value of the difference between the average gray value and the preset target gray value to obtain the gray value deviation value; if the gray value deviation value is greater than the preset error tolerance, the LED drive current is adjusted through the PID controller.

[0031] The adjustment formula for a PID controller is: In the formula, This represents the output value of the pulse width modulation control signal, used to control the brightness of the light source; This is the grayscale deviation value. The proportional coefficient determines the degree of brightness adjustment of the light source. It is a preset value and can be set to 0.6. This represents the integral coefficient, used to eliminate steady-state error. It is a preset value and can be set to 0.05. It is a differential coefficient used to predict future trends and avoid over-adjustment of the light source brightness. It is a preset value and can be set to 0.1.

[0032] For example, the image defect detection model can be an existing YOLOv8 object detection model. The model input is an image of the bamboo board to be detected, and the model output is a predicted bounding box reflecting the location of the defects on the surface of the bamboo board to be detected (the visual defect range is the range contained in the predicted bounding box), as well as the confidence of the predicted bounding box (i.e., visual confidence). It can be understood that the model output may also include the defect type corresponding to the predicted bounding box.

[0033] For example, the existing YOLOv8 object detection model calculates the visual confidence score of each predicted bounding box using the following formula: In the formula, Indicates visual confidence level. This represents the probability of including a defect. It should be noted that the YOLOv8 object detection model was trained using bamboo board image data with pre-labeled surface defect locations (i.e., ground truth boxes) and defect types. The specific training method is a conventional technique in this field and will not be elaborated upon here.

[0034] For example, before inputting the acquired target image into the image defect detection model, the target image can be preprocessed to reduce interference caused by bamboo texture and uneven lighting.

[0035] Preprocessing may include weighted grayscale conversion and adaptive histogram contrast enhancement: (1) Weighted grayscale conversion: A weighted grayscale conversion formula is used to convert the target image (color) into a grayscale image. The weighted grayscale conversion formula is as follows: Gray=0.299×R+0.587×G+0.114×B In the formula, Gray represents the pixel grayscale value of the grayscale image; R, G, and B are the values ​​of the pixels in the red, green, and blue channels of the target image, respectively.

[0036] This treatment preserves the details of the bamboo texture while reducing color noise interference.

[0037] (2) Adaptive histogram contrast enhancement The image is divided into 8×8 local sub-regions, and local histogram equalization is performed on each sub-region to enhance contrast.

[0038] The enhanced pixel grayscale value is calculated as follows: In the formula, The enhanced pixel grayscale value; The original grayscale value of the pixel; The maximum gray level of the sub-region; This represents the minimum gray level of the sub-region. This is the brightness compensation coefficient, used to control the overall enhancement of image brightness. It is a preset value, which can be between 0.1 and 0.3, with 0.2 being preferred. This represents the grayscale level, which is a preset value and can be 256.

[0039] The image processed by weighted grayscale and adaptive histogram contrast enhancement can suppress problems such as insufficient brightness at the bamboo edges and significantly enhance the contrast between the cracked area and the bamboo background texture.

[0040] The preprocessed image serves as input data for the image defect detection model, which is then used for subsequent defect feature extraction and recognition.

[0041] S202. Acquire multiple ultrasonic echo signals of the bamboo board to be tested in multiple detection areas, and determine multiple ultrasonic confidence levels corresponding to the bamboo board to be tested in multiple detection areas based on the multiple ultrasonic echo signals.

[0042] Among them, multiple detection areas correspond one-to-one with multiple ultrasonic echo signals.

[0043] For example, multiple ultrasonic transmitting probes can be used to transmit ultrasonic signals to the bamboo board to be tested, and corresponding ultrasonic receiving probes can be used to receive the ultrasonic echo signals; it can be understood that multiple ultrasonic transmitting probes correspond one-to-one with multiple detection areas.

[0044] For example, for each detection area, the maximum transmitted pulse amplitude in the ultrasonic echo signal can be extracted using the pulse penetration method. Then, the ratio of the maximum transmitted pulse amplitude in that detection area to a preset reference amplitude is calculated to obtain the transmittance of that detection area. Next, the difference between a preset transmittance threshold and the transmittance of that detection area is calculated to obtain the difference result. Finally, the product of the difference result and a preset sensitivity coefficient is calculated to obtain the product result. The product result is then substituted into a Sigmoid function and mapped to the [0,1] interval to obtain the ultrasonic confidence level corresponding to that detection area. It can be understood that multiple ultrasonic confidence levels corresponding to multiple detection areas can ultimately be obtained.

[0045] S203. Based on the visual defect range, visual confidence level, multiple detection areas, and multiple ultrasonic confidence levels, determine the defect confidence level of the bamboo board to be tested in multiple detection areas.

[0046] Specifically, S203 may include: Based on the visual defect range and multiple detection areas, a first region and a second region are determined. The first region is the detection area with the largest overlap with the visual defect range, and the second region consists of other detection areas. Based on visual confidence and ultrasonic confidence, the defect confidence of the first region and the defect confidence of the second region are determined.

[0047] It can be understood that the visual defect range is the range contained within the predicted bounding box of the bamboo board surface defect in the bamboo board image output by the model.

[0048] For example, the visual confidence score and the ultrasonic confidence score can be weighted and summed to obtain the defect confidence score of the first region; wherein the visual inspection weight corresponding to the visual confidence score and the ultrasonic inspection weight corresponding to the ultrasonic confidence score can both be preset values, and the sum of the two is equal to 1.

[0049] For example, the ultrasonic confidence level of the second region can be directly determined as the defect confidence level of the second region.

[0050] Further, based on visual confidence and ultrasonic confidence, the defect confidence in the first region and the defect confidence in the second region are determined, including: Modal consistency is determined based on visual and ultrasonic confidence levels, which indicate the consistency between visual and ultrasonic inspections. Visual and ultrasonic inspection weights are determined based on modal consistency. Defect confidence levels for the first and second regions are determined based on visual, ultrasonic, visual, and ultrasonic confidence levels.

[0051] For example, the absolute value of the difference between visual confidence and ultrasonic confidence can be used to reflect the consistency between visual and ultrasonic detection, i.e., modal consistency.

[0052] For example, the ultrasound detection weights can be determined by the following formula: In the formula, Indicates the weight of ultrasound detection. This indicates the maximum participation rate of ultrasound detection (preset value, such as 0.8). This indicates the minimum participation ratio of ultrasound detection (preset value, such as 0.4). This indicates modal consistency.

[0053] For example, the confidence level of a defect in any region can be determined by the following formula: In the formula, Indicates the confidence level of the defect; Indicates the confidence level of ultrasound; Indicates visual confidence level; This represents the mutual verification enhancement coefficient, which is a preset value greater than or equal to zero. It can be 0.2. When both visual and ultrasound indicate abnormalities, the confidence level is increased through the corresponding product term.

[0054] It is understandable that when there is a significant difference between visual and ultrasonic detection results, it indicates that the location may be in a structurally complex area or that a single detection modality is being interfered with. To prevent a single detection modality from dominating the final decision due to abnormal response, the ultrasonic detection weight is automatically adjusted to maintain a relative balance between visual and ultrasonic results, thereby improving the stability and robustness of the overall decision.

[0055] If only ultrasonic detection produces a significant response while visual detection fails to detect the corresponding defect target, it indicates that the defect is more likely to be an internal defect of the board material (such as voids, delamination, etc.). Therefore, the weight of ultrasonic detection results in the fusion process is automatically increased to avoid internal defects being missed because they are not visually visible.

[0056] Conversely, when visual inspection produces a significantly high confidence response, while ultrasonic inspection response is weak or close to the background level, the system determines that the defect is more likely to be a shallow crack, surface damage, or other defect type that is not sensitive to ultrasonic propagation. In this case, the system automatically suppresses the influence of ultrasonic inspection results in the fusion calculation to prevent the effective judgment of visual inspection from being diluted due to insufficient ultrasonic signal attenuation.

[0057] S204. Determine the defect detection results of the bamboo board to be inspected based on the defect confidence levels of multiple detection areas.

[0058] For example, since the quality of bamboo boards is determined by the most serious defect, the maximum risk criterion can be used to determine the final defect confidence of the bamboo board to be tested. That is, the maximum value of the defect confidence among multiple detection areas is determined as the final defect confidence of the bamboo board to be tested, and a preset defect confidence threshold is used for comparison: when the final defect confidence is greater than the preset defect confidence threshold, it is determined that the bamboo board to be tested has a serious defect, and when the final defect confidence is less than the preset defect confidence threshold, it is determined that the bamboo board to be tested does not have a serious defect.

[0059] It is understandable that when the final defect confidence level equals the preset defect confidence level threshold, it can be determined whether the bamboo board to be tested has serious defects or not, and no restrictions are imposed on this.

[0060] This embodiment of the invention performs image defect detection on a target image of a bamboo board to be inspected. Based on a pre-trained image defect detection model that outputs a prediction box reflecting the location of surface defects in the bamboo board and the confidence level of the prediction box, it quickly and accurately obtains the visual defect range (i.e., the range contained in the prediction box) and visual confidence level (i.e., the confidence level of the prediction box) reflecting the surface defects of the bamboo board. By performing ultrasonic detection on the bamboo board with multiple detection areas, and utilizing the different echo characteristics of the ultrasonic echo signals in the normal and defective parts of the bamboo board, it quickly and accurately obtains multiple ultrasonic confidence levels reflecting the internal defects of the bamboo board. Based on the visual defect range, visual confidence level, multiple detection areas, and multiple ultrasonic confidence levels, it determines the defect confidence level corresponding to each detection area, which integrates surface and internal defects and reflects the overall defect situation of the bamboo board in the detection area. Based on the defect confidence levels of multiple detection areas, it quickly and accurately determines the defect detection result of the bamboo board, improving the efficiency and accuracy of bamboo board defect detection.

[0061] Traditional ultrasonic algorithms only consider amplitude, but the internal structure of bamboo is not homogeneous like that of metal. As a laminated composite material, bamboo exhibits a ternary heterogeneous structure consisting of "bamboo matrix - bamboo joint reinforcement - adhesive layer interface." Conventional algorithms are prone to misidentifying normal adhesive layer interfaces as defects or failing to distinguish between normal glue lines and delamination. Therefore, in some possible implementations, refer to... Figure 3 Based on multiple ultrasonic echo signals, multiple ultrasonic confidence levels are determined for the bamboo board to be tested in multiple detection areas, which may include S301 to S303.

[0062] S301. Based on multiple ultrasonic echo signals, determine the multiple instantaneous amplitude envelopes corresponding to the multiple ultrasonic echo signals respectively.

[0063] Among them, multiple instantaneous amplitude envelopes correspond one-to-one with multiple detection areas.

[0064] For example, the ultrasonic echo signal acquired by the ultrasonic receiving probe Ultrasonic echo signals typically contain high-frequency oscillations, making them difficult to analyze directly. Therefore, a Hilbert transform can be performed on the original ultrasonic echo signal to extract the envelope signal that reflects energy fluctuations. The specific calculation formula is as follows: First, the Hilbert transform component of the original ultrasonic echo signal is calculated using the following formula. : In the formula, It is the integral variable.

[0065] Next, the instantaneous amplitude envelope is synthesized using the Pythagorean theorem. .

[0066] S302. Determine the plate structure type of multiple detection areas based on multiple instantaneous amplitude envelopes.

[0067] Specifically, S302 may include: For each instantaneous amplitude envelope, based on the characteristic parameters corresponding to the instantaneous amplitude envelope and according to preset evaluation rules, the material structure type of the detection area corresponding to the instantaneous amplitude envelope is determined. The characteristic parameters corresponding to the instantaneous amplitude envelope include peak height, pulse width, and rise time. The preset evaluation rules include: when the rise time is less than the product of the sound velocity difference coefficient and the preset reference rise time, and the pulse width is less than or equal to the product of the impedance mismatch coefficient and the preset reference pulse width, the material structure type of the detection area is determined to be bamboo-dominant; when the pulse width is greater than the product of the impedance mismatch coefficient and the preset reference pulse width, and the rise time is less than or equal to the product of the sound velocity difference coefficient and the preset reference pulse width, the material structure type of the detection area is determined to be bamboo-dominant. If the rise time is greater than or equal to the product of the sound velocity difference coefficient and the preset reference rise time, the board structure type of the detection area is determined to be the adhesive layer dominant type; if the rise time is less than the product of the sound velocity difference coefficient and the preset reference rise time, and the pulse width is greater than the product of the impedance mismatch coefficient and the preset reference pulse width, the board structure type of the detection area is determined to be the bamboo joint adhesive layer co-dominant type; if the rise time is greater than or equal to the product of the sound velocity difference coefficient and the preset reference rise time, and the pulse width is less than or equal to the product of the impedance mismatch coefficient and the preset reference pulse width, the board structure type of the detection area is determined to be the bamboo flesh dominant type.

[0068] After synthesizing the instantaneous amplitude envelope using S301, the chaotic oscillations of the ultrasonic echo signal are transformed into a smooth contour curve. Subsequent analysis will be based entirely on the peak height of this curve. Pulse width and rise time This process suppresses interference from high-frequency phase noise.

[0069] crest height Characterizes the ability of a medium to transmit and reflect ultrasound; when there are cavities or loose structures inside, the sound energy attenuation is significant. Significantly reduced. Pulse width. The time spread characteristics of the echo energy distribution are reflected; if there are multi-interface structures such as delamination and delamination, multiple reflections and refractions will occur, leading to... Increase. Rise time Characterizing the rate of change of acoustic impedance in a medium; highly dense, rigid structures (such as bamboo joints) cause sound waves to establish echoes rapidly, thus... Significantly reduced.

[0070] For example, determining the type of plate structure in the detection area can include: when and It was determined to be the bamboo-dominant type; among them, The preset baseline rise time, The sound speed difference coefficient (typical value 0.5~0.7); To preset the reference pulse width, The impedance mismatch factor is typically 1.2 to 1.5. The theoretical basis is that the vascular bundle density at the bamboo node is extremely high and laterally interwoven, resulting in high sound velocity. when and The type was determined to be dominated by the adhesive layer. The theoretical basis is that there is a difference in acoustic impedance between normal bamboo and the adhesive layer. The normal adhesive interface will produce a slight interface reflection wave, which causes the echo signal to broaden in the time domain.

[0071] when and At that time, it can be determined that the bamboo joint adhesive layer is the dominant type.

[0072] when and At that time, it was determined to be the bamboo flesh-dominant type.

[0073] S303. Based on the plate structure type and instantaneous amplitude envelope corresponding to the multiple detection areas, determine the multiple ultrasonic confidence levels corresponding to the multiple detection areas.

[0074] Specifically, S303 may include: For each detection area, when the corresponding board structure type is bamboo flesh-dominant or bamboo joint-dominant, the energy severity is determined based on the peak height of the detection area and a preset reference peak height. The energy severity indicates the proportion of the peak height of the detection area relative to the energy tolerance limit. When the corresponding board structure type is adhesive-dominant, the pulse width severity is determined based on the pulse width of the detection area and a preset reference pulse width. The pulse width severity indicates the proportion of the pulse width of the detection area relative to the pulse width tolerance limit. When the corresponding board structure type is a bamboo joint and adhesive-dominant combination, the energy severity and pulse width severity are determined based on the peak height of the detection area, the preset reference peak height, the pulse width, and the preset reference pulse width. Based on the energy severity and / or pulse width severity of each detection area, multiple ultrasonic confidence levels are determined for the multiple detection areas.

[0075] For example, for each detection area, the energy severity can be determined by the following formula: In the formula, Indicates the severity of energy levels. Indicates the preset reference peak height. Indicates the height of the wave crest. This refers to the energy tolerance coefficient corresponding to whether the board structure is dominated by bamboo flesh, bamboo nodes, or a combination of bamboo nodes and adhesive layers. When the detection area is dominated by bamboo flesh, the energy attenuation during ultrasonic propagation is relatively small due to the relatively uniform structure of the bamboo flesh. The value can be taken as 0.9~1.1; when the detection area is dominated by bamboo nodes or bamboo nodes and adhesive layers, due to the dense and complex fibrous tissue in the bamboo node area, ultrasonic scattering and energy attenuation are easily caused, and the energy tolerance coefficient is... A value of 1.2 to 1.4 is acceptable. It is understood that the energy tolerance coefficients corresponding to the bamboo-dominant type or the bamboo-joint adhesive layer co-dominant type of the board structure can be the same or different, and there are no restrictions on this.

[0076] For example, for each detection region, the pulse width severity can be determined by the following formula: In the formula, Indicates the severity of the pulse. Indicates the preset reference pulse width. Indicates the pulse width. This represents the pulse tolerance coefficient corresponding to either the adhesive-dominated type or the bamboo-joint adhesive-dominated type of the board structure, ranging from 1.1 to 1.3. It is understood that the pulse tolerance coefficients corresponding to the adhesive-dominated type or the bamboo-joint adhesive-dominated type of the board structure can be the same or different; there is no restriction on this.

[0077] For example, for each detection area, the ultrasonic confidence level can be determined by the following formula: In the formula, Indicates the confidence level of ultrasound. This represents the sensitivity adjustment coefficient, ranging from 2 to 5. Indicates the final severity.

[0078] For example, when the board structure type corresponding to the detection area is bamboo flesh-dominant or bamboo joint-dominant, the energy severity can be determined as the final severity; when the board structure type corresponding to the detection area is adhesive layer-dominant, the pulse width severity can be determined as the final severity; when the board structure type is bamboo flesh-adhesive layer-dominant, the larger value of energy severity and pulse width severity can be determined as the final severity.

[0079] It's understandable, when , Approaching 0; when , Approaching 1.

[0080] Thus, the ultrasonic confidence level corresponding to the detection area can be accurately determined by using the energy severity and pulse severity, which reflect the proportion of the current deviation relative to the corresponding process tolerance limit.

[0081] Furthermore, before determining the board structure type of the detection area corresponding to the instantaneous amplitude envelope based on the characteristic parameters corresponding to the instantaneous amplitude envelope and according to preset evaluation rules, the bamboo board defect detection method may also include: Based on the peak height, pulse width, and rise time corresponding to the instantaneous amplitude envelope, as well as the preset reference rise time, preset reference pulse width, and preset reference peak height, determine the peak height fluctuation value, pulse width fluctuation value, and rise time fluctuation value. Based on the fluctuation values ​​of peak height, pulse width, and rise time, as well as the corresponding fluctuation safety thresholds for peak height, pulse width, and rise time, update the preset reference rise time, preset reference pulse width, and preset reference peak height.

[0082] For example, the peak height fluctuation value can be obtained by calculating the ratio of the difference between the peak height and the preset reference peak height to the preset reference peak height; the pulse width fluctuation value can be obtained by calculating the ratio of the difference between the pulse width and the preset reference pulse width to the preset reference pulse width; and the rise time fluctuation value can be obtained by calculating the ratio of the difference between the rise time and the preset reference rise time to the preset reference rise time.

[0083] For example, the fluctuation values ​​corresponding to the characteristic parameters (including peak height fluctuation values, pulse width fluctuation values, and rise time fluctuation values) can be determined by the following formula: In the formula, Representing characteristic parameters The corresponding fluctuation value; Represents characteristic parameters, ; Representing characteristic parameters The corresponding preset baseline value.

[0084] Understandable, peak height The corresponding preset reference value, i.e., the preset reference peak height Pulse width The corresponding preset reference value, i.e., the preset reference pulse width Ascent time The corresponding preset baseline value, i.e., the preset baseline rise time. .

[0085] For example, a preset benchmark value can be updated only when the fluctuation value does not meet the fluctuation safety threshold to avoid introducing defective data. For instance, only when... (in Representing characteristic parameters Only when the corresponding fluctuation safety threshold (which is a preset value) can the preset benchmark value be updated.

[0086] For example, the characteristic parameters can be expressed by the following formula. The corresponding preset benchmark values ​​are updated to accommodate the natural differences in moisture content, density, and other factors among different batches of bamboo boards. In the formula, Representing characteristic parameters The corresponding preset baseline value; This represents the update coefficient, which is a preset value (e.g., 0.02) used to limit the magnitude of a single update.

[0087] In this way, by setting a fluctuation safety threshold, the update constraint of the preset benchmark value can be realized, so as to avoid introducing defective data and improve the accuracy of the preset benchmark value.

[0088] In some possible implementations, the image defect detection model is a YOLOv8 object detection optimization model, which includes multiple anisotropic texture suppression modules; The anisotropic texture suppression module is used to extract horizontal crack features and vertical texture features from the feature map output by the backbone network. Based on the extracted horizontal crack features and vertical texture features, the directional difference weights are determined. The feature map output by the backbone network is updated according to the directional difference weights, and the updated feature map output by the backbone network is used as the input of the neck network.

[0089] The existing YOLOv8 object detection model includes a backbone network, a neck network, and a detection head. The backbone network is used to extract multi-layer features from the input image to be detected; the neck network is used to fuse, enhance, and scale-align the multi-layer features extracted by the backbone network, outputting high-quality feature maps suitable for detecting targets of different sizes to the detection head; the detection head is used to perform object detection on the feature maps of the three scales output by the neck network.

[0090] Because the fibers on the surface of bamboo boards exhibit a highly consistent longitudinal distribution, this regular background texture can easily obscure defect information such as tiny cracks during feature extraction. Addressing the two core challenges of bamboo surface defect detection—"strong background texture interference" and "inaccurate localization of fine cracks"—this paper proposes targeted architectural improvements to the existing YOLOv8 target detection model to reduce the interference of background texture on defect identification. This improvement leverages the characteristic that cracks or defects are typically horizontal or oblique.

[0091] like Figure 4As shown, this embodiment introduces an anisotropic texture suppression module based on directional feature differences for texture suppression processing between the backbone network and the neck network of the existing YOLOv8 object detection model, thus obtaining an optimized YOLOv8 object detection model.

[0092] The specific processing flow of the anisotropic texture suppression module includes: Let the feature map of a certain layer output by the backbone network be... .

[0093] (a) To separate background texture from defect features, asymmetric convolution kernels are used to extract gradient responses in the horizontal and vertical directions, respectively: In the formula, Indicates horizontal fracture characteristics; Represents vertical texture features; This represents a gradient convolution operator along the horizontal direction, used to extract horizontal crack and edge features; This represents a gradient convolution operator along the vertical direction, used to extract longitudinal fiber texture features; The scale value of the convolution kernel; This represents the ReLU activation function.

[0094] (b) Calculate the directional difference weights: In the formula, Directional difference weight, with a numerical range of [0,1]. This represents a minimal constant to prevent the denominator from being zero, and is taken as 10. -6 Up to 10 -8 .

[0095] It can be seen that in areas with strong background, Significantly greater than Both the numerator and denominator are relatively large. It tends towards a specific value; in defective areas, Significantly greater than This generates extremely strong directional difference signals.

[0096] (c) Texture suppression feature output: In the formula, This represents the output feature map of the anisotropic texture suppression module, which is also the feature map output by the updated backbone network. The suppression intensity coefficient is a preset value (e.g., 0.8).

[0097] Based on the above formula, Regions with larger values ​​(i.e., strong background textures) are significantly suppressed, while Regions with smaller values ​​(i.e. potential defects) are preserved, thereby increasing the salience of anomalous structures in the feature map.

[0098] Furthermore, the target loss function applied during the training process of the image defect detection model is shown in the following equation: in, In the formula, Represents the target loss function. Indicates the loss of a complete intersection and union. This indicates a form perception penalty. This represents the balance coefficient, which is set to 0.5. , These represent the width and height of the actual bounding box, respectively. , These represent the width and height of the prediction box, respectively. This represents the focusing factor, which is set to 3. This represents the intersection-union ratio (IoU) between the predicted bounding box and the ground truth bounding box.

[0099] It is understandable that the true bounding boxes are pre-annotated in the bamboo board image during the training process to reflect the actual location of surface defects.

[0100] Complete intersection and union loss The calculation formula is shown below: In the formula, Represents the coordinates of the center point of the prediction box. Coordinates of the center point of the real bounding box The square of the Euclidean distance between them The diagonal length of the minimum bounding rectangle covering the predicted bounding box and the ground truth bounding box. As a penalty term for aspect ratio, This indicates the difference in aspect ratio. This indicates the aspect ratio penalty weight.

[0101] Given that bamboo cracks generally exhibit a slender shape and significant differences in aspect ratio, traditional IoU-type bounding box regression loss functions (i.e., complete intersection-union loss) are not suitable for this type of defect. When the overlap between the predicted and ground truth bounding boxes is small or there are subtle deviations in shape, the gradient feedback effect is limited, making it difficult to stably constrain the boundary shape of slender defects. This embodiment introduces an additional constraint term based on the defect shape characteristics (i.e., a shape-aware penalty) on top of the existing complete intersection-union loss. This forms a morphology-aware composite loss calculation method, which is used to drive the bounding box regression training of the detection model.

[0102] Using complete intersection-union ratio loss As a basic localization term, it is used to simultaneously constrain the overlap, center point distance, and aspect ratio consistency between the predicted bounding box and the ground truth bounding box; a shape-aware penalty is employed. Further improve the fitting accuracy for extremely fine and long crack-like defects.

[0103] Specifically, when the predicted bounding box exhibits a shape difference of "too wide" or "too short," even if its center point position is accurately predicted, this term will still generate a large quadratic error gradient. In this embodiment, the shape-aware penalty term drives the network to adjust the side length parameter of the predicted bounding box during backpropagation, causing its geometry to converge towards a "slender" feature, thereby more closely matching the actual shape of the crack.

[0104] This application also provides a defect detection system for bamboo boards. Figure 5 This is a schematic diagram of a bamboo board defect detection system, including a frame 501, a conveying mechanism 502, an image acquisition module 503, an ultrasonic flaw detection module 504, a photoelectric sensor 505, and a central control module 506; the central control module 506 is electrically connected to the image acquisition module 503, the ultrasonic flaw detection module 504, and the photoelectric sensor 505.

[0105] The conveying mechanism 502 is mounted on the frame 501 and is used to convey the bamboo board to be tested. Image acquisition module 503 includes an industrial camera and an auxiliary light source (which may include a shadowless dome light source and a low-angle strip light source) for acquiring target images of the bamboo board to be inspected; The ultrasonic flaw detection module 504 includes an ultrasonic probe for acquiring multiple ultrasonic echo signals of the bamboo board to be inspected in multiple detection areas; The photoelectric sensor 505 is installed at the left end of the frame 501 and is used to send detection signals to the central control module 506; The central control module 506 is used to control the image acquisition module 503 to acquire the target image according to the detection signal, and to control the ultrasonic flaw detection module 504 to acquire multiple ultrasonic echo signals. The central control module 506 is also used to determine the visual defect range and visual confidence level of the bamboo board to be inspected based on the target image and the image defect detection model; to determine multiple ultrasonic confidence levels of the bamboo board to be inspected in multiple detection areas based on multiple ultrasonic echo signals, with each detection area corresponding to one of the multiple ultrasonic echo signals; to determine the defect confidence level of the bamboo board to be inspected in multiple detection areas based on the visual defect range, visual confidence level, multiple detection areas, and multiple ultrasonic confidence levels; and to determine the defect detection result of the bamboo board to be inspected based on the defect confidence level of the multiple detection areas (i.e., to execute the aforementioned bamboo board defect detection method).

[0106] It can be understood that the central control module 506 in the bamboo board defect detection system is the main body that executes the aforementioned bamboo board defect detection method. Specifically, it can be a terminal (such as an industrial computer or PLC) or a server that executes the aforementioned bamboo board defect detection method, and there are no restrictions on this.

[0107] For example, the ultrasonic flaw detection module 504 can adopt a roller-type coupling probe structure, including multiple transmitting probes and multiple receiving probes. The outer layers of the multiple transmitting probes and multiple receiving probes are all covered with elastic polyester to improve coupling effect and reduce energy loss caused by the roughness of the bamboo surface. The transmitting and receiving probes are arranged laterally symmetrically above the conveyor belt to cover the lateral width of the bamboo, enabling detection of multiple detection areas in the bamboo board to be inspected. The signal enters the interior of the bamboo through the probes, and defects such as internal voids, delamination, and cracks are detected through pulse penetration. The ultrasonic flaw detection module 504 may also include a signal amplifier and a filtering circuit for bandpass filtering and amplification of the ultrasonic echo signal, and transmission of the processed ultrasonic echo signal to the central control module.

[0108] For example, when the photoelectric sensor 505 detects that the bamboo board has reached the shooting position of the image acquisition module 503, it sends a detection signal to the central control module 506.

[0109] For example, a rotary encoder may be installed on the drive wheel shaft of the conveyor mechanism 502 to obtain the conveyor linear speed of the conveyor mechanism 502 in real time and output a pulse signal to the central control module 506.

[0110] Because the image acquisition module 503 and the ultrasonic flaw detection module 504 are spaced at a fixed distance. To ensure spatial synchronization between image detection and ultrasonic detection, when the photoelectric sensor 505 detects the bamboo board, the central control module 506 controls the image acquisition module to take a picture, and the central control module 506 begins to record the number of pulse signals sent from the rotary encoder.

[0111] When the pulse count reaches At the same time, the central control module 506 controls the ultrasonic flaw detection module 504 to acquire multiple ultrasonic echo signals, thereby achieving a correspondence between the image acquisition position and the ultrasonic detection position at the same physical location on the bamboo board, ensuring the spatial consistency of multi-source detection data. The displacement represented by a single pulse.

[0112] In some possible embodiments, the bamboo board defect detection system may also include a pneumatic actuator 507, and the central control module 506 is also electrically connected to the pneumatic actuator 507. The central control module 506 is also used to send a rejection signal to the pneumatic push rod 507 when the defect detection result of the bamboo board to be inspected is that there is a serious defect. A pneumatic push rod 507 is installed at the right end of the frame 501 and is used to move the push rod according to the rejection signal to push off and reject bamboo boards with serious defects on the conveying mechanism 502.

[0113] For example, when the defect detection result of the bamboo board to be inspected is that there is a serious defect, the central control module 506 sends a rejection signal to the pneumatic push rod 507. The central control module 506, combined with the pulse value counted by the encoder, calculates the precise delay time for the defective board to reach the rejection station, thus achieving accurate rejection. Precise delay time. The calculation formula is as follows: in, This refers to the physical distance from the photoelectric sensor to the centerline of the pneumatic actuator. The linear velocity of the conveyor mechanism is measured by a rotary encoder. This refers to the inherent processing time of the central control module from completing the fusion calculation to issuing the control signal.

[0114] For example, the bamboo board defect detection system may also include an audible and visual alarm. When the central control module 506 determines that there is a serious defect in the bamboo board to be inspected, it sends an alarm signal to the audible and visual alarm, which then sounds an alarm.

[0115] For example, the bamboo board defect detection system may also include a human-machine interface for displaying system data to operators.

[0116] For example, Figure 6 The electrical control principle block diagram of the bamboo board defect detection system provided in the embodiments of this application is shown. Figure 7 This is a schematic diagram of the detection logic flow of the bamboo board defect detection system provided in the embodiments of this application.

[0117] Based on the same inventive concept, this application also provides a bamboo board defect detection device for implementing the bamboo board defect detection method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations of one or more bamboo board defect detection device embodiments provided below can be found in the limitations of the bamboo board defect detection method described above, and will not be repeated here.

[0118] In one exemplary embodiment, such as Figure 8 As shown, a defect detection device for bamboo boards is provided, comprising: The first processing module 801 is used to acquire the target image of the bamboo board to be detected, and based on the target image and the image defect detection model, determine the visual defect range and visual confidence level of the bamboo board to be detected. The second processing module 802 is used to acquire multiple ultrasonic echo signals of the bamboo board to be tested in multiple detection areas, and determine multiple ultrasonic confidence levels of the bamboo board to be tested in multiple detection areas based on the multiple ultrasonic echo signals, with each detection area corresponding to a single ultrasonic echo signal. The fusion module 803 is used to determine the defect confidence of the bamboo board to be tested in multiple detection areas based on the visual defect range, visual confidence, multiple detection areas, and multiple ultrasonic confidence. The determination module 804 is used to determine the defect detection results of the bamboo board to be inspected based on the defect confidence levels of multiple detection areas.

[0119] The specific implementation methods and beneficial effects of this device embodiment can be found in the foregoing method embodiments, and will not be repeated here.

[0120] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 9As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network. When the computer program is executed by the processor, it implements the aforementioned method for detecting defects in bamboo panels.

[0121] Those skilled in the art will understand that Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0122] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0123] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0124] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0125] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0126] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0127] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0128] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0129] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for detecting defects in bamboo-based panels, characterized in that, The method for detecting defects in bamboo boards includes: Acquire a target image of the bamboo board to be inspected, and based on the target image, determine the visual defect range and visual confidence level of the bamboo board to be inspected using an image defect detection model. Multiple ultrasonic echo signals of the bamboo board to be tested in multiple detection areas are acquired. Based on the multiple ultrasonic echo signals, multiple ultrasonic confidence levels of the bamboo board to be tested in multiple detection areas are determined. The multiple detection areas correspond one-to-one with the multiple ultrasonic echo signals. Based on the visual defect range, the visual confidence level, the multiple detection areas, and the multiple ultrasonic confidence levels, the defect confidence level of the bamboo board to be tested in the multiple detection areas is determined; The defect detection results of the bamboo board to be tested are determined based on the defect confidence levels of the multiple detection areas.

2. The method for detecting defects of a bamboo board according to claim 1, wherein, The step of determining multiple ultrasonic confidence levels for the bamboo board to be tested in multiple detection areas based on the multiple ultrasonic echo signals includes: Based on the multiple ultrasonic echo signals, multiple instantaneous amplitude envelopes corresponding to the multiple ultrasonic echo signals are determined, and the multiple instantaneous amplitude envelopes correspond one-to-one with the multiple detection areas; The plate structure type of the multiple detection areas is determined based on the multiple instantaneous amplitude envelopes; Based on the plate structure type corresponding to the multiple detection areas and the multiple instantaneous amplitude envelopes, multiple ultrasonic confidence levels corresponding to the multiple detection areas are determined.

3. The method for detecting defects of a bamboo board according to claim 2, wherein, The step of determining the plate structure type of the multiple detection regions based on the multiple instantaneous amplitude envelopes includes: For each instantaneous amplitude envelope, based on the characteristic parameters corresponding to the instantaneous amplitude envelope and the preset evaluation rules, the plate structure type of the detection area corresponding to the instantaneous amplitude envelope is determined. The characteristic parameters corresponding to the instantaneous amplitude envelope include peak height, pulse width and rise time. The preset evaluation rules include: when the rise time is less than the product of the sound velocity difference coefficient and the preset reference rise time, and the pulse width is less than or equal to the product of the impedance mismatch coefficient and the preset reference pulse width, the board structure type of the detection area is determined to be bamboo-dominant; when the pulse width is greater than the product of the impedance mismatch coefficient and the preset reference pulse width, and the rise time is greater than or equal to the product of the sound velocity difference coefficient and the preset reference rise time, the board structure type of the detection area is determined to be adhesive-dominant; when the rise time is less than the product of the sound velocity difference coefficient and the preset reference rise time, and the pulse width is greater than the product of the impedance mismatch coefficient and the preset reference pulse width, the board structure type of the detection area is determined to be a bamboo-joint-adhesive-layer co-dominant type; when the rise time is greater than or equal to the product of the sound velocity difference coefficient and the preset reference rise time, and the pulse width is less than or equal to the product of the impedance mismatch coefficient and the preset reference pulse width, the board structure type of the detection area is determined to be bamboo-flesh-dominant.

4. The method for detecting defects of a bamboo board according to claim 3, wherein, The step of determining multiple ultrasonic confidence levels corresponding to the multiple detection areas based on the plate structure type corresponding to the multiple detection areas and the multiple instantaneous amplitude envelopes includes: For each detection area, when the board structure type corresponding to the detection area is bamboo flesh-dominant or bamboo joint-dominant, the energy severity is determined based on the peak height corresponding to the detection area and the preset benchmark peak height. The energy severity is used to indicate the proportion of the peak height corresponding to the detection area relative to the energy tolerance limit. When the board structure type corresponding to the detection area is the adhesive layer dominant type, the pulse width severity is determined according to the pulse width corresponding to the detection area and the preset reference pulse width. The pulse width severity is used to indicate the proportion of the pulse width corresponding to the detection area relative to the pulse width tolerance limit. When the board structure type corresponding to the detection area is the bamboo joint adhesive layer type, the energy severity and pulse width severity are determined according to the peak height, preset reference peak height, pulse width and preset reference pulse width corresponding to the detection area. Based on the energy severity and / or pulse width severity corresponding to each detection region, multiple ultrasound confidence levels are determined for the multiple detection regions.

5. The method for detecting defects in bamboo boards according to claim 4, characterized in that, Before determining the board structure type of the detection area corresponding to the instantaneous amplitude envelope based on the feature parameters corresponding to the instantaneous amplitude envelope and according to a preset evaluation rule, the bamboo board defect detection method further includes: Based on the peak height, pulse width, and rise time corresponding to the instantaneous amplitude envelope, as well as the preset reference rise time, the preset reference pulse width, and the preset reference peak height, determine the peak height fluctuation value, pulse width fluctuation value, and rise time fluctuation value. Based on the peak height fluctuation value, pulse width fluctuation value, rise time fluctuation value, and the fluctuation safety thresholds corresponding to peak height, pulse width, and rise time, update the preset reference rise time, the preset reference pulse width, and the preset reference peak height.

6. The method for detecting defects in bamboo boards according to claim 1, characterized in that, The step of determining the defect confidence level of the bamboo board to be tested in multiple detection areas based on the visual defect range, the visual confidence level, the multiple detection areas, and the multiple ultrasonic confidence levels includes: Based on the visual defect range and the multiple detection areas, a first area and a second area are determined, wherein the first area is the detection area with the largest overlap with the visual defect range, and the second area is the other detection areas; The defect confidence levels for the first region and the second region are determined based on visual and ultrasonic confidence levels.

7. The method for detecting defects in bamboo boards according to claim 6, characterized in that, Determining the defect confidence levels of the first region and the second region based on visual and ultrasonic confidence levels includes: Modal consistency is determined based on visual confidence and ultrasonic confidence, wherein the modal consistency is used to indicate the consistency between visual and ultrasonic detection. Based on modal consistency, determine the weights for visual detection and ultrasonic detection; The defect confidence levels of the first region and the second region are determined based on the visual inspection weight, ultrasonic inspection weight, visual confidence level, and ultrasonic confidence level.

8. The method for detecting defects in bamboo boards according to claim 1, characterized in that, The image defect detection model is a YOLOv8 target detection optimization model, which includes multiple anisotropic texture suppression modules. The anisotropic texture suppression module is used to extract horizontal crack features and vertical texture features from the feature map output by the backbone network. Based on the extracted horizontal crack features and vertical texture features, the module determines the directional difference weights and updates the feature map output by the backbone network according to the directional difference weights. The updated feature map output by the backbone network is then used as the input to the neck network.

9. The method for detecting defects in bamboo boards according to claim 8, characterized in that, The target loss function used in the training process of the image defect detection model is shown in the following formula: in, In the formula, Represents the target loss function. Indicates the loss of a complete intersection and union. This indicates a form perception penalty. Represents the balance coefficient. , These represent the width and height of the actual bounding box, respectively. , These represent the width and height of the prediction box, respectively. Indicates the focusing coefficient. This represents the intersection-union ratio (IoU) between the predicted bounding box and the ground truth bounding box.

10. A defect detection system for bamboo-based panels, characterized in that, The bamboo board defect detection system includes: The system includes a frame, a conveying mechanism, an image acquisition module, an ultrasonic flaw detection module, a photoelectric sensor, and a central control module; the central control module is electrically connected to the image acquisition module, the ultrasonic flaw detection module, and the photoelectric sensor. The conveying mechanism is mounted on the frame and is used to transport the bamboo boards to be tested; The image acquisition module includes an industrial camera and an auxiliary light source, used to acquire target images of the bamboo board to be inspected; The ultrasonic flaw detection module includes an ultrasonic probe, which is used to acquire multiple ultrasonic echo signals in multiple detection areas of the bamboo board to be tested; The photoelectric sensor is installed at the left end of the frame and is used to send detection signals to the central control module; The central control module is used to control the image acquisition module to acquire the target image based on the detection signal, and to control the ultrasonic flaw detection module to acquire multiple ultrasonic echo signals. The central control module is also used to determine the visual defect range and visual confidence level of the bamboo board to be inspected based on the target image and the image defect detection model; to determine multiple ultrasonic confidence levels of the bamboo board to be inspected in multiple detection areas based on multiple ultrasonic echo signals, with each detection area corresponding to a specific ultrasonic echo signal; to determine the defect confidence level of the bamboo board to be inspected in multiple detection areas based on the visual defect range, visual confidence level, multiple detection areas, and multiple ultrasonic confidence levels; and to determine the defect detection result of the bamboo board to be inspected based on the defect confidence levels of multiple detection areas.

Images