Plastic product strength detection method and device based on non-contact measurement

Abstract

The present invention discloses a plastic product strength detection method and device based on non-contact measurement, which relates to the technical field of wind turbine monitoring. The method comprises: scanning a target detection object and constructing a three-dimensional model thereof, and then segmenting the target according to thickness and shape changes. The area proportion and morphological continuity relationship of the segmentation result are analyzed to determine the detection direction and sequence. The detection module is activated to perform ultrasonic detection, the waveform and propagation speed of the detection ultrasonic wave are collected and analyzed, and a mapping relationship between the waveform and propagation speed and the segmented area is established. Waveform alignment and comparison are performed according to a data list to obtain fitting deviation information. The hardness test result of the target detection object is determined according to the fitting deviation information, and the deviation position is positioned and fed back based on the three-dimensional model. The technical problems of low detection efficiency and high detection cost are solved, and the technical effect of reducing detection cost and improving detection efficiency is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of wind turbine monitoring, and in particular to a plastic product strength detection method and device based on non-contact measurement. Background Art

[0002] Non-contact measurement technology has been widely used in many fields, usually using optical, electromagnetic or acoustic technologies to measure the physical or chemical properties of objects at a distance.

[0003] Traditional methods for testing the strength of plastic products mainly include tensile testing, bending testing, compression testing, etc. Existing contact methods may damage the sample, and the operation process is complicated and needs to be operated by professionals. There are technical problems such as low testing efficiency and high testing cost. Summary of the invention

[0004] The present invention provides a plastic product strength detection method and device based on non-contact measurement to solve the technical problems of low detection efficiency and high detection cost in the prior art, and achieves the technical effect of reducing detection cost and improving detection efficiency.

[0005] In a first aspect, the present invention provides a plastic product strength detection method based on non-contact measurement, wherein the method comprises:

[0006] Scanning the target object to construct a three-dimensional model of the target object, and segmenting the target according to thickness distribution and shape change based on the three-dimensional model;

[0007] According to the target segmentation results, the area ratio and morphological continuity relationship analysis are performed, and the detection direction and detection order of the segmented area are configured;

[0008] Based on the detection direction and detection order of the segmented area, the detection module is activated to perform ultrasonic detection of the segmented area and collect detection ultrasonic waves;

[0009] Analyzing the waveform and propagation speed of the detection ultrasonic wave to obtain waveform change characteristics and ultrasonic wave propagation speed;

[0010] Establishing a mapping relationship between the segmented area and the waveform change characteristics and ultrasonic wave propagation speed, constructing a data list, performing waveform alignment and comparison according to the data list, performing differential matching fitting, and obtaining fitting deviation information;

[0011] The hardness test result of the target test object is determined based on the fitting deviation information, and the deviation position is positioned and fed back based on the three-dimensional model.

[0012] In this method, non-contact hardness testing of the target object is achieved by combining three-dimensional model construction and ultrasonic testing technology. Accurate acquisition and reliability evaluation of hardness test results are achieved by using waveform alignment comparison and differential matching fitting. Positioning feedback based on the three-dimensional model can accurately locate the location of hardness anomalies, providing an important reference for subsequent processing. The comprehensive use of multiple technologies such as three-dimensional model construction, ultrasonic testing and waveform analysis improves the accuracy and reliability of hardness testing.

[0013] In a second aspect, the present invention further provides a plastic product strength detection device based on non-contact measurement, wherein the device comprises:

[0014] A model segmentation module, the model segmentation module is used to scan the target detection object, build a three-dimensional model of the target detection object, and perform target segmentation based on the three-dimensional model according to thickness distribution and shape change;

[0015] A detection configuration module, which is used to perform area ratio and morphological continuity relationship analysis based on the target segmentation results, and configure the detection direction and detection order of the segmented area;

[0016] A detection and collection module, the detection and collection module is used to activate the detection module based on the detection direction and detection order of the segmented area, perform ultrasonic detection of the segmented area, and collect detection ultrasonic waves;

[0017] An autonomous analysis module, the autonomous analysis module is used to analyze the waveform and propagation speed of the detection ultrasonic wave to obtain waveform change characteristics and ultrasonic wave propagation speed;

[0018] A deviation matching module, which is used to establish a mapping relationship between the segmented area and the waveform change characteristics and the ultrasonic wave propagation speed, construct a data list, perform waveform alignment and comparison according to the data list, perform differential matching fitting, and obtain fitting deviation information;

[0019] A deviation feedback module is used to determine the hardness test result of the target test object based on the fitting deviation information, and to provide positioning feedback on the deviation position based on the three-dimensional model.

[0020] The present invention discloses a plastic product strength detection method and device based on non-contact measurement, comprising: scanning the target detection object and constructing its three-dimensional model, and then segmenting the target according to thickness and shape changes. Analyzing the area ratio and morphological continuity relationship of the segmentation result to determine the detection direction and sequence. Activating the detection module to perform ultrasonic detection, collecting and analyzing the waveform and propagation speed of the detection ultrasonic wave, and establishing a mapping relationship between the waveform and propagation speed and the segmented area. Performing waveform alignment and comparison according to the data list to obtain fitting deviation information. Determine the hardness test result of the target detection object based on the fitting deviation information, and provide positioning feedback for the deviation position based on the three-dimensional model. The plastic product strength detection method and device based on non-contact measurement disclosed by the present invention solve the technical problems of low detection efficiency and high detection cost, and achieve the technical effect of reducing detection cost and improving detection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a schematic flow chart of a plastic product strength detection method based on non-contact measurement of the present invention;

[0022] Figure 2 It is a schematic structural diagram of a plastic product strength detection device based on non-contact measurement according to the present invention.

[0023] Explanation of the accompanying drawings: model segmentation module 11, detection configuration module 12, detection acquisition module 13, autonomous analysis module 14, deviation matching module 15, deviation feedback module 16. DETAILED DESCRIPTION

[0024] The technical solution provided in the embodiments of the present invention is to solve the technical problems of low detection efficiency and high detection cost in the prior art. The overall idea adopted is as follows:

[0025] First, scan the target object, build its three-dimensional model, and segment the target according to the thickness distribution and shape change. Next, analyze the area ratio and morphological continuity relationship based on the segmentation results, and configure the detection direction and order of the segmented area. Then, based on the configured detection direction and order, activate the detection module, perform ultrasonic detection on the segmented area, and collect the detection ultrasonic wave. Next, analyze the waveform and propagation velocity of the detection ultrasonic wave to obtain the waveform change characteristics and ultrasonic propagation velocity. Subsequently, establish a mapping relationship between the segmented area and the waveform change characteristics and ultrasonic propagation velocity, build a data list, and perform waveform alignment and comparison based on the data list, perform differential matching fitting, and obtain fitting deviation information. Finally, determine the hardness test result of the target object based on the fitting deviation information, and provide positioning feedback for the deviation position based on the three-dimensional model.

[0026] The above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods of the specification to better understand the above technical solution. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments used only to explain the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In addition, it should be noted that, for the convenience of description, only the parts related to the present invention are shown in the drawings, rather than all of them.

[0027] Embodiment 1

[0028] Figure 1 The present invention is a schematic flow chart of a plastic product strength detection method based on non-contact measurement; wherein, it includes:

[0029] Scanning the target object to construct a three-dimensional model of the target object, and segmenting the target according to thickness distribution and shape change based on the three-dimensional model;

[0030] The target object to be tested refers to the target plastic product to be tested for strength. Optionally, scan and build a three-dimensional model of the target object. First, through interaction with designers or relevant departments, obtain the design data of the target object, including design drawings, CAD models, etc. Then, parse the acquired design data to extract key geometric information and parameters. Next, based on the geometric information and parameters obtained by the analysis, use a three-dimensional scanning device (such as a laser scanner) to scan the target object to generate three-dimensional point cloud data. Finally, combine the design data of the target object with the three-dimensional point cloud data, and generate a three-dimensional model of the target object through three-dimensional modeling software or algorithm.

[0031] Through the above steps, the generation process from design data to three-dimensional point cloud data can be realized, providing basic data for subsequent three-dimensional model construction.

[0032] In some embodiments, segmenting the target based on the three-dimensional model according to thickness distribution and shape change includes:

[0033] According to the three-dimensional model, plane image recognition is performed to determine shape change nodes;

[0034] Based on the plane image, determine the graph nodes, perform layered superposition of the plane image, and determine the thickness difference nodes;

[0035] The target object is segmented according to the shape change nodes and the thickness difference nodes to obtain a target segmentation result, wherein the target segmentation result includes a plurality of segmentation areas, and each segmentation area has the same shape and orientation and consistent thickness.

[0036] Specifically, according to the three-dimensional model, before performing plane image recognition, first, projection or rendering technology projects the three-dimensional information onto a two-dimensional plane, and then converts the three-dimensional model into a plane image.

[0037] Optionally, the generated plane image is processed using image recognition or computer vision technology to determine the shape change node. Exemplarily, the edge in the plane image is identified by an edge detection algorithm, a shape descriptor is used to describe and match the shape in the image, a deep learning model such as a convolutional neural network (CNN) is used to automatically extract and identify features in the image, and the shape change node is determined based on the mutation point of the feature or edge.

[0038] Optionally, the graph nodes are key points in the image, such as corner points, edge intersections, etc. The graph nodes are determined by feature detection algorithms (such as Harris corner detection, SIFT, SURF, etc.). After the graph nodes are determined, multiple plane images are layered and superimposed. Specifically, based on image registration and image fusion technology, first, the graph nodes of two or more images are aligned so that the plane images can be compared or combined in the same coordinate system. Then, the registered images are cumulatively combined to obtain a fused image with thickness features and determine the thickness difference nodes.

[0039] Exemplarily, the target object is segmented according to the shape change node and the thickness difference node. First, the shape change node and the thickness difference node are used as the initial segmentation points to pre-segment the target object. Then, based on the shape matching, shape alignment, shape regularization and other technologies, the pre-segmentation result is further refined so that each segmented area has the same shape orientation and thickness, ensuring the consistency of each segmented area in shape and orientation, and obtaining the target segmentation result. The target segmentation result includes multiple segmented areas, each of which has the same shape orientation and thickness.

[0040] According to the target segmentation results, the area ratio and morphological continuity relationship analysis are performed, and the detection direction and detection order of the segmented area are configured;

[0041] In some embodiments, according to the target segmentation result, the area ratio of each segmented region is calculated, and the regions are arranged in positive order according to the area ratio;

[0042] Extract the segmented region with the largest area as the starting region, extend the adjacent segmented regions with the starting region as the center, and determine the continuous segmented regions;

[0043] Determine the detection direction according to the size distribution and extension scheme of the starting area, perform spatial dimension position relationship analysis on the continuous segmented areas based on the detection direction, and select segmented areas of the same spatial dimension for orientation marking;

[0044] Based on the orientation marking, the marked continuous segmented areas are analyzed for area ratio priority, and the first priority of the area ratio of each continuous segmented area is determined. The segmented area with the largest first priority is used as the second segmented area, and then the second segmented area is used as the center to identify continuous segmented areas and determine the third segmented area, and so on, to obtain the detection direction and detection order of all segmented areas.

[0045] Optionally, first, the number of pixels in each segmented area is calculated to obtain the area of ​​multiple segmented areas, and then their area proportions in the entire image are calculated. Then, the segmented area with the largest area proportion is selected as the starting area, and then with this area as the center, adjacent segmented areas are determined according to the spatial position relationship to form a continuous segmented area.

[0046] Optionally, the detection direction is determined based on the size distribution of the starting area and the extension scheme, including calculating the centroid of the starting area and then determining the extension direction based on the position of the centroid and the shape of the area. The extension direction refers to the direction in which the shape of the starting area changes fastest.

[0047] Furthermore, based on the determined detection direction, the spatial dimension position relationship of the continuous segmented areas is analyzed to determine the areas with consistent detection directions, and the orientation of the segmented areas with consistent detection directions is marked, thereby reducing the control trajectory distance of the detection discussion.

[0048] Optionally, based on the orientation annotation, the marked continuous segmented areas are analyzed for area ratio priority, and the first priority of the area ratio of each continuous segmented area is determined. Then, the segmented area with the largest first priority area ratio is used as the second segmented area, and then the continuous segmented areas are iteratively identified and determined with the second segmented area as the center, and the above steps are repeated until the detection direction and detection order of all segmented areas are determined.

[0049] In some implementations, obtaining the detection directions and detection orders of all segmented regions includes:

[0050] Based on the three-dimensional model and the target segmentation result, a segmentation topology structure is constructed with the segmented area as a node, and a topological connection direction is set according to the detection direction of the segmented area;

[0051] Adding the area ratio of each segmented area to the node and setting it as the node coefficient;

[0052] Taking the starting area as the center, taking the minimum node distance as the first classification feature, and detecting the consistency of the direction as the second classification feature, constructing a classification decision tree, and obtaining multiple association chains;

[0053] The node coefficients in the multiple associated chains are superimposed to determine the chain coefficient of each relationship chain, and the node relationship with the largest chain coefficient is obtained to determine the detection direction and detection order of the segmented area, wherein the node order is the detection order and the topological connection direction is the detection direction.

[0054] Exemplarily, the detection direction and detection order of all segmented areas are obtained. First, based on the three-dimensional model and the target segmentation results, each segmented area is regarded as a node, and the topological connection direction is set according to the spatial position relationship and detection direction of the segmented area. Then, node empowerment is performed, and the area proportion of each segmented area is added to the node, which is set as the node coefficient. The node coefficient defines the area weight of the node based on the segmented area. Next, with the starting area as the center point, the distance from each node to the starting area is calculated using the Euclidean distance or other distance measurement method as the first classification feature, and then the consistency of the detection direction is used as the second classification feature. If the moving direction of the node on different layers is consistent, it can be regarded as a group, and then these two features are used to construct a classification decision tree. The decision tree divides the nodes into different categories according to the first classification feature and the second classification feature, thereby achieving the segmentation of the target. Among them, the classification decision tree includes multiple association chains, and the nodes in each association chain have consistent feature categories.

[0055] Through the above steps, target segmentation based on the starting area, the distance from the node to the starting area and the consistency of the detection direction can be achieved, and the nodes are divided into different categories by the classification decision tree, thereby achieving effective target segmentation.

[0056] Based on the detection direction and detection order of the segmented area, the detection module is activated to perform ultrasonic detection of the segmented area and collect detection ultrasonic waves;

[0057] Specifically, after determining the detection direction and detection order of the segmented area, first, an activation signal is sent or an activation function is called to start the ultrasonic detection module. The ultrasonic detection module includes one or more groups of ultrasonic sensors for non-contact measurement.

[0058] Optionally, activating the detection module includes configuring or moving ultrasonic sensors at corresponding positions based on the detection direction and detection order of the segmented areas, and defining the movement order and activation order of the ultrasonic sensors in the ultrasonic detection module.

[0059] Further, after the detection module is activated, the activated detection module will perform ultrasonic detection on the segmented area according to the detection direction and detection order of the segmented area, including emitting ultrasonic waves using the ultrasonic detection module and receiving reflected ultrasonic waves according to a certain path and speed, following the detection direction and detection order.

[0060] Optionally, the detected ultrasonic waves are collected and converted into electrical signals, which are then collected using a data acquisition device and converted into a data format that can be further analyzed and processed. These signals contain information about the internal structure, density, shape, etc. of the target object.

[0061] Analyzing the waveform and propagation speed of the detection ultrasonic wave to obtain waveform change characteristics and ultrasonic wave propagation speed;

[0062] Optionally, the waveform and propagation speed of the ultrasonic wave reflect the internal properties of the target object, and the received ultrasonic wave signal is subjected to waveform analysis, including analysis of the signal's amplitude, frequency, waveform shape and other characteristics, to obtain information about the target object's internal structure, such as existing defects or foreign matter.

[0063] For example, Fourier Transform or Wavelet Transform is used to analyze the frequency component and time-frequency characteristics of ultrasound, and to detect the waveform and propagation speed of ultrasound. At the same time, the waveform change characteristics are obtained by comparing the peak value, valley value, period, amplitude and other parameters of the ultrasound.

[0064] Exemplarily, the propagation time and path length of the ultrasonic signal are obtained to calculate the propagation speed of the ultrasonic wave in the target object. The propagation speed is affected by the material and structure of the target object, so it can be used to identify areas of different materials or structures. Specifically, the propagation speed is faster in materials with larger density and elastic modulus.

[0065] By acquiring the waveform change characteristics and ultrasonic propagation speed, we can obtain more detailed information about the internal structure and properties of the target detection object, which helps to understand the state and characteristics of the target object more accurately.

[0066] Establishing a mapping relationship between the segmented area and the waveform change characteristics and ultrasonic wave propagation speed, constructing a data list, performing waveform alignment and comparison according to the data list, performing differential matching fitting, and obtaining fitting deviation information;

[0067] Optionally, based on the correspondence between the detection direction and detection sequence and the detection ultrasonic wave, and the correspondence between the detection ultrasonic wave and the waveform change characteristics and the ultrasonic wave propagation speed, the waveform change characteristics and propagation speed data of each segmented area are associated with the corresponding segmented area, and then the position information, waveform change characteristics, and ultrasonic wave propagation speed of each area are recorded in a data list to construct a data list. The data list is a structured data set for subsequent analysis and prediction.

[0068] In some embodiments, waveform alignment and comparison are performed according to the data list, and differential matching fitting is performed to obtain fitting deviation information, including:

[0069] Based on the segmented areas, obtaining the waveform and propagation speed of each segmented area;

[0070] Align and compare the waveforms and propagation velocities of the segmented regions to determine the comparison difference values ​​of the segmented regions;

[0071] Extracting the thickness and shape of each segmented area and determining a target difference value;

[0072] The comparison difference value and the target difference value are used to perform fitting to obtain the fitting deviation information.

[0073] Optionally, first, the data list is interacted with, and the corresponding waveform and propagation speed are called according to the segmented area; then, the waveform and propagation speed of each segmented area are aligned and compared by calculating the similarity or difference between the waveform and the propagation speed. Exemplarily, the difference between the waveform and the propagation speed is calculated using a metric such as a correlation coefficient, Euclidean distance, or Manhattan distance, and the comparison difference value of each segmented area is determined.

[0074] Optionally, the thickness and shape of each segmented area are extracted to determine a target difference value, wherein the target difference value refers to the maximum difference value allowed under the segmented area quality control requirement. If the difference value of the target detection object is greater than the target difference value, it means that the deviation of the target detection object is too large.

[0075] Optionally, fitting is performed using the comparison difference value and the target difference value. The fitting is performed by linear regression, nonlinear regression, machine learning algorithm, etc., so as to obtain fitting deviation information. Exemplarily, the fitting deviation information is the integral of the target difference value of multiple segmented areas of the target detection object over the detection path.

[0076] In some implementations, extracting the thickness and shape of each segmented area and determining the target difference value includes:

[0077] According to the material of the test object, multiple experimental samples are constructed according to thickness and shape, and ultrasonic testing experiments are carried out to obtain experimental record data sets, which include material thickness, shape information, size information and corresponding ultrasonic signals;

[0078] Performing conventional, reflection, diffraction waveform analysis and propagation velocity analysis on the ultrasonic signal to obtain ultrasonic characteristics and add them into the experimental record data set;

[0079] Based on the experimental record data set, a neural network model is trained to perform ultrasonic signal difference relationship recognition learning to obtain a difference analysis model, wherein the difference analysis model is used to identify the difference relationship between material thickness, shape and ultrasonic signal parameters, and output an ultrasonic difference value;

[0080] The difference values ​​of the thickness and shape of each segmented area are input into the difference analysis model to obtain the target difference value.

[0081] Specifically, first, multiple experimental samples are constructed according to the thickness and shape of different materials, and experimental measurements are performed. By performing ultrasonic detection on sample materials with known material thickness, shape information, and size information, the corresponding ultrasonic signals are obtained, and the correlation between the material thickness, shape information, and size information and the corresponding ultrasonic signals is established and stored as an experimental record data set. The experimental record data set reflects the correspondence between various thicknesses and shapes that may appear in the material and the ultrasonic signal. Then, the recorded ultrasonic signal is subjected to conventional, reflection, and diffraction waveform analysis and propagation velocity analysis to obtain the characteristics of the ultrasonic signal, such as amplitude, frequency, phase, attenuation, etc., and stored in the experimental record data set; then, based on the experimental record data set, a neural network model is constructed and trained to learn the complex mapping relationship between the acoustic wave signal and the material thickness and shape, and to obtain a difference analysis model, wherein the difference analysis model takes the ultrasonic signal as input, and outputs the ultrasonic difference value after end-to-end processing and analysis.

[0082] Through the training of the neural network model, complex difference relationships can be automatically learned, thereby improving the accuracy and efficiency of detection.

[0083] Optionally, based on a preset update cycle, the differential analysis model is automatically updated and optimized to adapt to new data and conditions, ensuring the long-term accuracy of differential analysis.

[0084] In some implementations, performing fitting using the comparison difference value and the target difference value to obtain the fitting deviation information includes:

[0085] Based on the size and detection direction of the segmented area, obtaining the detection direction difference distance of the segmented area;

[0086] Obtaining the ultrasonic loss coefficient of the material of the test object;

[0087] Determining the ultrasonic loss amount according to the detection direction difference distance and the ultrasonic loss coefficient;

[0088] The target difference value is corrected according to the ultrasonic wave loss amount.

[0089] Optionally, the comparison difference value and the target difference value are used for fitting to evaluate the accuracy and deviation of the model. First, based on the size of the segmented area and the detection direction, the detection direction difference distance of the segmented area is obtained. The difference distance reflects the propagation difference of the ultrasonic wave in the detection direction. Then, based on the material of the detection object, the ultrasonic loss coefficient is extracted. When the ultrasonic wave passes through the material, it will encounter resistance, resulting in energy loss. The ultrasonic loss coefficient reflects the energy loss of the ultrasonic wave during the propagation process. Different materials have different ultrasonic loss coefficients. Then, according to the detection direction difference distance and the ultrasonic loss coefficient, the ultrasonic loss amount is determined. The ultrasonic loss amount is the product of the detection direction difference distance and the ultrasonic loss coefficient. The ultrasonic loss amount is used to correct the target difference value to reduce the impact of propagation differences and improve the accuracy of detection.

[0090] Optionally, the target difference value is corrected according to the ultrasonic loss amount, including adding the ultrasonic loss amount to the target difference value.

[0091] The hardness test result of the target test object is determined based on the fitting deviation information, and the deviation position is positioned and fed back based on the three-dimensional model.

[0092] Optionally, based on the fitting deviation information of multiple segmented areas, the hardness of multiple positions of the target object is judged to obtain the hardness test results. Specifically, if the fitting deviation information is large, it means that the position of the target object corresponding to the fitting deviation information may have defects such as poor uniformity, structural defects, and substandard density.

[0093] Optionally, locating the deviation position based on the three-dimensional model helps relevant technical personnel understand the specific location where the deviation occurs. Specifically, the fitting deviation information is mapped to the three-dimensional model, and this area is marked on the three-dimensional model for corresponding feedback and adjustment.

[0094] In summary, the plastic product strength detection method based on non-contact measurement provided by the present invention has the following technical effects:

[0095] By scanning the target object, a three-dimensional model of the target object is constructed, and the target is segmented according to the thickness distribution and shape change based on the three-dimensional model; the area ratio and morphological continuity relationship analysis are performed according to the target segmentation result, and the detection direction and detection order of the segmented area are configured; based on the detection direction and detection order of the segmented area, the detection module is activated, and ultrasonic detection of the segmented area is performed, and the detection ultrasonic wave is collected; the waveform and propagation speed of the detection ultrasonic wave are analyzed to obtain the waveform change characteristics and ultrasonic wave propagation speed; the mapping relationship between the segmented area and the waveform change characteristics and ultrasonic wave propagation speed is established, and a data list is constructed. According to the data list, waveform alignment and comparison are performed, and differential matching fitting is performed to obtain fitting deviation information; the hardness test result of the target object is determined based on the fitting deviation information, and the deviation position is positioned and fed back based on the three-dimensional model. The technical problems of low detection efficiency and high detection cost are solved, and the technical effect of reducing detection cost and improving detection efficiency is achieved.

[0096] Embodiment 2

[0097] Figure 2 Schematic diagram of the structure of the plastic product strength detection device based on non-contact measurement of the present invention. Figure 1 The schematic flow chart of the plastic product strength detection method based on non-contact measurement of the present invention can be shown as follows: Figure 2 The structure shown is implemented.

[0098] Based on the same concept as the plastic product strength detection method based on non-contact measurement in the above embodiment, the present invention also provides a plastic product strength detection device based on non-contact measurement, including:

[0099] The model segmentation module 11 is used to scan the target object, construct a three-dimensional model of the target object, and segment the target according to thickness distribution and shape change based on the three-dimensional model;

[0100] A detection configuration module 12 is used to analyze the area ratio and morphological continuity relationship according to the target segmentation result, and configure the detection direction and detection order of the segmented area;

[0101] A detection and collection module 13 is used to activate the detection module based on the detection direction and detection order of the segmented area, perform ultrasonic detection of the segmented area, and collect detection ultrasonic waves;

[0102] The autonomous analysis module 14 is used to analyze the waveform and propagation speed of the detection ultrasonic wave to obtain the waveform change characteristics and ultrasonic wave propagation speed;

[0103] The deviation matching module 15 is used to establish a mapping relationship between the segmented area and the waveform change characteristics and the ultrasonic wave propagation speed, construct a data list, perform waveform alignment and comparison according to the data list, perform differential matching fitting, and obtain fitting deviation information;

[0104] The deviation feedback module 16 is used to determine the hardness test result of the target test object based on the fitting deviation information, and to provide positioning feedback on the deviation position based on the three-dimensional model.

[0105] Among them, the model segmentation module 11 includes:

[0106] A plane image recognition unit, used to perform plane image recognition according to the three-dimensional model and determine shape change nodes;

[0107] A layered overlay unit is used to determine the graph nodes based on the plane image, perform layered overlay of the plane image, and determine the thickness difference nodes;

[0108] The target object segmentation unit is used to segment the target object according to the shape change node and the thickness difference node to obtain a target segmentation result, wherein the target segmentation result includes multiple segmentation areas, each segmentation area has the same shape and orientation and consistent thickness.

[0109] Furthermore, the detection configuration module 12 includes:

[0110] A positive sequence arrangement unit, used for calculating the area ratio of each segmented area according to the target segmentation result, and performing positive sequence arrangement according to the area ratio;

[0111] A region extension unit is used to extract the segmented region with the largest area as the starting region, extend adjacent segmented regions with the starting region as the center, and determine continuous segmented regions;

[0112] A position relationship analysis unit, configured to determine the detection direction according to the size distribution and extension scheme of the starting area, perform spatial dimension position relationship analysis on the continuous segmented areas based on the detection direction, and select segmented areas of the same spatial dimension for orientation marking;

[0113] A priority analysis unit is used to perform an area ratio priority analysis on the marked continuous segmented areas based on the orientation marking, determine the first priority of the area ratio of each continuous segmented area, take the segmented area with the largest first priority as the second segmented area, and then use the second segmented area as the center to identify the continuous segmented areas and determine the third segmented area, and so on, to obtain the detection direction and detection order of all segmented areas.

[0114] In some implementations, the detection configuration module 12 includes:

[0115] A segmentation topology unit, configured to construct a segmentation topology structure based on the three-dimensional model and the target segmentation result, using the segmented area as a node, and setting a topological connection direction according to a detection direction of the segmented area;

[0116] A node coefficient unit, used to add the area ratio of each segmented area to the node and set it as a node coefficient;

[0117] A classification decision unit is used to construct a classification decision tree with the starting area as the center, take the minimum node distance as the first classification feature, and detect the consistency of the direction as the second classification feature to obtain multiple association chains;

[0118] A coefficient superposition unit is used to superimpose the node coefficients in the multiple associated chains to determine the chain coefficient of each relationship chain, obtain the node relationship with the largest chain coefficient, and determine the detection direction and detection order of the segmented area, wherein the node order is the detection order and the topological connection direction is the detection direction.

[0119] Furthermore, the deviation matching module 15 includes:

[0120] A data extraction unit, used for acquiring the waveform and propagation speed of each segmented area based on the segmented areas;

[0121] An alignment and comparison unit, used for aligning and comparing the waveforms and propagation speeds of the segmented regions to determine the comparison difference values ​​of the segmented regions;

[0122] A difference value unit, used to extract the thickness and shape of each segmented area and determine a target difference value;

[0123] The deviation fitting unit is used to perform fitting using the comparison difference value and the target difference value to obtain the fitting deviation information.

[0124] In some implementations, the difference value unit in the deviation matching module 15 includes:

[0125] The sample detection unit is used to construct multiple experimental samples according to the material of the test object, thickness and shape, conduct ultrasonic detection experiments, and obtain experimental record data sets, which include material thickness, shape information, size information and corresponding ultrasonic signals;

[0126] A feature analysis unit, used to perform conventional, reflection, diffraction waveform analysis and propagation velocity analysis on the ultrasonic signal to obtain ultrasonic features and add them to the experimental record data set;

[0127] A relationship recognition learning unit is used to train a neural network model based on the experimental record data set, perform ultrasonic signal difference relationship recognition learning, obtain a difference analysis model, and the difference analysis model is used to identify the difference relationship between material thickness, shape and ultrasonic signal parameters, and output an ultrasonic difference value;

[0128] The difference analysis unit is used to input the difference values ​​of the thickness and shape of each segmented area into the difference analysis model to obtain the target difference value.

[0129] In some implementations, the deviation fitting unit in the deviation matching module 15 includes:

[0130] A difference distance unit, used for obtaining a detection direction difference distance of the segmented area based on the size and detection direction of the segmented area;

[0131] The loss coefficient unit is used to obtain the ultrasonic loss coefficient of the material of the test object;

[0132] A loss amount calculation unit, used to determine the ultrasonic loss amount according to the detection direction difference distance and the ultrasonic loss coefficient;

[0133] The difference value correction unit is used to correct the target difference value according to the ultrasonic wave loss amount.

[0134] It should be understood that the embodiments mentioned in this specification focus on their differences from other embodiments. The specific embodiments in the aforementioned embodiment one are also applicable to the plastic product strength detection device based on non-contact measurement described in embodiment two. For the sake of brevity of the specification, they will not be further elaborated here.

[0135] It should be understood that the embodiments disclosed in the present invention and the above description can enable those skilled in the art to use the present invention to implement the present invention. At the same time, the present invention is not limited to the above-mentioned embodiments. It should be understood that those skilled in the art can still modify the technical solutions recorded in the above-mentioned embodiments, or replace some of the technical features therein by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A plastic product strength detection method based on non-contact measurement, characterized in that: The method comprises: Scanning the target object to construct a three-dimensional model of the target object, and segmenting the target according to thickness distribution and shape change based on the three-dimensional model; According to the target segmentation results, the area ratio and morphological continuity relationship analysis are performed, and the detection direction and detection order of the segmented area are configured; Based on the detection direction and detection order of the segmented area, the detection module is activated to perform ultrasonic detection of the segmented area and collect detection ultrasonic waves; Analyzing the waveform and propagation speed of the detection ultrasonic wave to obtain waveform change characteristics and ultrasonic wave propagation speed; Establishing a mapping relationship between the segmented area and the waveform change characteristics and ultrasonic wave propagation speed, constructing a data list, performing waveform alignment and comparison according to the data list, performing differential matching fitting, and obtaining fitting deviation information; Based on the fitting deviation information, the intensity of multiple positions of the target detection object is judged to determine the intensity detection result of the target detection object. If the fitting deviation information is large, it means that the position of the target detection object corresponding to the fitting deviation information has poor uniformity, structural defects and density defects that do not meet the standard, and positioning feedback is performed on the deviation position based on the three-dimensional model; Perform waveform alignment and comparison according to the data list, perform differential matching fitting, and obtain fitting deviation information, including: Based on the segmented areas, obtaining the waveform and propagation speed of each segmented area; Align and compare the waveforms and propagation velocities of the segmented regions to determine the comparison difference values ​​of the segmented regions; Extracting the thickness and shape of each segmented area and determining a target difference value; The comparison difference value and the target difference value are used to perform fitting to obtain the fitting deviation information.

2. The method according to claim 1, characterized in that The target is segmented based on the three-dimensional model according to thickness distribution and shape change, including: According to the three-dimensional model, plane image recognition is performed to determine shape change nodes; Based on the plane image, determine the graph nodes, perform layered superposition of the plane image, and determine the thickness difference nodes; The target object is segmented according to the shape change nodes and the thickness difference nodes to obtain a target segmentation result, wherein the target segmentation result includes a plurality of segmentation areas, and each segmentation area has the same shape and orientation and consistent thickness.

3. The method according to claim 2, characterized in that According to the target segmentation results, the area ratio and morphological continuity relationship analysis are performed, and the detection direction and detection order of the segmented area are configured, including: According to the target segmentation result, the area ratio of each segmented area is calculated, and the areas are arranged in positive order according to the area ratio; Extract the segmented region with the largest area as the starting region, extend the adjacent segmented regions with the starting region as the center, and determine the continuous segmented regions; Determine the detection direction according to the size distribution and extension scheme of the starting area, perform spatial dimension position relationship analysis on the continuous segmented areas based on the detection direction, and select segmented areas of the same spatial dimension for orientation marking; Based on the orientation marking, the marked continuous segmented areas are analyzed for area ratio priority, and the first priority of the area ratio of each continuous segmented area is determined. The segmented area with the largest first priority is used as the second segmented area, and then the second segmented area is used as the center to identify continuous segmented areas and determine the third segmented area, and so on, to obtain the detection direction and detection order of all segmented areas.

4. The method according to claim 3, characterized in that The obtaining of the detection directions and detection orders of all segmented regions includes: Based on the three-dimensional model and the target segmentation result, a segmentation topology structure is constructed with the segmented area as a node, and a topological connection direction is set according to the detection direction of the segmented area; Adding the area ratio of each segmented area to the node and setting it as the node coefficient; Taking the starting area as the center, taking the minimum node distance as the first classification feature, and detecting the consistency of the direction as the second classification feature, constructing a classification decision tree, and obtaining multiple association chains; The node coefficients in the multiple associated chains are superimposed to determine the chain coefficient of each relationship chain, and the node relationship with the largest chain coefficient is obtained to determine the detection direction and detection order of the segmented area, wherein the node order is the detection order and the topological connection direction is the detection direction.

5. The method according to claim 1, characterized in that Extracting the thickness and shape of each segmented area and determining the target difference value includes: According to the material of the test object, multiple experimental samples are constructed according to thickness and shape, and ultrasonic testing experiments are carried out to obtain experimental record data sets, which include material thickness, shape information, size information and corresponding ultrasonic signals; Performing conventional, reflection, diffraction waveform analysis and propagation velocity analysis on the ultrasonic signal to obtain ultrasonic characteristics and add them into the experimental record data set; Based on the experimental record data set, a neural network model is trained to perform ultrasonic signal difference relationship recognition learning to obtain a difference analysis model, wherein the difference analysis model is used to identify the difference relationship between material thickness, shape and ultrasonic signal parameters, and output an ultrasonic difference value; The difference values ​​of the thickness and shape of each segmented area are input into the difference analysis model to obtain the target difference value.

6. The method according to claim 5, characterized in that Using the comparison difference value and the target difference value to perform fitting to obtain the fitting deviation information includes: Based on the size and detection direction of the segmented area, obtaining the detection direction difference distance of the segmented area; Obtaining the ultrasonic loss coefficient of the material of the test object; Determining the ultrasonic loss amount according to the detection direction difference distance and the ultrasonic loss coefficient; The target difference value is corrected according to the ultrasonic wave loss amount.

7. A plastic product strength detection device based on non-contact measurement, characterized in that: The device comprises: A model segmentation module, the model segmentation module is used to scan the target detection object, build a three-dimensional model of the target detection object, and perform target segmentation based on the three-dimensional model according to thickness distribution and shape change; A detection configuration module, which is used to perform area ratio and morphological continuity relationship analysis based on the target segmentation results, and configure the detection direction and detection order of the segmented area; A detection and collection module, the detection and collection module is used to activate the detection module based on the detection direction and detection order of the segmented area, perform ultrasonic detection of the segmented area, and collect detection ultrasonic waves; An autonomous analysis module, the autonomous analysis module is used to analyze the waveform and propagation speed of the detection ultrasonic wave to obtain waveform change characteristics and ultrasonic wave propagation speed; A deviation matching module, which is used to establish a mapping relationship between the segmented area and the waveform change characteristics and the ultrasonic wave propagation speed, construct a data list, perform waveform alignment and comparison according to the data list, perform differential matching fitting, and obtain fitting deviation information; A deviation feedback module, wherein the deviation feedback module is used to perform intensity discrimination of multiple positions of the target detection object based on the fitting deviation information, and determine the intensity detection result of the target detection object. If the fitting deviation information is large, it means that the position of the target detection object corresponding to the fitting deviation information has poor uniformity, structural defects and substandard density defects, and positioning feedback is performed on the deviation position based on the three-dimensional model; The deviation matching module includes: A data extraction unit, used for acquiring the waveform and propagation speed of each segmented area based on the segmented areas; An alignment and comparison unit, used for aligning and comparing the waveforms and propagation speeds of the segmented regions to determine the comparison difference values ​​of the segmented regions; A difference value unit, used to extract the thickness and shape of each segmented area and determine a target difference value; The deviation fitting unit is used to perform fitting using the comparison difference value and the target difference value to obtain the fitting deviation information.

Images