Method and System for Extracting the Width of a Laser Metal Directed Energy Deposition Molten Pool Based on Edge Erosion

Abstract

The present invention provides a method and system for extracting the width of a laser metal directed energy deposition molten pool based on edge corrosion. First, the original molten pool image is grayscaled to generate a grayscale image of the molten pool; then, binarization processing is performed to obtain a binarized molten pool image; an edge detection algorithm is used to traverse the binarized molten pool image to extract image edge points, denoted as the contour point pixel set (x k , y k ); an image erosion convolution algorithm is used to perform erosion operation on the binarized molten pool image, and the last eroded white pixel points are retained; the pixel point boundaries are connected, and based on the largest inscribed circle of the convex hull within the boundary range and the minimum distance from the largest inscribed circle to the edge of the binarized molten pool image, the width of the molten pool is determined. Through the method of the present invention, the width information of the molten pool in the metal laser directed energy deposition process can be quickly and accurately extracted, and its geometric features can be accurately and real-time obtained, which is beneficial to controlling the morphology and shape of the molten pool and improving the quality of laser cladding parts.

Description

Technical Field

[0001] The present invention relates to the technical field of metal laser additive manufacturing, especially to the extraction of the molten pool width therein, and specifically relates to a method and system for extracting the width of a laser metal direct energy deposition molten pool based on edge corrosion. Background Art

[0002] During the process of laser metal direct energy deposition, the stability of the molten pool represents the stability of the forming process. Controlling the morphology of the molten pool is a key factor in ensuring the quality of laser cladding parts. During the real-time monitoring of the molten pool and the printing process, accurately extracting the edge of the molten pool is a prerequisite for extracting the geometric information of the molten pool. The width information of the molten pool can reflect the characteristics of the molten pool more than other information. Therefore, it is very important to monitor the width of the molten pool during the laser metal direct energy deposition process.

[0003] In the prior art, algorithms based on machine learning are tried to be used to detect the width of the molten pool. For example, according to historical molten pool images for annotation and training, a network model for detecting the size of the molten pool is trained, such as a convolutional neural network model, a segmentation network model, etc., and real-time detection output is performed on the input molten pool image, such as the detection methods proposed in CN115170545A and CN113554587A. We can see that when using a network model to detect the geometric size of the molten pool, on the one hand, it is limited by the training data and the accuracy of the model. The morphology, size, and characteristics presented by the molten pool are diverse and show great randomness, resulting in great obstacles in data annotation and feature analysis. How to prepare the annotation is a difficult problem. On the other hand, it is restricted by the clarity, complexity, and angle problems of the image caused by interference such as splashing and arc light during on-site detection of the molten pool, resulting in large errors in image recognition and size recognition.

[0004] Another way to detect the geometric size of the molten pool in the prior art is based on image processing. For example, in the patent application disclosed in CN115018816A, a method for identifying and extracting the molten pool based on median filtering and threshold segmentation is proposed, and then the mean value calculation method is used to process the grayscale image with clear edges to obtain the molten pool width data. In the patent application disclosed in CN115187567A, based on obtaining the high-pixel-value area image and the low-pixel-value area image of the molten pool after binarization of the molten pool image, the center point of the bright area of the molten pool and the center point of the dark area of the molten pool are calculated, and the direction angle and direction line of the molten pool are calculated according to the center points. Then, the direction perpendicular line of the molten pool is calculated according to the direction line of the molten pool, and combined with the segmented image of the molten pool, the width of the molten pool is calculated.

[0005] Prior Art:

[0006] Patent Document 1: CN115170545A A Method for Detecting Dynamic Molten Pool Size and Determining Forming Direction

[0007] Patent Document 2: CN113554587A A Method and System for Extracting Geometric Features of Molten Pool Images Based on Deep Learning

[0008] Patent Document 3: CN115018816A An Image Processing Method, Device, Equipment and Storage Medium for Real-Time Detection of Molten Pool Width

[0009] Patent Document 4: CN115187567A A Method for Detecting Molten Pool Forming Direction and Width in Metal Additive Manufacturing Summary of the Invention

[0010] The object of the present invention is to provide a method and system for extracting the width of a laser metal direct energy deposition molten pool based on edge erosion, eliminating the influence of noise such as molten pool splashing and burrs, relatively accurately calculating the width of the molten pool, providing accurate support for the shape and size control of the molten pool, and ensuring the quality of laser cladding processed parts.

[0011] According to the first aspect of the object of the present invention, a method for extracting the width of a laser metal direct energy deposition molten pool based on edge erosion is proposed, including the following steps:

[0012] Perform grayscale processing on the original molten pool image to generate a grayscale image of the molten pool;

[0013] Perform binarization processing on the grayscale image of the molten pool to obtain a binarized molten pool image;

[0014] Use an edge detection algorithm to traverse the binarized molten pool image, extract the image edge points, denoted as the contour point pixel set (x k , y k ), k = 1, 2,..n, and calculate the grayscale value of the single-channel image pixel points;

[0015] Use an image erosion convolution algorithm to perform erosion operation on the binarized molten pool image, and retain the white pixel points of the last erosion;

[0016] Connect the pixel point boundaries, and determine the molten pool width based on the largest inscribed circle of the convex hull within the boundary and the minimum distance from the largest inscribed circle to the edge of the binarized molten pool image.

[0017] As an optional implementation manner, the use of an image erosion convolution algorithm to perform erosion operation on the binarized molten pool image and retain the white pixel points of the last erosion includes:

[0018] Perform convolution on the binarized molten pool image using a preset convolution kernel, and the convolution kernel is defined as follows: 1 1 1 1 1 1 1 1 1

[0022] Then, use a preset convolutional kernel to perform convolution on the binarized molten pool image. During the process of using the convolutional kernel to perform convolution on the binarized molten pool image, when the convolution result y' of the target area is less than 255*8, the target value y'' of the target area is set to 0:

[0023]

[0024] Repeat the convolution operation until the target values of all pixel points become 0. Record the target area of the last convolution, and denote the set of edge points of the target area as (m k , n k ).

[0025] As an optional implementation manner, determining the molten pool width based on the largest inscribed circle of the convex hull within the boundary range and the minimum distance from the largest inscribed circle to the edge of the binarized molten pool image includes:

[0026] First, calculate the radius of the largest inscribed circle within the set of edge points (m k , n k ) of the target area. Denote the radius of the largest inscribed circle as R1;

[0027] Then, calculate the minimum distance from the points on the largest inscribed circle to the set of contour point pixels (x k , y k ), and denote it as R2;

[0028] Finally, calculate the molten pool width D:

[0029] D = 2 * (R1 + R2).

[0030] As an optional implementation manner, calculating the radius of the largest inscribed circle within the set of edge points (m k , n k ) of the target area includes:

[0031] Within the target area after the last convolution, find a point that satisfies: the sum of the distances from this point to all edge points of the target area after the last convolution operation is the smallest. Then, take this point as the center of the circle, and determine the largest inscribed circle and its radius R1 by calculating the largest inscribed circle of the triangle.

[0032] According to the second aspect of the object of the present invention, a system for extracting the molten pool width of laser metal directed energy deposition based on edge erosion is also proposed, including:

[0033] One or more processors;

[0034] A memory stores instructions that can be operated, and when the instructions are executed by the one or more processors, the one or more processors are caused to perform operations, including the process of the method for extracting the width of a laser metal directed energy deposition molten pool based on edge erosion as described above.

[0035] According to a third aspect of the object of the present invention, there is also provided a computer-readable medium storing software, the software including instructions executable by one or more computers, and when the instructions are executed, the one or more computers are caused to perform operations, including the process of the method for extracting the width of a laser metal directed energy deposition molten pool based on edge erosion as described above.

[0036] Thus, through the above-mentioned image processing and continuous convolutional erosion proposed by the present invention, the width information of the molten pool in the metal laser directed energy deposition process can be quickly and accurately extracted, and its geometric features can be accurately obtained in real time, which is beneficial to controlling the morphology and shape of the molten pool and improving the quality of laser cladding parts.

[0037] It should be understood that all combinations of the foregoing concepts and additional concepts described in more detail below are considered to be part of the inventive subject matter of the present disclosure as long as such concepts do not conflict with each other. In addition, all combinations of the claimed subject matter are considered to be part of the inventive subject matter of the present disclosure.

[0038] The foregoing and other aspects, embodiments, and features of the teachings of the present invention can be more fully understood from the following description in conjunction with the accompanying drawings. Other additional aspects of the present invention, such as the features and / or beneficial effects of exemplary embodiments, will be apparent from the following description or will be learned through the practice of specific embodiments according to the teachings of the present invention. Description of the Drawings

[0039] Figure 1 is a flowchart of the method for extracting the width of a laser metal directed energy deposition molten pool based on edge erosion according to an embodiment of the present invention.

[0040] Figure 2 is a schematic diagram of pixel changes in the corrosion convolution of the method for extracting the width of a laser metal directed energy deposition molten pool based on edge erosion according to an embodiment of the present invention.

[0041] Figures 3a - 3d is an exemplary process diagram of the method for extracting the width of a laser metal directed energy deposition molten pool based on edge erosion according to an embodiment of the present invention, where 3a - 3d respectively represent the processes of grayscale conversion, binarization, corrosion convolution, and molten pool width extraction.

[0042] Figure 4 is a schematic diagram of the principle of the method for extracting the width of a laser metal directed energy deposition molten pool based on edge erosion according to an embodiment of the present invention. Detailed Embodiments

[0043] To better understand the technical content of the present invention, specific embodiments are hereby given and described in conjunction with the accompanying drawings as follows.

[0044] In the present disclosure, aspects of the present invention are described with reference to the accompanying drawings, in which many illustrative embodiments are shown. The embodiments of the present disclosure are not necessarily intended to cover all aspects of the present invention. It should be understood that the various concepts and embodiments introduced above, as well as those concepts and embodiments described in more detail below, can be implemented in any of many ways, because the concepts and embodiments disclosed in the present invention are not limited to any implementation manner. Additionally, some aspects of the present invention can be used alone, or in any suitable combination with other aspects of the present invention.

[0045] Combined Figure 1 As shown in FIGS. 1, 3, and 4, the object of the present invention aims to address the problem that it is difficult to accurately extract the width of the molten pool due to the existence of interferences such as spatter, noise, and arc light. A monitoring system for the molten pool in the laser metal directed energy deposition process is proposed to achieve online monitoring of the molten pool, extract its geometric information, especially the width information of the molten pool, so as to reflect the characteristics of the molten pool in real time, facilitate controlling the shape of the molten pool, and ensure the quality of the laser cladding parts.

[0046] In an embodiment of the present invention, a method for extracting the width of the molten pool based on an image-based corrosion algorithm is proposed. Generally, first, the image of the molten pool is extracted; then, the binarized molten pool image is obtained by means of threshold segmentation; after that, the corrosion algorithm is used to erode the image, and the erosion operation is repeated, and the white pixel points of the last erosion are retained. Finally, the boundaries of the pixel points are connected, and the maximum inscribed circle of the convex hull within the boundary is calculated to obtain the radius R1 of the inscribed circle. Finally, the minimum distance R2 from the inscribed circle to the edge of the binarized image is calculated, and the sum of 2*(R1 + R2) is the width of the molten pool.

[0047] Thus, the width information of the molten pool can be extracted quickly and accurately.

[0048] Combined Figure 1 The exemplary figure shown represents a specific implementation process of the method for extracting the width of the laser metal directed energy deposition molten pool based on edge corrosion, including the following steps:

[0049] The original molten pool image is grayscale processed to generate a grayscale image of the molten pool;

[0050] The grayscale image of the molten pool is binarized to obtain a binarized molten pool image;

[0051] The binarized molten pool image is traversed using an edge detection algorithm to extract the image edge points, denoted as the contour point pixel set (x k , y k), k = 1, 2,..n, and calculate the grayscale values of the pixels in the single-channel image;

[0052] Use the image erosion convolution algorithm to perform erosion operation on the binary molten pool image, and retain the white pixels in the last erosion;

[0053] Connect the pixel boundaries, and determine the width of the molten pool based on the maximum inscribed circle of the convex hull within the boundary and the minimum distance from the maximum inscribed circle to the edge of the binary molten pool image.

[0054] Next, we combine Figure 2 and Figures 3a - 3d as shown below to further illustrate the implementation process of the foregoing method.

[0055] For the frame images extracted from the on-site molten pool monitoring video, through image processing, the color image data of the RGB channels can be obtained; then through grayscale processing and binary processing of the image, the finally binary-processed molten pool is obtained as follows, and convolution calculation is performed using the erosion algorithm of the present invention, as Figures 3a - 3d shown.

[0056] As an optional embodiment, the original molten pool image is grayscaled to generate a grayscale image of the molten pool, as Figure 3a shown, and it is grayscaled based on the following grayscale processing function:

[0057] y gray = (R + G + B) / 3

[0058] where R, G, and B are the grayscale values of the corresponding pixels in the red, green, and blue channel images of the original molten pool image, and y gray is the grayscale value of each pixel after grayscale processing.

[0059] In the embodiment of the present invention, the grayscale image of the molten pool is binary-processed to obtain a binary molten pool image, as Figure 3b shown.

[0060] Based on a preset binary threshold and according to the following binary function, the image is binary-processed:

[0061]

[0062] where, when the grayscale value y gray of a certain pixel is greater than the binary threshold a, the pixel value y is set to 255, otherwise the pixel value y is set to 0, thereby obtaining a binary molten pool image.

[0063] As Figure 3b shown in the binary image processing result, where the binary threshold a is taken as 94.

[0064] Combination Figure 2 As shown, an image erosion convolution algorithm is used to perform an erosion operation on the binarized molten pool image, and the last eroded white pixel points are retained, including:

[0065] Use a preset convolution kernel to perform convolution on the binarized molten pool image. During the process of using the convolution kernel to perform convolution on the binarized molten pool image, when the convolution result y' of the target area is less than 255*8, the target value y" of the target area is set to 0:

[0066]

[0067] Repeat the convolution operation until the target values y of all pixel points ” All become 0, record the last convolution target area, and the set of target area edge points is denoted as (m k ,n k ).

[0068] As an optional example, in the embodiment of the present invention, the convolution kernel is defined as follows: 1 1 1 1 1 1

[0071] 1 1 1.

[0072] Such as Figure 3c The schematic diagram of the process of the erosion convolution process shown.

[0073] In the embodiment of the present invention, based on the largest inscribed circle of the convex hull within the boundary range and the minimum distance from the largest inscribed circle to the edge of the binarized molten pool image, the width of the molten pool is determined, including:

[0074] First, calculate the radius of the largest inscribed circle within the set of target area edge points (m k ,n k ), and the radius of the largest inscribed circle is denoted as R1;

[0075] Then, calculate the minimum distance from the points on the largest inscribed circle to the set of contour point pixels (x k ,y k ), denoted as R2;

[0076] Finally, calculate the width D of the molten pool:

[0077] D = 2*(R1 + R2).

[0078] Combination Figure 4 As shown in the schematic diagram of the calculation principle, in the embodiment of the present invention, calculate the radius of the largest inscribed circle within the set of target area edge points (m k ,n k ), including:

[0079] Within the target area after the last convolution, find a point that satisfies the condition that the sum of the distances from this point to all the edge points of the target area after the last convolution operation is the smallest. Then, use this point as the center of a circle, and determine the largest inscribed circle and its radius R1 by calculating the largest inscribed circle of a triangle.

[0080] According to the embodiments disclosed in the present invention, a system for extracting the width of a laser metal directed energy deposition molten pool based on edge erosion is also proposed, including: one or more processors and a memory.

[0081] The aforementioned memory is configured to store executable instructions. When the instructions are executed by the one or more processors, the one or more processors perform operations, and the operations include the process of the molten pool width extraction method in the foregoing embodiments.

[0082] According to the embodiments disclosed in the present invention, a computer-readable medium storing software is also proposed. The software includes instructions that can be executed by one or more computers. When the instructions are executed, the one or more computers perform operations, and the operations include the process of the molten pool width extraction method in the foregoing embodiments.

[0083] Although the present invention has been disclosed above with preferred embodiments, it is not intended to limit the present invention. Those with ordinary knowledge in the technical field to which the present invention pertains can make various modifications and refinements without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be determined by what is defined in the claims.

Claims

1. A method for extracting the width of a laser metal direct energy deposition molten pool based on edge corrosion, characterized in that, It includes the following steps: Grayscale the original molten pool image to generate a grayscale image of the molten pool; Binarize the grayscale image of the molten pool to obtain a binarized molten pool image; Traverse the binarized molten pool image using an edge detection algorithm, extract the image edge points, denoted as the contour point pixel set (x k , y k ), k = 1, 2,..n, and calculate the gray value of the pixel points in the single-channel image; Use an image erosion convolution algorithm to perform an erosion operation on the binarized molten pool image, and retain the white pixel points of the last erosion; Connect the pixel point boundaries, and determine the molten pool width based on the largest inscribed circle of the convex hull within the boundary range and the minimum distance from the largest inscribed circle to the edge of the binarized molten pool image; Among them, using an image erosion convolution algorithm to perform an erosion operation on the binarized molten pool image and retaining the white pixel points of the last erosion includes the following process: Use a preset convolution kernel to convolve the binarized molten pool image. During the process of using the convolution kernel to convolve the binarized molten pool image, when the convolution result y' of the target area is less than 255*8, the target value y" of the target area is set to 0: Repeat the convolution operation until the target values y” of all pixel points all become 0, record the target area of the last convolution, and denote the set of edge points of the target area as (m k , n k ); Determining the molten pool width based on the largest inscribed circle of the convex hull within the boundary range and the minimum distance from the largest inscribed circle to the edge of the binarized molten pool image includes the following process: First, calculate the radius of the largest inscribed circle within the set of edge points (m k , n k ) of the target area. The radius of the largest inscribed circle is denoted as R1; Then, calculate the minimum distance from the points on the largest inscribed circle to the set of contour point pixels (x k , y k ), denoted as R2; Finally, calculate the molten pool width D: D = 2*(R1 + R2); Among them, calculating the radius of the largest inscribed circle within the set of edge points of the target area (m k , n k ) includes the following process: In the target area after the last convolution, find a point that satisfies: the sum of the distances from this point to all edge points of the target area after the last convolution operation is the smallest, then use this point as the center of the circle, and determine the largest inscribed circle and its radius R1 by calculating the largest inscribed circle of the triangle.

2. The method for extracting the width of a laser metal directed energy deposition molten pool based on edge corrosion according to claim 1, wherein The grayscaling of the original molten pool image to generate a grayscale image of the molten pool includes: Perform grayscaling based on the following grayscale processing function: y gray = (R + G + B) / 3 Wherein, R, G, and B are the gray values of the corresponding pixels in the red, green, and blue channel images of the original molten pool image, and y gray represents the gray value of each pixel after gray processing.

3. The method for extracting the width of a laser metal directed energy deposition molten pool based on edge corrosion according to claim 1, characterized in that, The binarization of the grayscale image of the molten pool to obtain a binarized molten pool image includes: Perform image binarization based on a preset binarization threshold and according to the following binarization function: Among them, when the grayscale value y of a certain pixel point gray is greater than the binarization threshold a, the pixel point value y is set to 255, otherwise the pixel point value y is set to 0, and thus a binarized molten pool image is obtained.

4. The method for extracting the width of the laser metal directed energy deposition molten pool based on edge corrosion according to claim 1, characterized in that The convolution kernel is defined as follows:

5. A system for extracting the width of a laser metal direct energy deposition molten pool based on edge erosion, characterized in that, It includes: One or more processors; A memory that stores operable instructions, and the instructions, when executed by the one or more processors, cause the one or more processors to perform operations, and the operations include the processes of the method described in any one of claims 1-4.

6. A computer-readable medium storing software, characterized in that, The software includes instructions that can be executed by one or more computers, and the instructions, when executed, cause the one or more computers to perform operations, and the operations include the processes of the method described in any one of claims 1-4.

Images