Low-coherence imaging quality evaluation method and apparatus, and electronic device
Patent Information
- Application Number
- PCT/CN2025/133707
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-03-25
- Filing Date
- 2025-11-10
- Publication Date
- 2026-10-01
Smart Images
Figure CN2025133707_01102026_PF_FP_ABST
Abstract
Description
Methods, apparatus and electronic equipment for evaluating low coherence imaging quality Technical Field
[0001] This application relates to the field of laser welding monitoring technology, and in particular to a method, apparatus and electronic device for evaluating low coherence imaging quality. Background Technology
[0002] Laser welding is a high-precision welding method that uses a high-energy-density laser beam as a heat source. During laser welding, the metal vaporizes in the molten pool, forming a narrow keyhole. A continuous and stable keyhole is crucial for achieving sufficient melting of the solder and reducing welding defects. The key parameters of the keyhole are its depth and width. When the depth and width are too small, the molten pool formation rate is slow and the component forming efficiency is low; when the depth and width are too large, the high-temperature liquid metal zone in the molten pool is too large, and the grain growth environment temperature is too high, thus reducing the dimensional and shape accuracy of the parts.
[0003] It is evident that online monitoring of keyhole depth during laser welding is crucial. Currently, the primary method for keyhole depth monitoring is through a monitoring system based on optical low-coherence imaging. However, due to potential unknown errors or inaccuracies in the calibration process, the accuracy of keyhole depth information acquired via optical low-coherence imaging is low, leading to keyhole depth monitoring failure. Therefore, evaluating the accuracy of keyhole depth information acquired via optical low-coherence imaging is essential. Summary of the Invention
[0004] The main objective of this application is to provide a method, apparatus, and electronic device for evaluating the quality of low-coherence imaging, so as to evaluate the accuracy of key depth information acquired by optical low-coherence imaging, and to provide a prerequisite for online monitoring of key depth in the laser welding process.
[0005] To achieve the above objectives, this application provides a method for evaluating the quality of low-coherence imaging, comprising:
[0006] Acquire point cloud data related to the keyhole during laser welding, and obtain the peripheral point cloud projection map of the keyhole based on the point cloud data;
[0007] The keyhole region fitting process is performed multiple times to determine the keyhole region based on the fitting parameters obtained in each fitting process; wherein, in each fitting process, a coordinate system is constructed based on the outer point cloud projection map and a grid is divided according to a preset scale, and the fitting parameters of the keyhole region are determined based on the grid parameters of each grid; the preset scale is different in different fitting processes;
[0008] The imaging quality evaluation result is determined based on the point cloud data within the keyhole area and the area of the keyhole region.
[0009] Optionally, obtaining the peripheral point cloud projection map of the keyhole based on the point cloud data includes: segmenting the point cloud data using a preset segmentation formula to obtain a peripheral point cloud data set and a keyhole point cloud data set; and projecting the peripheral point cloud data set onto a preset welding surface to obtain the peripheral point cloud projection map.
[0010] Optionally, the preset segmentation formula is:
[0011]
[0012] In the formula, For the point cloud data of the j-th point before segmentation, The peripheral point cloud data set, The keyhole point cloud data set, Let be the perpendicular distance between the j-th point and the center point of the laser emitter. The vertical distance from the center point of the laser emitter to the preset welding surface. This is a preset distance threshold.
[0013] Optionally, the grid parameters include the number of point clouds within the grid, the grid area, and the distance between the grid and the central grid, where the central grid is the grid containing the origin of the coordinate system. Determining the fitting parameters for the keyhole region based on the grid parameters of each grid includes: for any grid, determining grid evaluation parameters using a preset index formula and based on the number of point clouds within the grid, the grid area, and the distance between the grid and the central grid; determining multiple target grids using a preset screening algorithm and based on the grid evaluation parameters of each grid; and inputting the point cloud data of the center points of each target grid into the Hough transform algorithm to obtain the fitting parameters for the keyhole region.
[0014] Optionally, determining multiple target grids using a preset filtering algorithm and based on the grid evaluation parameters of each grid includes: sorting the grid evaluation parameters in descending order; selecting grid evaluation parameters within a preset percentage range from the side with smaller grid evaluation parameters using the preset filtering algorithm, and using the grid corresponding to the selected grid evaluation parameters as the target grids.
[0015] Optionally, the preset index formula is:
[0016]
[0017] In the formula, Let i be the grid evaluation parameters for the i-th grid. Let i be the number of point clouds in the i-th grid. Let i be the area of the i-th grid. The distance between the i-th grid center point and the center grid center point is given.
[0018] Optionally, the point cloud data within the keyhole area includes the number of point clouds in the keyhole area, the signal strength of each point within the keyhole area, and the coordinate data of each point; determining the imaging quality evaluation result based on the point cloud data within the keyhole area and the area of the keyhole area includes: using a preset quality evaluation formula and based on the number of point clouds in the keyhole area, the signal strength of each point within the keyhole area, the coordinate data of each point, the coordinate data of the center point of the keyhole area, and the area of the keyhole area to determine the imaging quality evaluation result.
[0019] Optionally, the preset quality evaluation formula is:
[0020]
[0021] In the formula, The imaging quality evaluation result is as follows. The area of the keyhole region is [area]. The number of point clouds in the keyhole region. The signal strength at the z-th point in the i-th grid within the keyhole region. The total signal strength within the keyhole area. The coordinate data of the z-th point in the i-th grid within the keyhole area. The coordinates of the center point of the keyhole area.
[0022] Furthermore, to achieve the above objectives, this application also provides a low-coherence imaging quality evaluation device, comprising: an acquisition module for acquiring point cloud data related to a keyhole during laser welding, and obtaining a peripheral point cloud projection map of the keyhole based on the point cloud data; a fitting module for performing a keyhole region fitting process multiple times, and determining the keyhole region based on the fitting parameters obtained in each fitting process; wherein, in each fitting process, a coordinate system is constructed based on the peripheral point cloud projection map and a grid is divided according to a preset scale, and the fitting parameters of the keyhole region are determined based on the grid parameters of each grid; the preset scale is different in different fitting processes; and an evaluation module for determining an imaging quality evaluation result based on the point cloud data within the keyhole region and the area of the keyhole region.
[0023] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the low coherence imaging quality evaluation method as described above.
[0024] The low-coherence imaging quality evaluation method of this application acquires point cloud data related to the keyhole during laser welding and obtains a projection map of the outer point cloud of the keyhole based on the point cloud data. Then, it performs a keyhole region fitting process multiple times at different preset scales. In each fitting process, a coordinate system is constructed based on the outer point cloud projection map and a grid is divided according to the preset scale. The fitting parameters of the keyhole region are determined based on the grid parameters of each grid. The keyhole region is then fitted based on the fitting parameters obtained in each fitting process. Finally, the imaging quality evaluation result is determined based on the point cloud data and the area of the keyhole region, thereby evaluating the accuracy of key depth information acquired by optical low-coherence imaging.
[0025] Attached Figure Description
[0026] Figure 1 is a schematic diagram of a laser-welded keyhole according to a specific example of this application;
[0027] Figure 2 is one of the flowcharts of the low coherence imaging quality evaluation method according to an embodiment of this application;
[0028] Figure 3 is a second flowchart of the low coherence imaging quality evaluation method according to an embodiment of this application;
[0029] Figure 4 is a point cloud distribution map of the keyhole in a specific example of this application;
[0030] Figure 5 is a projection of the peripheral point cloud of a specific example of this application;
[0031] Figure 6 is a flowchart of the low coherence imaging quality evaluation method according to an embodiment of this application;
[0032] Figure 7 is a schematic diagram of mesh generation for a specific example of this application;
[0033] Figure 8 is a schematic diagram of the structure of a low-coherence imaging quality evaluation device according to an embodiment of this application;
[0034] Figure 9 illustrates a schematic diagram of the physical structure of an electronic device;
[0035] In the figure: 800, low coherence imaging quality evaluation device; 810, acquisition module; 820, fitting module; 830, evaluation module; 910, processor; 920, communication interface; 930, memory; 940, communication bus.
[0036] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Embodiments of the present invention
[0037] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0038] Laser welding uses a high-energy-density laser beam to bombard the workpiece, causing the weld joint to melt and form a molten pool. The weld seam is then joined by controlling parameters. Laser welding features low heat input, high speed, large weld depth-to-width ratio, and a small heat-affected zone, and is widely used in automotive manufacturing, electronics, aerospace, and other fields.
[0039] Figure 1 is a schematic diagram of a laser welding keyhole according to a specific example of this application. As shown in Figure 1, during laser welding, the metal vaporizes in the molten pool to form a long and narrow keyhole. A continuous and stable keyhole is an important factor in achieving sufficient melting of the solder and reducing the formation of welding defects. The key parameters of the keyhole are its depth (the vertical distance from the keyhole surface to the deepest point of the keyhole) and width. When the width and depth are too small, the molten pool forming rate is slow and the component forming efficiency is low; when the width and depth are too large, the high-temperature liquid metal zone of the molten pool is too large, and the grain growth environment temperature is high, thereby reducing the forming size and shape accuracy of the parts.
[0040] Therefore, online monitoring of keyhole depth during laser welding is crucial. Current keyhole depth monitoring systems primarily employ optical low-coherence imaging (OLI) technology. However, accurate keyhole depth monitoring using OLI relies on the system's ability to accurately and effectively acquire keyhole information during welding. However, during data acquisition, errors such as incorrect welding process parameter settings, sensor installation errors, or low reflectivity of the welded object can render the acquired results unreliable, leading to the failure of keyhole depth monitoring.
[0041] Therefore, this application provides a method, apparatus, and electronic device for evaluating the quality of low-coherence imaging. By converting the point cloud data around the keyhole during welding into a two-dimensional projection map of the surrounding point cloud, and then fitting the keyhole region based on the projection map, the keyhole region in the optical low-coherence imaging is determined. Finally, based on the point cloud data and area of the keyhole region in the optical low-coherence imaging, the quality of the low-coherence imaging is evaluated, and the quality of the optical low-coherence imaging is judged. This can serve as a basis for investigating negative factors such as improper welding process parameter settings, large sensor installation errors, sensor incompatibility, or low reflectivity of the welded object, thereby improving the reliability of online monitoring of laser welding keyhole depth.
[0042] Figure 2 is a flowchart of one of the low-coherence imaging quality evaluation methods according to an embodiment of this application. This low-coherence imaging quality evaluation method can be executed by a processor in an optical low-coherence imaging monitoring device. As shown in Figure 2, the low-coherence imaging quality evaluation method may include the following steps:
[0043] Step 210: Obtain point cloud data related to the keyhole during laser welding, and obtain the outer point cloud projection map of the keyhole based on the point cloud data.
[0044] Step 220: Perform the keyhole region fitting process multiple times to determine the keyhole region based on the fitting parameters obtained in each fitting process; wherein, in each fitting process, a coordinate system is constructed based on the outer point cloud projection map and a grid is divided according to a preset scale, and the fitting parameters of the keyhole region are determined based on the grid parameters of each grid; the preset scale is different in different fitting processes.
[0045] Step 230: Determine the imaging quality evaluation result based on the point cloud data within the keyhole area and the area of the keyhole region.
[0046] It should be noted that the point cloud data related to the keyhole can include point cloud data of the keyhole space and the welding plane space. In this embodiment, the positional relationship between the welding plane space and the keyhole space is represented in the form of point cloud. In addition, the keyhole space can refer to the keyhole space marked in Figure 1, and the welding plane space can be the space above the keyhole space and below the laser emitter in Figure 1.
[0047] In this embodiment, the laser welding equipment can be pre-set with parameters and then started to begin welding on the test piece. The laser welding equipment may include a laser emitter, a water-cooling system, a gas protection system, and a high-speed galvanometer. Specifically, manually set process parameters are first input, and the controller in the laser welding equipment controls the laser emitter to weld on the test piece according to these parameters; the set welding path can be a straight line, and the welding surface of the test piece can be horizontal.
[0048] During welding, metal vaporizes and forms a keyhole in the molten pool. At this point, an optical low-coherence imaging monitoring device can be controlled to perform a three-dimensional scan of the keyhole surface. During the scan, the device records the interference signal resulting from the mixing of the reference light and the light from the keyhole surface. This interference signal reflects the intensity changes of the light reflected from the keyhole surface. Furthermore, by performing a Fourier transform on the acquired interference signal, the frequency information of the light reflected from the keyhole surface can be extracted, and the depth information of the keyhole surface can be calculated. Finally, based on the depth information of each measurement point and combined with the scanning position, three-dimensional coordinate data of the keyhole space and the welding plane space (i.e., the outer space between the keyhole and the laser emitter head) can be obtained. This three-dimensional coordinate data is the point cloud data.
[0049] In this embodiment, the optical low-coherence imaging monitoring equipment can continuously acquire keyhole-related point cloud data throughout the laser welding process, ultimately obtaining multiple sets of point cloud data. When evaluating the quality of optical low-coherence imaging, one set of point cloud data can be selected from these multiple sets for evaluation. Furthermore, after obtaining the keyhole-related point cloud data, it can be visualized to generate two-dimensional or three-dimensional point cloud images, facilitating the viewing of the point cloud data distribution by staff or simplifying subsequent processing of the point cloud data.
[0050] After obtaining the point cloud data related to the keyhole, this point cloud data can be divided into point cloud data in the keyhole space and point cloud data around the keyhole (i.e., the welding plane space). Furthermore, the point cloud data around the keyhole can be projected onto the welding plane to obtain the outer point cloud projection map.
[0051] In step 220, after obtaining the outer point cloud projection map of the keyhole, the shape curve of the keyhole region can be fitted based on the outer point cloud projection map to obtain a mathematical equation that can be used to characterize the shape of the keyhole region. Then, the keyhole region can be determined according to the fitted mathematical equation.
[0052] Specifically, a Cartesian coordinate system can be first constructed on the outer point cloud projection map, with the origin of the coordinate system being the projection point of the laser emitter's center point. Further, it can be scaled according to a preset scale. The grid is divided on the outer point cloud projection map of the established coordinate system, and finally the fitting parameters of the keyhole region are determined based on the grid parameters of each grid.
[0053] It should be noted that the preset scale refers to the side length of a single grid. The preset scale can be set by the operator according to actual needs, and the value range of the preset scale can be 10% to 20% of the maximum width at the edge of the keyhole along the welding direction. For example, if the maximum width at the edge of the keyhole is 10mm, the preset scale can be set to 1×1mm. In addition, the fitting parameters of the keyhole region refer to the fitting equation obtained by fitting the outer contour of the keyhole region, which characterizes the outer contour of the keyhole region. The parameters of the fitting equation are the fitting parameters described in this application.
[0054] In this embodiment, after one fitting process of the keyhole region, a preset scale can be reset. And according to the newly set preset scale The mesh is re-divided on the outer point cloud projection map. After re-dividing the mesh at different preset scales, the fitting parameters for a keyhole region are determined again based on the mesh parameters of each mesh. This process can be repeated continuously, changing the preset scale multiple times to divide the mesh and performing multiple keyhole region fitting processes, resulting in multiple sets of fitting parameters. Finally, the optimal set of fitting parameters can be selected as the fitting equation parameters characterizing the shape of the keyhole region.
[0055] In step 230, after obtaining the fitting parameters of the keyhole region, a mathematical equation composed of the fitting parameters can be used to characterize the shape of the keyhole region. In this embodiment, by fitting the keyhole region and finding the keyhole region in the outer point cloud projection map, the optical low-coherence imaging quality evaluation result can be obtained based on the point cloud data falling within the keyhole region and the area of the keyhole region.
[0056] Specifically, this embodiment primarily uses three parameters—point cloud density, point cloud signal strength, and center concentration—in the keyhole region as evaluation indicators for optical low-coherence imaging quality. Therefore, the point cloud data in the keyhole region can include these three parameters. Finally, a preset quality evaluation formula is used to calculate the optical low-coherence imaging quality score, which serves as the imaging quality evaluation result. If the point cloud density in the keyhole region is low, or the point cloud signal strength is weak, or the point cloud center concentration is low, it can be considered that the welding process parameters are improperly set, the sensor installation error is large, the sensor is incompatible, or the reflectivity of the welded object is low. In this case, the accuracy of the key depth information acquired by the optical low-coherence imaging monitoring device is low. Therefore, the operator can adjust the optical low-coherence imaging monitoring device and re-acquire the key depth information.
[0057] Figure 3 is a second flowchart of the low coherence imaging quality evaluation method according to an embodiment of this application. As shown in Figure 3, in some embodiments, step 210, which obtains the peripheral point cloud projection map of the keyhole based on point cloud data, may include the following steps:
[0058] Step 310: Use a preset segmentation formula to segment the point cloud data to obtain the outer point cloud data set and the keyhole point cloud data set.
[0059] Step 320: Project the peripheral point cloud data set onto the preset welding surface to obtain the peripheral point cloud projection map.
[0060] In the description of the foregoing embodiments, the point cloud data related to the keyhole during laser welding obtained by the processor in the optical low-coherence imaging monitoring device mainly includes point cloud data of the keyhole space and point cloud data of the keyhole's peripheral space (i.e., point cloud data of the welding plane space). Figure 4 is a keyhole-related point cloud distribution diagram of a specific example of this application. As can be seen from Figures 1 and 4, the point cloud data collected by the optical low-coherence imaging monitoring device is mainly distributed in the keyhole region and the region between the keyhole and the laser emitter, and the point clouds in these two regions can be clearly distinguished spatially. The embodiments of this application mainly segment the point cloud data by using the coordinates of each point in the point cloud data to obtain the peripheral point cloud dataset and the keyhole point cloud dataset.
[0061] Referring again to Figure 4, specifically, a three-dimensional coordinate system can be established with the center point of the laser emitter as the origin O. The positive X-axis of the three-dimensional coordinate system represents the welding direction at the current position, the positive Y-axis represents the direction towards the keyhole, and the positive Z-axis can be any direction perpendicular to the plane formed by the X and Y axes. The processor of the optical low-coherence imaging monitoring device can obtain the required point cloud data based on this three-dimensional coordinate system. The point cloud data includes the three-dimensional coordinates of all points in the point cloud. It should be noted that the welding positions corresponding to different sets of point cloud data may be different. Connecting all welding positions from front to back can form the aforementioned welding path. The current position is the welding position corresponding to the currently selected set of point cloud data.
[0062] Furthermore, as can be seen from Figure 4, the peripheral point cloud (the horizontal strip point cloud distribution in the upper part of Figure 4) and the keyhole point cloud (the inverted triangle point cloud distribution in the lower part of Figure 4) can be distinguished by the Y value. Based on this, the embodiments of this application can pre-construct a segmentation formula (i.e., a preset segmentation formula) to segment the point cloud data.
[0063] In some implementations, the preset segmentation formula can be:
[0064]
[0065] In the formula, For the point cloud data of the j-th point before segmentation, For the peripheral point cloud data set, For keyhole point cloud data set, Let be the perpendicular distance between the j-th point and the center point of the laser emitter. The vertical distance from the center point of the laser emitter to the preset welding surface. This is a preset distance threshold. It should be noted that the preset welding surface can be defined as: a distance of [value missing] from the origin O. The plane whose normal vector is parallel to the Y-axis; in addition, the preset distance threshold can be set manually by the staff according to actual needs, and the preset distance threshold is not limited here.
[0066] Therefore, the Y-coordinate in the point cloud can be less than or equal to + The point cloud data corresponding to the points is segmented into the outer point cloud data set, and the point cloud with a Y coordinate greater than 0 is segmented into the outer point cloud data set. + The point cloud data corresponding to the points is segmented into the keyhole point cloud data set, thus realizing the segmentation of point cloud data.
[0067] After obtaining the peripheral point cloud dataset, the point cloud data can be projected onto a preset welding surface based on the 3D coordinates of each point in the dataset to obtain a peripheral point cloud projection map. Specifically, since the preset welding surface is a plane parallel to the plane formed by the X and Z axes, in short, the Y values of points on the preset welding surface are the same. Therefore, the Y values of all points in the peripheral point cloud dataset can be removed or set to the Y values corresponding to the preset welding surface, thus converting the 3D coordinates of each point in the peripheral point cloud dataset into 2D coordinates to obtain the desired peripheral point cloud projection map.
[0068] Figure 5 is a projection of the peripheral point cloud of a specific example of this application. As shown in Figure 5, the peripheral point cloud in the projection is shaped like an elliptical ring, and the central region of the elliptical ring is the keyhole region. Therefore, the outline of the keyhole region is also elliptical, and the mathematical equation parameters obtained when fitting the keyhole region in the subsequent process are also elliptical equation parameters.
[0069] Figure 6 is a flowchart of the low coherence imaging quality evaluation method according to an embodiment of this application. As shown in Figure 6, in some embodiments, the grid parameters may include the number of point clouds within the grid, the grid area, and the distance between the grid and the central grid, where the central grid is the grid containing the origin of the coordinate system; step 220, determining the fitting parameters of the keyhole region based on the grid parameters of each grid, may include the following steps:
[0070] Step 610: For any grid, determine the grid evaluation parameters using a preset index formula and based on the number of point clouds within the grid, the grid area, and the distance between the grid and the central grid.
[0071] Step 620: Determine multiple target grids using a preset screening algorithm and based on the grid evaluation parameters of each grid.
[0072] Step 630: Input the point cloud data of the center points of each target grid into the Hough transform algorithm to obtain the fitting parameters of the keyhole region.
[0073] In this embodiment, after obtaining the projection map of the peripheral point cloud, the projection point of the laser emitter's center point (i.e., the center point of the peripheral point cloud) can be used as the origin O. proj Construct an X proj Y proj The coordinate system (as shown in Figure 5) is based on X. proj Y proj The coordinate system can obtain the two-dimensional coordinates (X, Y, X) of each point in the outer point cloud. proj Y proj After establishing the coordinate system, it can be used according to the preset scale. Divide the outer point cloud projection map into a grid, and make the center point of one of the grids coincide with the origin O. proj They overlap, and this grid can be defined as the center grid.
[0074] Figure 7 is a schematic diagram of mesh division in a specific example of this application. As shown in Figure 7, after the outer point cloud projection map is meshed, each point in the outer point cloud is distributed in each mesh. The number of points in the mesh on the outer point cloud is relatively large, while the number of points in the mesh in the middle and outside of the outer point cloud is relatively small.
[0075] Based on this, the number of point clouds within each grid can be determined according to the two-dimensional coordinates of each point, the coordinates of the grid center point, and the preset scale; the grid area can be directly calculated using the preset scale. The grid area is The distance between the grid and the center grid can be calculated using the coordinates of the center point of the grid and the center point of the center grid.
[0076] Furthermore, after dividing the grid according to a preset scale, the grid evaluation parameters can be determined using a preset index formula based on the number of point clouds within the grid, the grid area, and the distance between the grid and the central grid. It should be noted that the grid evaluation parameters are comprehensive evaluation indicators used to characterize the positional relationship between the grid and the keyhole region. In this embodiment, the comprehensive evaluation index of the grid is mainly determined from two aspects: the point cloud density within the grid and the distance between the grid and the central grid. It can be understood that when the point cloud density within the grid is high and the distance between the grid and the central grid is far, it indicates that the grid may be located in the outer point cloud area; when the point cloud density within the grid is low and the distance between the grid and the central grid is close, it indicates that the grid may be located in the keyhole region. Therefore, the relative positional relationship between the grid and the keyhole region can be characterized by these two indicators: the point cloud density within the grid and the distance between the grid and the central grid.
[0077] Taking a single grid as an example, the number of point clouds within that grid, its area, and the distance between the grid and the central grid can be substituted into a preset index formula to obtain the grid evaluation parameters for that grid. The grid evaluation parameters for all grids are calculated in the same way, and then the target grid is selected based on these parameters.
[0078] In some implementations, the preset index formula can be:
[0079]
[0080] In the formula, Let i be the grid evaluation parameters for the i-th grid. Let i be the number of point clouds in the i-th grid. Let i be the area of the i-th grid. Let be the distance between the center point of the i-th grid and the center point of the central grid. A smaller value indicates a lower point cloud density within the grid and a closer proximity to the center grid. A larger value indicates a higher grid density and a greater distance from the center grid.
[0081] After obtaining the grid evaluation parameters for all grids, a preset filtering algorithm can be used to determine multiple target grids based on the grid evaluation parameters of each grid. It should be noted that the preset filtering algorithm can be a percentage filtering method; the selected target grids are mainly those located at the boundary between the keyhole region and the surrounding area, i.e., the grids containing the keyhole region's outline. Only through the point cloud data within these target grids can the mathematical equation characterizing the keyhole region's outline be better fitted, thus enabling a more accurate determination of the keyhole region.
[0082] In some implementations, step 620, which uses a preset filtering algorithm and based on the grid evaluation parameters of each grid, determines multiple target grids. This may include: sorting the grid evaluation parameters in descending order; selecting grid evaluation parameters within a preset percentage range from the side with smaller grid evaluation parameters using the preset filtering algorithm; and using the grids corresponding to the selected grid evaluation parameters as target grids.
[0083] Specifically, all grid evaluation parameters can be sorted from largest to smallest, and then a percentage filtering method can be used to select grid evaluation parameters within a preset percentage range from the smaller end of the parameter range. It should be noted that after sorting all grid evaluation parameters, the position of each parameter in the sequence can be expressed as a percentage. In this embodiment, the preset percentage range can be set to 8%–12%. The preset percentage range can be specifically set by the operator based on the specifications of the peripheral point cloud projection map and the distribution of the peripheral point cloud; no specific restrictions are imposed here.
[0084] As an example, after sorting all grid evaluation parameters in descending order, the percentage filtering method can be used to select the grid evaluation parameters that are at the end of this sequence and fall within the 8% to 12% range. The grids corresponding to these grid evaluation parameters are then identified as target grids. Target grids have relatively low location cloud density and are relatively closer to the center grid.
[0085] As another example, after sorting all grid evaluation parameters in descending order, the percentage filtering method can be used to select the grid evaluation parameters at the end of the sequence, and the grids corresponding to these grid evaluation parameters can be determined as the target grids.
[0086] Furthermore, the point cloud data of the center points of each target grid are then input into the Hough transform algorithm to obtain the fitting parameters for the keyhole region. Since the keyhole region is elliptical from a top-down view (as shown in Figure 7), the fitting curve is also elliptical, and the fitting equation is an elliptical equation. Specifically, point cloud data of the center points of each target grid can be randomly selected to inversely calculate the parameters of the elliptical equation, obtaining a set of fitting parameters for the elliptical equation. Then, respectively using... The mesh is redefined using a preset scale, and the steps for determining the fitting parameters described above are repeated. Finally, n sets of fitting parameters for the elliptic equations are obtained. It should be noted that n is the number of iterations, and the value of n can be manually set by the staff according to actual needs. The larger n is, the more accurate the fitting parameters for the keyhole region will be.
[0087] For the n sets of fitting parameters obtained above, this embodiment can use a hard voting method to determine the optimal fitting parameters. The optimal fitting parameters constitute an elliptical equation for characterizing the shape of the keyhole region. The hard voting rule is as follows: if all n sets of fitting parameters are inconsistent, the average value can be taken as the optimal fitting parameter; if multiple sets of fitting equation parameters are consistent, the mode result can be taken as the optimal fitting parameter.
[0088] As an example, suppose there are 99 sets of fitting parameters. Among them, the fitting parameters of the 2nd and 3rd sets are the same, the fitting parameters of the 30th, 31st and 32nd sets are the same, and the fitting parameters of the 96th, 97th, 98th and 99th sets are the same. Then the fitting parameters of the 96th, 97th, 98th and 99th sets are selected as the mode and used as the final fitting parameters for the keyhole area.
[0089] Once the fitting parameters for the keyhole region are determined, the keyhole region can be determined according to the fitting equation. Furthermore, the optical low-coherence imaging quality can be evaluated based on the relevant data of the fitted keyhole region.
[0090] In some implementations, the point cloud data within the keyhole area may include the number of point clouds in the keyhole area, the signal strength of each point within the keyhole area, and the coordinate data of each point. Step 230, determining the imaging quality evaluation result based on the point cloud data within the keyhole area and the area of the keyhole area, may include: using a preset quality evaluation formula and based on the number of point clouds in the keyhole area, the signal strength of each point within the keyhole area, the coordinate data of each point, the coordinate data of the center point of the keyhole area, and the area of the keyhole area, to determine the imaging quality evaluation result.
[0091] This application's embodiments primarily evaluate the optical low-coherence imaging quality of the keyhole region from three perspectives: point cloud density, signal strength, and center concentration. Specifically, point cloud density is the ratio between the number of points in the keyhole region in the peripheral point cloud projection map and the area of the keyhole region; point cloud signal strength is the signal strength returned to the optical low-coherence imaging monitoring device; and center concentration is the degree of dispersion between the point cloud data and the center point of the fitted region.
[0092] It is understandable that the higher the point cloud density in the keyhole area, the stronger the point cloud signal intensity, or the higher the point cloud center concentration, the higher the quality of the acquired point cloud data and the higher the quality of optical low coherence imaging.
[0093] Based on this, in some implementations, the following preset quality evaluation formula can be constructed to quantitatively evaluate the quality of low-coherence optical imaging:
[0094]
[0095] In the formula, For the image quality evaluation results, This represents the area of the keyhole region. The number of point clouds in the keyhole area. Let Z be the signal strength at point z in the i-th grid within the keyhole region. The total signal strength within the keyhole area. This represents the coordinate data of the z-th point in the i-th grid within the keyhole area. The coordinates of the center point of the keyhole area. This represents the sum of distances from the point cloud within the fitted region to the center point of the fitted region. It should be noted that... as well as All are X proj O proj Y proj Coordinate data in a coordinate system.
[0096] In this embodiment, the number of point clouds in the keyhole region, the signal strength of each point in the keyhole region, the coordinate data of each point, the coordinate data of the center point of the keyhole region, and the area of the keyhole region can be substituted into the above-mentioned preset quality evaluation formula to obtain the imaging quality evaluation result. , A smaller value indicates lower quality in low-coherence imaging. The larger the value, the higher the quality of low-coherence imaging.
[0097] Therefore, a quantitative evaluation of the imaging quality of low-coherence optical images can be achieved, and the imaging quality evaluation results can be obtained. The level of image quality directly reflects the quality of the image, thus improving the accuracy of online monitoring of laser welding keyhole depth. Secondly, by calculating the grid evaluation parameters based on the point cloud density of the grid within the keyhole area and the distance between the grid and the central grid, a percentage filtering method is used to select target grids with relatively low point cloud density and relatively close distance to the central grid. The center point of the selected target grid is then substituted into the Hough transform to obtain the fitting equation for the keyhole area, which greatly improves the computational efficiency of the Hough transform and enables rapid determination of the laser welding keyhole position. Finally, the obtained image quality evaluation results can serve as a basis for investigating negative factors such as improper welding process parameter settings, large sensor installation errors, sensor incompatibility, or low reflectivity of the welded object, thus improving the reliability of online monitoring of laser welding keyhole depth.
[0098] Based on the above embodiments, this application also provides a low-coherence imaging quality evaluation device. Figure 8 is a schematic diagram of the structure of the low-coherence imaging quality evaluation device according to an embodiment of this application. As shown in Figure 8, the low-coherence imaging quality evaluation device 800 may include an acquisition module 810, a fitting module 820, and an evaluation module 830.
[0099] The acquisition module 810 is used to acquire point cloud data related to the keyhole during laser welding and obtain a projection map of the outer point cloud of the keyhole based on the point cloud data; the fitting module 820 is used to perform the keyhole region fitting process multiple times to determine the keyhole region based on the fitting parameters obtained in each fitting process; wherein, in each fitting process, a coordinate system is constructed based on the outer point cloud projection map and a grid is divided according to a preset scale, and the fitting parameters of the keyhole region are determined based on the grid parameters of each grid; the preset scale is different in different fitting processes; the evaluation module 830 is used to determine the imaging quality evaluation result based on the point cloud data in the keyhole region and the area of the keyhole region.
[0100] Therefore, the acquisition module 810 acquires point cloud data related to the keyhole during laser welding and obtains the outer point cloud projection map of the keyhole based on the point cloud data; the fitting module 820 then performs the keyhole region fitting process multiple times at different preset scales. In each fitting process, a coordinate system is constructed based on the outer point cloud projection map and a grid is divided according to the preset scale. The fitting parameters of the keyhole region are determined based on the grid parameters of each grid, and the keyhole region is fitted based on the fitting parameters obtained in each fitting process; finally, the evaluation module 830 determines the imaging quality evaluation result based on the point cloud data and the area of the keyhole region, thereby realizing the evaluation of the accuracy of key depth information acquired by optical low coherence imaging.
[0101] In some implementations, the acquisition module 810 is specifically used to: segment the point cloud data using a preset segmentation formula to obtain a peripheral point cloud data set and a keyhole point cloud data set; and project the peripheral point cloud data set onto a preset welding surface to obtain a peripheral point cloud projection map.
[0102] In some implementations, the preset segmentation formula is:
[0103]
[0104] In the formula, For the point cloud data of the j-th point before segmentation, For the peripheral point cloud data set, For keyhole point cloud data set, Let be the perpendicular distance between the j-th point and the center point of the laser emitter. The vertical distance from the center point of the laser emitter to the preset welding surface. This is a preset distance threshold.
[0105] In some implementations, the grid parameters include the number of point clouds within the grid, the grid area, and the distance between the grid and the central grid, where the central grid is the grid containing the origin of the coordinate system. The fitting module 820 is specifically used to: determine the grid evaluation parameters for any grid using a preset index formula and based on the number of point clouds within the grid, the grid area, and the distance between the grid and the central grid; determine multiple target grids using a preset screening algorithm and based on the grid evaluation parameters of each grid; and input the point cloud data of the center point of each target grid into the Hough transform algorithm to obtain the fitting parameters of the keyhole region.
[0106] In some implementations, the fitting module 820 is further specifically used to: sort the grid evaluation parameters in descending order; select grid evaluation parameters within a preset percentage range from the side with smaller grid evaluation parameters using a preset screening algorithm, and use the grid corresponding to the selected grid evaluation parameter as the target grid.
[0107] In some implementations, the preset index formula is:
[0108]
[0109] In the formula, Let i be the grid evaluation parameters for the i-th grid. Let i be the number of point clouds in the i-th grid. Let i be the area of the i-th grid. Let be the distance between the center point of the i-th grid and the center point of the central grid.
[0110] In some implementations, the point cloud data within the keyhole area includes the number of point clouds in the keyhole area, the signal strength of each point within the keyhole area, and the coordinate data of each point; the evaluation module 830 is specifically used to: determine the imaging quality evaluation result by using a preset quality evaluation formula and based on the number of point clouds in the keyhole area, the signal strength of each point within the keyhole area, the coordinate data of each point, the coordinate data of the center point of the keyhole area, and the area of the keyhole area.
[0111] In some implementations, the preset quality evaluation formula is:
[0112]
[0113] In the formula, For the image quality evaluation results, This represents the area of the keyhole region. The number of point clouds in the keyhole area. Let Z be the signal strength at point z in the i-th grid within the keyhole region. The total signal strength within the keyhole area. This represents the coordinate data of the z-th point in the i-th grid within the keyhole area. This refers to the coordinates of the center point of the keyhole area.
[0114] It should be noted that for details not disclosed in the low coherence imaging quality evaluation device of this embodiment, please refer to the details disclosed in the embodiments of the low coherence imaging quality evaluation method in this specification, which will not be repeated here.
[0115] Based on the above embodiments, Figure 9 illustrates a schematic diagram of the physical structure of an electronic device. As shown in Figure 9, the electronic device may include: a processor 910, a communication interface 920, a memory 930, and a communication bus 940. The processor 910, communication interface 920, and memory 930 communicate with each other via the communication bus 940. The processor 910 can call logical instructions in the memory 930 to execute a low-coherence imaging quality evaluation method. This method includes: acquiring point cloud data related to the keyhole during laser welding, and obtaining a projection map of the outer point cloud of the keyhole based on the point cloud data; performing a keyhole region fitting process multiple times to determine the keyhole region based on the fitting parameters obtained in each fitting process; wherein, in each fitting process, a coordinate system is constructed based on the outer point cloud projection map and a grid is divided according to a preset scale, and the fitting parameters of the keyhole region are determined based on the grid parameters of each grid; the preset scale is different in different fitting processes; and the imaging quality evaluation result is determined based on the point cloud data within the keyhole region and the area of the keyhole region.
[0116] Furthermore, the logical instructions in the aforementioned memory 930 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0117] Based on the above embodiments, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the low-coherence imaging quality evaluation method provided by the above methods. The method includes: acquiring point cloud data related to the keyhole during laser welding, and obtaining a peripheral point cloud projection map of the keyhole based on the point cloud data; performing a keyhole region fitting process multiple times to determine the keyhole region based on the fitting parameters obtained in each fitting process; wherein, in each fitting process, a coordinate system is constructed based on the peripheral point cloud projection map and a grid is divided according to a preset scale, and the fitting parameters of the keyhole region are determined based on the grid parameters of each grid; the preset scale is different in different fitting processes; and the imaging quality evaluation result is determined based on the point cloud data in the keyhole region and the area of the keyhole region.
[0118] Based on the above embodiments, in another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the low-coherence imaging quality evaluation method provided by the above methods. The method includes: acquiring point cloud data related to the keyhole during laser welding, and obtaining a peripheral point cloud projection map of the keyhole based on the point cloud data; performing a keyhole region fitting process multiple times to determine the keyhole region based on the fitting parameters obtained in each fitting process; wherein, in each fitting process, a coordinate system is constructed based on the peripheral point cloud projection map and a grid is divided according to a preset scale, and the fitting parameters of the keyhole region are determined based on the grid parameters of each grid; the preset scale is different in different fitting processes; and the imaging quality evaluation result is determined based on the point cloud data in the keyhole region and the area of the keyhole region.
[0119] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.
[0120] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.
Claims
1. A method for evaluating the quality of low-coherence imaging, characterized in that, include: Acquire point cloud data related to the keyhole during laser welding, and obtain the peripheral point cloud projection map of the keyhole based on the point cloud data; The keyhole region fitting process is performed multiple times to determine the keyhole region based on the fitting parameters obtained in each fitting process; wherein, in each fitting process, a coordinate system is constructed based on the outer point cloud projection map and a grid is divided according to a preset scale, and the fitting parameters of the keyhole region are determined based on the grid parameters of each grid; the preset scale is different in different fitting processes; The imaging quality evaluation result is determined based on the point cloud data within the keyhole area and the area of the keyhole region.
2. The method for evaluating low-coherence imaging quality according to claim 1, characterized in that, The step of obtaining the peripheral point cloud projection map of the keyhole based on the point cloud data includes: The point cloud data is segmented using a preset segmentation formula to obtain a peripheral point cloud data set and a keyhole point cloud data set. The peripheral point cloud data set is projected onto a preset welding surface to obtain the peripheral point cloud projection map.
3. The method for evaluating low-coherence imaging quality according to claim 2, characterized in that, The preset segmentation formula is: In the formula, For the point cloud data of the j-th point before segmentation, The peripheral point cloud data set, The keyhole point cloud data set, Let be the perpendicular distance between the j-th point and the center point of the laser emitter. The vertical distance from the center point of the laser emitter to the preset welding surface. This is a preset distance threshold.
4. The method for evaluating low-coherence imaging quality according to claim 1, characterized in that, The grid parameters include the number of point clouds within the grid, the grid area, and the distance between the grid and the central grid, where the central grid is the grid containing the origin of the coordinate system. The process of determining the fitting parameters for the keyhole region based on the grid parameters of each of the grids includes: For any given grid, a grid evaluation parameter is determined using a preset index formula and based on the number of point clouds within the grid, the grid area, and the distance between the grid and the central grid. Multiple target grids are determined by using a preset screening algorithm and based on the grid evaluation parameters of each grid. Substitute the point cloud data of the center points of each target grid into the Hough transform algorithm to obtain the fitting parameters of the keyhole region.
5. The method for evaluating low-coherence imaging quality according to claim 4, characterized in that, The process of determining multiple target grids using a preset screening algorithm and based on the grid evaluation parameters of each grid includes: The grid evaluation parameters are sorted in descending order of size. The preset filtering algorithm is used to select grid evaluation parameters within a preset percentage range from the side with smaller grid evaluation parameters, and the grid corresponding to the selected grid evaluation parameters is taken as the target grid.
6. The method for evaluating low-coherence imaging quality according to claim 4, characterized in that, The preset index formula is: In the formula, Let i be the grid evaluation parameters for the i-th grid. Let i be the number of point clouds in the i-th grid. Let i be the area of the i-th grid. The distance between the i-th grid center point and the center grid center point is given.
7. The method for evaluating low-coherence imaging quality according to claim 1, characterized in that, The point cloud data within the keyhole area includes the number of point clouds in the keyhole area, the signal strength of each point within the keyhole area, and the coordinate data of each point. The determination of the imaging quality evaluation result based on the point cloud data within the keyhole area and the area of the keyhole area includes: The imaging quality evaluation result is determined by using a preset quality evaluation formula and based on the number of point clouds in the keyhole region, the signal strength of each point in the keyhole region, the coordinate data of each point, the coordinate data of the center point of the keyhole region, and the area of the keyhole region.
8. The method for evaluating low-coherence imaging quality according to claim 7, characterized in that, The preset quality evaluation formula is: In the formula, The imaging quality evaluation result is as follows. The area of the keyhole region is [area]. The number of point clouds in the keyhole region. The signal strength at the z-th point in the i-th grid within the keyhole region. The total signal strength within the keyhole area. The coordinate data of the z-th point in the i-th grid within the keyhole area. The coordinates of the center point of the keyhole area.
9. A low-coherence imaging quality evaluation device, characterized in that, include: The acquisition module is used to acquire point cloud data related to the keyhole during the laser welding process, and to obtain the peripheral point cloud projection map of the keyhole based on the point cloud data. The fitting module is used to perform the keyhole region fitting process multiple times to determine the keyhole region based on the fitting parameters obtained in each fitting process; wherein, in each fitting process, a coordinate system is constructed based on the outer point cloud projection map and a grid is divided according to a preset scale, and the fitting parameters of the keyhole region are determined based on the grid parameters of each grid; the preset scale is different in different fitting processes; The evaluation module is used to determine the imaging quality evaluation result based on the point cloud data within the keyhole area and the area of the keyhole area.
10. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the low coherence imaging quality evaluation method as described in any one of claims 1 to 8.