A method for online prediction of equivalent elastic modulus of bamboo splints for a bamboo splint lengthening system

Abstract

The application provides a bamboo splint equivalent elastic modulus online prediction method for a bamboo splint lengthening system, and specifically comprises the following steps: step one, a front-end double-view angle defect detection unit acquires upper surface images and side surface images of bamboo splints before lengthening, a double-view angle detection unit acquires upper surface images and side surface images of the lengthened bamboo splint segments and joints after gluing, and a width detection assembly acquires upper surface images of the bamboo splint segments and joints; the obtained images are processed to obtain image feature vectors; step two, original mechanical detection data detected by an independent modulus detection module are acquired, and the original mechanical detection data are processed to obtain mechanical response feature vectors; step three, lengthening process parameter vectors of each bamboo splint segment or joint are acquired; step four, the obtained image feature vectors, mechanical response feature vectors and lengthening process parameter vectors are standardized to obtain fused feature vectors; and step five, the fused feature vectors are input into a modulus prediction model to predict the equivalent elastic modulus of the lengthened bamboo splints.

Description

Technical Field

[0001] This invention relates to the field of bamboo strip lengthening, and more particularly to an online method for predicting the equivalent elastic modulus of bamboo strips in a bamboo strip splicing system. Background Technology

[0002] Bamboo strips are an important basic raw material for bamboo furniture, bamboo woven products, bamboo-based composite materials, and bamboo structural components. Due to the limited length of a single bamboo strip, and the fact that natural bamboo is prone to defects such as cracks, insect holes, mold, abnormal nodes, missing edges, and uneven thickness during growth, transportation, drying, and mechanical processing, shorter bamboo strips usually need to be continuously spliced ​​to meet the requirements of continuous production and long-size applications.

[0003] The applicant filed a utility model patent application with application number 20251117168X on October 28, 2025. The patent describes the specific structure of the bamboo strip splicing system and explains how to achieve splicing, but does not mention the detection of relevant data during the splicing process, making it impossible to obtain the quality of the spliced ​​bamboo strips in a timely manner. Summary of the Invention

[0004] The main objective of this invention is to provide an online prediction method for the equivalent elastic modulus of bamboo strips in a bamboo strip splicing system, in order to solve the above-mentioned technical problems.

[0005] To achieve the above objectives, the technical solution adopted by this invention is as follows: an online prediction method for the equivalent elastic modulus of bamboo strips in a bamboo strip splicing system. The bamboo strip splicing system includes, along the direction of bamboo strip movement, a front-end dual-view defect detection unit, a cutting, gluing and compaction unit, a post-gluing dual-view detection unit, a width-fixing component, a width detection component, an independent modulus detection module, and a position synchronization module, arranged sequentially at the front end. The method specifically includes the following steps: Step 1: The front-end dual-view defect detection unit acquires the top surface and side images of the bamboo strips before splicing. After gluing, the dual-view detection unit acquires the top surface and side images of the spliced ​​bamboo strip segments and joints. The width-fixed detection component acquires the top surface images of the bamboo strip segments and the machine head. The obtained images are processed to obtain image feature vectors. ; Step 2: Obtain the raw mechanical test data from the independent modulus testing module and process the raw mechanical test data to obtain the mechanical response feature vector. ; Step 3: Obtain the splicing process parameter vector for each bamboo strip segment or joint. ; Step 4: Obtain the image feature vector Mechanical response eigenvectors and extension process parameter vector The fused feature vector is obtained by standardization. ; Step 5: Fuse the feature vectors The equivalent elastic modulus of the long bamboo strips is predicted by inputting it into the modulus prediction model.

[0006] Preferably, step one specifically includes the following steps: Step 11: The front-end dual-view defect detection unit acquires the upper surface image and side image of the bamboo strip before splicing. After gluing, the dual-view detection unit acquires the upper surface image and side image of the spliced ​​bamboo strip segment and joint. The obtained images are then matched with the corresponding bamboo strip segment or joint image using the position synchronization module. Step 12: Input all images corresponding to the k-th bamboo strip segment or joint obtained in Step 11 into the YOLOv8-Seg instance segmentation model to obtain the defect tuple: , (1) in, The nth defect tuple output by the YOLOv8-Seg instance segmentation model. Let n be the defect type of the nth defect. Let n be the detection box for the nth defect. For the segmentation mask of the nth defect, Let n be the confidence level of the nth defect; Step 13: Obtain the image feature vector for each bamboo strip segment or joint based on the obtained defect tuple. : , (2) in, Let k be the image feature vector of the k-th bamboo strip segment or joint. This represents the sum of the defect areas of all defects within the k-th bamboo strip segment or joint. This is the sum of the crack lengths of all defects within the k-th bamboo strip segment or joint. This represents the maximum edge breakage depth of all defects within the k-th bamboo strip segment or joint. The maximum fiber warping height is the highest among all defects within the k-th bamboo strip segment or joint. Let be the coverage rate of the adhesive layer within the k-th bamboo strip segment or joint. Let be the glue overflow rate in the k-th bamboo strip segment or joint. This represents the overlap length of the k-th bamboo strip segment or joint. In practice, only the joint has an overlap length, but the number k is assigned to both the bamboo strip segment and the joint as a whole. Let be the misalignment amount of the k-th bamboo strip segment or joint. Let be the joint deviation angle of the k-th bamboo strip segment or joint. Let represent the change in joint thickness at the k-th bamboo strip segment or joint. Let be the width of the bamboo strip at the k-th bamboo segment or joint. Let be the width fluctuation of the k-th bamboo strip segment or joint. The area of ​​the edge burrs of the k-th bamboo strip segment or joint.

[0007] Preferably, step two specifically includes the following steps: Step 21: Obtain the loading force collected by the force sensor through the independent modulus detection module. The displacement or deflection of the bamboo strips or joints is collected by a laser displacement sensor. Vertical displacement of the loading mechanism acquired by the magnetic ruler The bamboo strip conveying position is fed back by the servo motor or servo driver in the independent modulus detection module. and conveying speed ; Step 22: Use a position synchronization module to apply force Displacement or deflection Vertical displacement Bamboo strip conveying location and conveying speed Match it with the corresponding bamboo strips or joints; Step 23, for the first Each bamboo strip or joint, based on the moment it reaches the mechanical testing area. Based on the baseline, extract the time window. Within the stated time window, the original mechanical test data is obtained. : (3), in, ; Step 24: Process the obtained raw mechanical test data Feature extraction is performed to obtain the mechanical response feature vector. : (4), in, This represents the maximum loading force of the k-th bamboo strip segment or joint within the time window. The average loading force of the k-th bamboo segment or joint within the time window. The fluctuation of the loading force of the k-th bamboo strip segment or joint within the time window is obtained by subtracting the minimum loading force from the maximum loading force. This refers to the maximum displacement or maximum deflection of the k-th bamboo segment or joint within the time window. This refers to the change in displacement or deflection. Force-displacement slope The average conveying speed of the k-th bamboo strip segment or joint within the time window. The energy response per unit length of the k-th bamboo segment or joint within the time window.

[0008] Preferably, step three involves extending the process parameter vector. It is expressed as follows: (5), in, This refers to the actual servo transmission speed. For servo acceleration or deceleration parameters, For servo-controlled fixed-length conveying, This refers to the amount of adhesive applied. For crimping time, The width of the bamboo strips, For the thickness of the bamboo strips, For the first The overlap length of each connector.

[0009] Preferably, step four specifically includes the following steps: Step 41: Process the image feature vector Mechanical response eigenvectors and extension process parameter vector Standardization is performed to obtain the standardized image feature vector. Mechanical response eigenvectors and extension process parameter vector ; Step 42: Convert the standardized image feature vector Mechanical response eigenvectors and extension process parameter vector Perform feature fusion to obtain a fused feature vector. : (6), in, , where are weighting coefficients, used to adjust the weights of image features, mechanical response features, and process parameters in modulus prediction, respectively. .

[0010] Preferably, in step five, the expression for the modulus prediction model is as follows: (7), in, For the first The coordinates of each bamboo strip segment or joint are corrected. Predicted equivalent elastic modulus at the location, For the first A tree of return, This represents the number of regression trees.

[0011] Preferably, the method further includes step six: determining the grade of the predicted value of the equivalent elastic modulus according to a preset elastic modulus threshold. (8), in, For the first The modulus grade of each bamboo strip segment or joint. and This is a preset modulus threshold.

[0012] Preferably, in step eight, when the modulus grade of multiple consecutive joints decreases, the process parameters are adjusted to compensate for the decrease. (9), in, This is the process parameter vector corresponding to the next joint. This is the current process parameter vector. For the compensation coefficient vector, For the target equivalent elastic modulus, This is the current predicted equivalent elastic modulus.

[0013] Compared with the prior art, the present invention has the following beneficial effects: This invention can predict the elastic modulus of bamboo strips after splicing based on the detected data, and takes into account multiple parameters in the prediction process, which can make the prediction more accurate. Attached Figure Description

[0014] Figure 1 It is a 3D diagram of the bamboo strip extension system; Figure 2 This is a magnified view of point A; Figure 3 This is a magnified view of point B; Figure 4 and Figure 5 These are 3D images of the independent modulus testing module from different angles; Figure 6 This is a cross-sectional view of the independent modulus testing module; Figure 7 This is a flowchart of the prediction method. Detailed Implementation

[0015] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0016] The bamboo strip lengthening system includes, along the direction of movement of the bamboo strip 10, a front-end dual-view defect detection unit 1100, a first conveying component 201, a cutting, gluing and compaction unit 300, a post-gluing dual-view detection unit 1200, a second conveying component 202, a sanding component 600, a third conveying component 203, a width fixing component 700, a width detection component 1300, a tension component 800, a winding component 900, and an independent modulus detection module 1400, all arranged sequentially at the front end.

[0017] The front-end dual-view defect detection unit 1100 includes a first camera 1101 and a second camera 1102. The first camera 1101 is positioned directly above the bamboo strip 10 and faces downwards to capture images of the upper surface of the bamboo strip. The second camera 1102 is positioned on one side of the bamboo strip 10 in the horizontal direction and faces the bamboo strip to capture images of the side of the bamboo strip. For ease of description, the image captured by the first camera 1101 is referred to as the first image, and the image captured by the second camera 1102 is referred to as the second image. By processing the first image and the second image, the defect characteristics of the bamboo strip before splicing can be obtained. The defect characteristics before splicing include one or more of the following: defect type, defect location, defect area, crack length, edge damage depth, wormhole area, abnormal bamboo node area, and fiber warping height. Based on the defect characteristics before splicing and a preset defect threshold, the bamboo strip 10 is sorted into qualified release, marked downgrade, or rejected. For bamboo strips that are deemed qualified or allowed to be downgraded, the first conveying group 201 continues to feed them into the cutting, gluing and compaction unit 300; for bamboo strips that are deemed unqualified, the defect sorting execution mechanism performs rejection, alarm or shutdown inspection.

[0018] The adhesive bonding area detection unit 1200 is located after the cutting, gluing, and compaction unit 300 and before the second conveying group 202, and is used to detect the joint area of ​​the bamboo strips after adhesive bonding. The adhesive bonding area detection unit 1200 includes a third camera 1201 and a fourth camera 1202. The third camera 1201 is positioned directly above the adhesive bonding area and shooting downwards to capture images of the upper surface of the joint; the fourth camera 1202 is positioned on one side of the adhesive bonding area in the horizontal direction and facing the hinge area to capture images of the side contour of the joint. For ease of description, the image captured by the third camera 1201 is referred to as the third image, and the image captured by the fourth camera 1202 is referred to as the fourth image. By processing the third and fourth images, the characteristics of the adhesive bonding area at the bamboo strip joint can be obtained. These characteristics include overlap length, glue coverage, glue overflow rate, joint misalignment, joint skew angle, thickness variation, and lateral warping.

[0019] The width detection unit 1300 is positioned after the width fixing component 700 and before the tension component 800, and is used to detect the dimensions of the bamboo strips after width fixing. The width detection unit 1300 includes a fifth camera 1301, which is positioned above the bamboo strip 10 to acquire images of the bamboo strips after width fixing, serving as the fifth image. Based on the fifth image, the width, width fluctuation, edge straightness, and burr area of ​​the bamboo strips are calculated as width-fixed dimensional characteristics.

[0020] The first conveying group 201, the cutting, gluing and compacting unit 300, the second conveying assembly 202, the grinding assembly 600, the third conveying assembly 203, the width fixing assembly 700, the tension assembly 800, and the winding assembly 900 all adopt the technology in the patent application number 20251117168X or other technologies in the prior art, and the focus of this patent is not on these structures.

[0021] The independent modulus detection module 1400 includes a support 1400 and, in the feeding direction, a feeding guide component 1401, a first positioning component 1408, a loading mechanism 1402, a second positioning component 1409, and a discharge guide component 1407, which are sequentially arranged on the support 1400. The feeding guide component 1401 and the discharge guide component 1407 are used to drive and guide the bamboo strips 10, respectively. The loading mechanism 1402 is used to load the bamboo strips 10 downward, causing them to bend within a certain range. The first positioning component 1408 and the second positioning component 1409 are symmetrically arranged on both sides of the loading mechanism 1402 along the feeding direction, and are used to position the bamboo strips 10 on both sides when the loading mechanism 1402 bends downward. When the loading mechanism 1402 presses down on the bamboo strips 10, only the bamboo strips between the first positioning component 1408 and the second positioning component 1409 bend downward. Furthermore, the loading mechanism 1402 manually rotates the lead screw 14021 to press down on the bamboo strip 10. A magnetic ruler 1405 and a force sensor 1403 are provided on the loading mechanism 1402. The magnetic ruler 1405 is used to detect the downward distance of the loading mechanism 1402, and the force sensor 1403 is used to detect the force between the bamboo strip 10 and the loading mechanism 1402 in real time.

[0022] The independent modulus detection module 1400 also includes a laser displacement sensor 1404, which is used to detect the surface displacement of the bamboo strip, the displacement of the loading roller, or the deflection of the bamboo strip.

[0023] The bamboo strip lengthening system also includes a position synchronization module 1500, used to map data collected from different workstations to the same bamboo strip segment or the same joint. This module includes a first fiber optic sensor located before the cutting station and a second fiber optic sensor located between the cutting station and the gluing station. The distance from the first fiber optic sensor to the cutting station and the distance from the second fiber optic sensor to the gluing station are predetermined. Through these two fiber optic sensors, accurate positioning of the bamboo strip 10 can be achieved when the signal from the two sensors changes. The bamboo strip segment is the bamboo strip located between two adjacent joints.

[0024] The location synchronization module 1500 operates as follows: 1) When the servo driver outputs the cumulative pulse count, the cumulative conveying length of the bamboo strips can be expressed as: , (1) in, For a moment The corresponding cumulative conveying length, This refers to the cumulative number of pulses fed back by the encoder or servo driver. The length of bamboo strips conveyed corresponding to a single pulse; 2) Because the bamboo strips are joined by overlapping glue, the effective length will be shortened at the joint; at the same time, compensation is needed for fixed-width cutting. Therefore, for the first... The cumulative length compensation of a single connector can be expressed as: , (2) in, For the first The overlap length of each connector is obtained from the fourth image. For the first The length correction amount formed during the cutting and width determination process of each joint or bamboo strip is the cutting length. The calculation starts when the end of the bamboo strip reaches the first fiber optic sensor, and the length corresponding to the encoder or servo driver is the cutting length. 3) For the first The first testing station, obtain the first Correction coordinates of each bamboo strip segment or joint : , (3) in, For the first The fixed distance between each testing station and the reference point. For the first Each bamboo strip segment or joint has undergone the first... The data collection time at each testing station. The cumulative conveying length fed back by the servo when the k-th bamboo strip or joint passes through the i-th inspection station; The inspection station includes the front-end dual-view defect inspection unit 1100, the adhesive bonding area inspection unit 1200, the width inspection unit 1300, and the positions corresponding to the first and second fiber optic sensors. 4) When the corrected position coordinates obtained from two testing stations satisfy the following formula, they are determined to correspond to the same bamboo strip segment or the same joint: , (4) in, This is the allowable error threshold; 5) If different workstations are matched by time, the time compensation amount... Represented as: , (5) in, For the first The equivalent length compensation amount generated during the lengthening, cutting, or width determination of a bamboo strip segment or joint, for a bamboo strip segment, For a joint, the length to be cut plus the length loss due to a fixed bandwidth is the total length. The length of the bamboo strip cut off at both ends plus the length of the overlap. and These are the distances of the i-th and j-th workstations relative to the same reference point, which can be determined manually. This refers to the conveying speed corresponding to the bamboo strip or joint.

[0025] The above steps can bind the front-end dual-view images, glued area images, fixed-width images, and mechanical response data collected by the independent modulus detection module to the same bamboo strip segment or the same joint.

[0026] The online prediction method for the equivalent elastic modulus of bamboo strips specifically includes the following steps: Step 1: Each camera acquires images of bamboo strips at different locations, obtaining the first, second, third, fourth, and fifth images. These images are then processed to obtain image feature vectors. Specifically, it includes the following steps: Step 11: Collect images of bamboo strips at different locations using various machines to obtain the first image, second image, third image, fourth image, and fifth image. Then, use the position synchronization module to match the collected first image, second image, third image, fourth image, and fifth image with the corresponding bamboo strip segment or joint image. Step 12: Input the first, second, third, fourth, and fifth images corresponding to the k-th bamboo strip segment or joint into the YOLOv8-Seg instance segmentation model in sequence to obtain the defect tuple: , (6) in, The nth defect tuple output by the YOLOv8-Seg instance segmentation model. Let n be the defect type of the nth defect. Let be the detection bounding box for the nth defect, where the detection bounding box is the outer rectangle of the corresponding defect. This is the segmentation mask for the nth defect, where the pixel value in the defect area is 1, and the pixel value in the remaining areas is 0. Let n be the confidence level of the nth defect; Step 13: Obtain the image feature vector for each bamboo strip segment or joint based on the obtained defect tuple. : , (7) in, Let k be the image feature vector of the k-th bamboo strip segment or joint. This represents the sum of the defect areas of all defects within the k-th bamboo strip segment or joint. This is the sum of the crack lengths of all defects within the k-th bamboo strip segment or joint. This represents the maximum edge breakage depth of all defects within the k-th bamboo strip segment or joint. The maximum fiber warping height is the highest among all defects within the k-th bamboo strip segment or joint. Let be the coverage rate of the adhesive layer within the k-th bamboo strip segment or joint. Let be the glue overflow rate in the k-th bamboo strip segment or joint. This represents the overlap length of the k-th bamboo strip segment or joint. In practice, only the joint has an overlap length, but the number k is assigned to both the bamboo strip segment and the joint as a whole. Let be the misalignment amount of the k-th bamboo strip segment or joint. Let be the joint deviation angle of the k-th bamboo strip segment or joint. Let represent the change in joint thickness at the k-th bamboo strip segment or joint. Let be the width of the bamboo strip at the k-th bamboo segment or joint. Let be the width fluctuation of the k-th bamboo strip segment or joint. The area of ​​the edge burrs of the k-th bamboo strip segment or joint.

[0027] further, It can be calculated using the following formula: , (8) in, The physical length corresponding to a single pixel is obtained from camera calibration. The area of ​​a single pixel. is the total number of defects in the k-th bamboo strip segment or joint, and m is the defect number in the k-th bamboo strip segment or joint.

[0028] The calculation formula is as follows: , (9) in, This is the set of defects containing cracks in the k-th bamboo strip segment or joint. The total number of pixels for the nth crack can be obtained by refining the skeleton of each crack mask and then extracting the total number of pixels along the crack centerline. .

[0029] The calculation formula is as follows: , (10) in, To obtain the baseline edge contour line of the bamboo strip body, which is obtained by fitting edge points from the acquired image, points at defective locations need to be removed during the fitting process. This is the set of mask edge points for edge damage defects. For pixels The vertical pixel distance to the reference edge contour line, max() is the value to be taken. The maximum value in.

[0030] The calculation formula is as follows: (11) in, The reference upper surface line of the bamboo strip matrix is ​​obtained by fitting it to the side view image. The reference upper surface line is obtained by fitting the upper edge points of the bamboo strips in the side view image. During the fitting process, points at defect locations need to be removed. This is the set of masking points for fiber warping defects. For pixels The distance to the reference upper surface line, To obtain The maximum value in.

[0031] The calculation formula is as follows: , (12) in, The effective adhesive layer area within the joint region is [area], and the overlapping area of ​​the two bamboo strips at the joint is [area], obtained through the third image. This is the area of ​​the preset joint area.

[0032] The calculation formula is as follows: , (13) in, The area of ​​adhesive overflow in the bonding region was obtained from the third and fourth images.

[0033] The calculation formula is as follows: , (14) in, and These are the maximum and minimum pixel coordinates of the overall mask of the connector in the conveying direction.

[0034] The calculation formula is as follows: , (15) in, and These refer to the centerline position or edge reference position of the two bamboo strips at the joint area.

[0035] The calculation formula is as follows: , (16) in, , These are the axial and lateral pixel differences of the longest side of the smallest bounding rectangle, respectively.

[0036] The calculation formula is as follows: , (17) in, For the thickness of the joint area, This is the normal thickness of bamboo strips.

[0037] The calculation formula is as follows: , (18) in, and These are the pixel coordinates of the right and left edges of the bamboo strip in the image, respectively. The pixel equivalent is obtained from the calibration of the fifth camera.

[0038] The calculation formula is as follows: , (19) Where m represents the number of sampling points taken at equal intervals along the conveying direction for the k-th bamboo strip segment or joint. For the first The width of each sampling point.

[0039] The calculation formula is as follows: , (20) in, This is a set of edge burr-like defects. The total number of pixels for the nth edge burr-type defect.

[0040] Step 2: Obtain the raw mechanical test data from the independent modulus testing module 1400 and process the raw mechanical test data to obtain the mechanical response feature vector. Specifically, it includes the following steps: Step 21: Obtain the loading force collected by the force sensor through the independent modulus detection module 1400. The displacement or deflection of the bamboo strips or joints is collected by a laser displacement sensor. Vertical displacement of the loading mechanism acquired by the magnetic ruler The bamboo strip conveying position is fed back by the servo motor or servo driver in the independent modulus detection module. and conveying speed ; Step 22: Use a position synchronization module to apply force Displacement or deflection Vertical displacement Bamboo strip conveying location and conveying speed Match it with the corresponding bamboo strips or joints; Step 23, for the first Each bamboo strip or joint, based on the moment it reaches the mechanical testing area. Based on the baseline, extract the time window. Within the stated time window, the original mechanical test data is obtained. : (twenty one), in, , and It can be determined manually based on the actual situation; Step 24: Process the obtained raw mechanical test data Feature extraction is performed to obtain the mechanical response feature vector. : (twenty two), in, This represents the maximum loading force of the k-th bamboo strip segment or joint within the time window. The average loading force of the k-th bamboo segment or joint within the time window. The fluctuation of the loading force of the k-th bamboo strip segment or joint within the time window is obtained by subtracting the minimum loading force from the maximum loading force. This refers to the maximum displacement or maximum deflection of the k-th bamboo segment or joint within the time window. This refers to the change in displacement or deflection. Force-displacement slope The average conveying speed of the k-th bamboo strip segment or joint within the time window. The energy response per unit length of the k-th bamboo strip segment or joint within the time window; further, The calculation formula is as follows: (twenty three), The calculation formula is as follows: (twenty four), in, The time window corresponds to the length of the bamboo strips being transported.

[0041] Step 3: Obtain the process parameter vector for each bamboo strip segment or joint. : (25), in, This refers to the actual servo transmission speed. For servo acceleration or deceleration parameters, For servo-controlled fixed-length conveying, This refers to the amount of adhesive applied. For crimping time, The width of the bamboo strips, For the thickness of the bamboo strips, For the first The overlap length of each joint and other process parameters are all pre-set according to the actual situation.

[0042] Step 4: Obtain the image feature vector Mechanical response eigenvectors and extension process parameter vector The fused feature vector is obtained by standardization. Specifically, it includes the following steps: Step 41: Process the image feature vector Mechanical response eigenvectors and extension process parameter vector Standardization is performed to obtain the standardized image feature vector. Mechanical response eigenvectors and extension process parameter vector The standardized calculation formula is as follows: (26) in, Let j be the standardized value of the j-th dimension feature of the k-th bamboo strip segment or joint. Let be the original feature value of the j-th dimension of the k-th bamboo strip segment or joint. This represents the average value of the j-th dimension feature corresponding to the bamboo strip segment and the joint. The standard deviation of the j-th dimension characteristic corresponding to the bamboo strip segment and the joint; Step 42: Convert the standardized image feature vector Mechanical response eigenvectors and extension process parameter vector Perform feature fusion to obtain a fused feature vector. : (27) in, , where are weighting coefficients, used to adjust the weights of image features, mechanical response features, and process parameters in modulus prediction, respectively. ; Step 5: Fuse the feature vectors The equivalent elastic modulus of the long bamboo strips is predicted by inputting it into the modulus prediction model. The modulus prediction model adopts the XGBoost regression model, and the expression of the modulus prediction model is as follows: (28) in, For the first The coordinates of each bamboo strip segment or joint are corrected. Predicted equivalent elastic modulus at the location, For the first A tree of return, This represents the number of regression trees.

[0043] Step Six: Train the obtained modulus prediction model. The image feature vector, mechanical response feature vector, and process parameter vector in the training data can be obtained through actual detection or preset data. The true equivalent elastic modulus is calculated using the following formula: (29) in, The applied force is obtained through detection. To support the span, the distance between the first positioning component 1408 and the second positioning component 1409 in the independent modulus detection module, The width of the bamboo strips, For the thickness of the bamboo strips, The mid-span deflection is obtained by laser displacement sensor 1404.

[0044] The training objective function is: (30) in, The true equivalent modulus of elasticity obtained from offline tensile or bending tests. This is the regularization term for the regression tree complexity.

[0045] Step 7: Determine the grade of the predicted equivalent elastic modulus value according to the preset elastic modulus threshold. (31), in, For the first The modulus grade of each bamboo strip segment or joint. and For the preset modulus threshold, if for If the level is set, the take-up assembly will be controlled to take up the winding mechanism normally; if for If the level is controlled, the marking mechanism or the grading and winding mechanism will be activated; if for If the level is high, then an alarm, removal, or shutdown inspection will be performed.

[0046] Step 8: When the modulus grade of multiple consecutive joints decreases, the process parameters can be adjusted to compensate for the decrease. (32), in, This is the process parameter vector corresponding to the next joint. This is the current process parameter vector. For the compensation coefficient vector, For the target equivalent elastic modulus, This is the current predicted equivalent elastic modulus.

[0047] 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 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 claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A method for online prediction of the equivalent elastic modulus of bamboo strips in a bamboo strip splicing system, the bamboo strip splicing system comprising, sequentially arranged along the direction of bamboo strip movement, a front-end dual-view defect detection unit, a cutting, gluing and compaction unit, a post-gluing dual-view detection unit, a width-fixing component, a width detection component, an independent modulus detection module, and a position synchronization module; characterized in that, The method specifically includes the following steps: Step 1: The front-end dual-view defect detection unit acquires the top surface and side images of the bamboo strips before splicing. After gluing, the dual-view detection unit acquires the top surface and side images of the spliced ​​bamboo strip segments and joints. The width-fixed detection component acquires the top surface images of the bamboo strip segments and the machine head. The obtained images are processed to obtain image feature vectors. ; Step 2: Obtain the raw mechanical test data from the independent modulus testing module and process the raw mechanical test data to obtain the mechanical response feature vector. ; Step 3: Obtain the splicing process parameter vector for each bamboo strip segment or joint. ; Step 4: Obtain the image feature vector Mechanical response eigenvectors and extension process parameter vector The fused feature vector is obtained by standardization. ; Step 5: Fuse the feature vectors The equivalent elastic modulus of the long bamboo strips is predicted by inputting it into the modulus prediction model.

2. The prediction method according to claim 1, characterized in that, Step one specifically includes the following steps: Step 11: The front-end dual-view defect detection unit acquires the upper surface image and side image of the bamboo strip before splicing. After gluing, the dual-view detection unit acquires the upper surface image and side image of the spliced ​​bamboo strip segment and joint. The obtained images are then matched with the corresponding bamboo strip segment or joint image using the position synchronization module. Step 12: Input all images corresponding to the k-th bamboo strip segment or joint obtained in Step 11 into the YOLOv8-Seg instance segmentation model to obtain the defect tuple: ， （1） in, The nth defect tuple output by the YOLOv8-Seg instance segmentation model. Let n be the defect type of the nth defect. Let n be the detection box for the nth defect. For the segmentation mask of the nth defect, Let n be the confidence level of the nth defect; Step 13: Obtain the image feature vector for each bamboo strip segment or joint based on the obtained defect tuple. : ， （2） in, Let k be the image feature vector of the k-th bamboo strip segment or joint. This represents the sum of the defect areas of all defects within the k-th bamboo strip segment or joint. This is the sum of the crack lengths of all defects within the k-th bamboo strip segment or joint. This represents the maximum edge breakage depth of all defects within the k-th bamboo strip segment or joint. The maximum fiber warping height is the highest among all defects within the k-th bamboo strip segment or joint. Let be the coverage rate of the adhesive layer within the k-th bamboo strip segment or joint. Let be the glue overflow rate in the k-th bamboo strip segment or joint. This represents the overlap length of the k-th bamboo strip segment or joint. In practice, only the joint has an overlap length, but the number k is assigned to both the bamboo strip segment and the joint as a whole. Let be the misalignment amount of the k-th bamboo strip segment or joint. Let be the joint deviation angle of the k-th bamboo strip segment or joint. Let represent the change in joint thickness at the k-th bamboo strip segment or joint. Let be the width of the bamboo strip at the k-th bamboo segment or joint. Let be the width fluctuation of the k-th bamboo strip segment or joint. The area of ​​the edge burrs of the k-th bamboo strip segment or joint.

3. The prediction method according to claim 1, characterized in that, Step two specifically includes the following steps: Step 21: Obtain the loading force collected by the force sensor through the independent modulus detection module. The displacement or deflection of the bamboo strips or joints is collected by a laser displacement sensor. Vertical displacement of the loading mechanism acquired by the magnetic ruler The bamboo strip conveying position is fed back by the servo motor or servo driver in the independent modulus detection module. and conveying speed ; Step 22: Use a position synchronization module to apply force Displacement or deflection Vertical displacement Bamboo strip conveying location and conveying speed Match it with the corresponding bamboo strips or joints; Step 23, for the first Each bamboo strip or joint, based on the moment it reaches the mechanical testing area. Based on the baseline, extract the time window. Within the stated time window, the original mechanical test data is obtained. : （3）， in, ; Step 24: Process the obtained raw mechanical test data Feature extraction is performed to obtain the mechanical response feature vector. : （4）， in, This represents the maximum loading force of the k-th bamboo strip segment or joint within the time window. The average loading force of the k-th bamboo segment or joint within the time window. The fluctuation of the loading force of the k-th bamboo strip segment or joint within the time window is obtained by subtracting the minimum loading force from the maximum loading force. This refers to the maximum displacement or maximum deflection of the k-th bamboo segment or joint within the time window. This refers to the change in displacement or deflection. Force-displacement slope The average conveying speed of the k-th bamboo strip segment or joint within the time window. The energy response per unit length of the k-th bamboo segment or joint within the time window.

4. The prediction method according to claim 1, characterized in that, Step 3: Extend the process parameter vector It is expressed as follows: （5）， in, This refers to the actual servo transmission speed. For servo acceleration or deceleration parameters, For servo-controlled fixed-length conveying, This refers to the amount of adhesive applied. For crimping time, The width of the bamboo strips, For the thickness of the bamboo strips, For the first The overlap length of each connector.

5. The prediction method according to claim 1, characterized in that, Step four specifically includes the following steps: Step 41: Process the image feature vector Mechanical response eigenvectors and extension process parameter vector Standardization is performed to obtain the standardized image feature vector. Mechanical response eigenvectors and extension process parameter vector ; Step 42: Convert the standardized image feature vector Mechanical response eigenvectors and extension process parameter vector Perform feature fusion to obtain a fused feature vector. : （6）， in, , where are weighting coefficients, used to adjust the weights of image features, mechanical response features, and process parameters in modulus prediction, respectively. .

6. The prediction method according to claim 1, characterized in that, Step 5: The expression for the modulus prediction model is as follows: （7）， in, For the first The coordinates of each bamboo strip segment or joint are corrected. Predicted equivalent elastic modulus at the location, For the first A tree of return, This represents the number of regression trees.

7. The prediction method according to claim 1, characterized in that, The process also includes step six: determining the grade of the predicted equivalent elastic modulus value according to a preset elastic modulus threshold. （8）， in, For the first The modulus grade of each bamboo strip segment or joint. and This is a preset modulus threshold.

8. The prediction method according to claim 1, characterized in that, Step 8: When the modulus grade of multiple consecutive joints decreases, compensate and adjust the process parameters: （9）， in, This is the process parameter vector corresponding to the next joint. This is the current process parameter vector. For the compensation coefficient vector, For the target equivalent elastic modulus, This is the current predicted equivalent elastic modulus.

Images