Welding Quality Detection Method and System Suitable for Overlapping Welded Parts

Abstract

The present invention relates to a welding quality detection method and system suitable for overlapping weldments. It includes: providing a target overlapping weldment, and respectively performing ultrasonic scanning on two opposite surfaces of the target overlapping weldment to respectively obtain a weld ultrasonic reflection signal group related to the weld seam and a weld surface ultrasonic reflection signal group related to the weld surface; generating a weld seam state image based on the weld ultrasonic reflection signal group, and generating a weld surface state image based on the weld surface ultrasonic reflection signal group, fusing the weld seam state image and the weld surface state image, and generating a welding fusion image, calculating the fusion ratio of the welding connection area in the welding quality recognition area based on the weld fusion image, and comparing the calculated fusion ratio with a quality detection ratio threshold to determine the welding quality state of the target overlapping weldment after the comparison. The present invention can perform ultrasonic detection on the welding quality of overlapping weldments and improve the accuracy of ultrasonic detection.

Description

Technical Field

[0001] The present invention relates to a detection method and system, and in particular to a welding quality detection method and system suitable for overlapping weldments. Background Art

[0002] Laser welding is widely used in the field of precision manufacturing due to its characteristics of high energy density and small heat affected zone. In order to understand the quality of laser welding, it is necessary to detect the welding quality of the weldments using laser welding. When detecting the welding quality of the weldments using laser welding, traditional detection methods can be used, such as the tensile shear method. When using the tensile shear method to detect the welding quality of the weldments, it will cause damage to the weldments. Therefore, the traditional detection method cannot perform real-time detection on the weldments on the production line.

[0003] In order to detect the welding quality of the weldments, ultrasonic detection technology can be used to detect the welding quality of the weldments. Specifically, although ultrasonic detection technology is a non-destructive detection, it has limitations in detecting laser welded parts. For example, when the weld seam is located at the contact surface between overlapping metal sheets, the existing ultrasonic detection technology is difficult to directly obtain the three-dimensional data of the weld seam area; the resolution of the transition area (such as the fusion zone and the non-fusion zone) is insufficient, that is, the existing ultrasonic detection technology is difficult to meet the welding quality detection requirements of overlapping weldments. Summary of the Invention

[0004] The purpose of the present invention is to overcome the deficiencies in the prior art and provide a welding quality detection method and system suitable for overlapping weldments, which can perform ultrasonic detection on the welding quality of overlapping weldments and improve the accuracy of ultrasonic detection.

[0005] According to the technical solution provided by the present invention, a welding quality detection method suitable for overlapping weldments, the welding quality detection method includes:

[0006] Providing a target overlapping weldment, and respectively performing ultrasonic scanning processing on two opposite surfaces of the target overlapping weldment to respectively obtain a weld ultrasonic reflection signal group related to the weld seam and a welding surface ultrasonic reflection signal group related to the welding surface;

[0007] Generating a weld state image based on the weld ultrasonic reflection signal group, and generating a welding surface state image based on the welding surface ultrasonic reflection signal group, wherein the weld state image includes a weld reference area representing the welding state, and the welding surface state image includes a welding surface reference area representing the range of the welding surface;

[0008] Fusing the weld state image and the welding surface state image, and generating a welding fusion image, wherein,

[0009] During image fusion, at least align the weld reference area with the weld surface reference area, and fuse the weld reference area within the weld surface reference area to form a welding quality identification area in the welded fusion image. Among them, the welding quality identification area at least includes a welding connection area and a welding transition area;

[0010] Based on the weld fusion image, calculate the fusion ratio of the welding connection area within the welding quality identification area, and compare the calculated fusion ratio with the quality inspection ratio threshold to determine the welding quality status of the target overlapping weldment after comparison.

[0011] Among the two opposite surfaces of the target overlapping weldment subjected to ultrasonic scanning processing, at least the first surface and the second surface corresponding to the weld. Among the first surface and the second surface, at least one surface is a welding bearing surface;

[0012] When performing ultrasonic scanning processing on each surface, it at least includes a surface scanning operation and a signal interception operation performed in sequence. Among them,

[0013] After performing the surface scanning operation on the current surface, generate a surface ultrasonic reflection source signal group;

[0014] When performing the signal interception operation, intercept the surface ultrasonic reflection source signal group to generate a surface ultrasonic reflection intercepted signal group;

[0015] When the current surface is the welding bearing surface, configure the surface ultrasonic reflection intercepted signal group as the weld surface ultrasonic reflection signal group, otherwise, configure the surface ultrasonic reflection intercepted signal group as the weld ultrasonic reflection signal group.

[0016] When performing ultrasonic scanning processing, it also includes a tilt correction operation. Among them,

[0017] Before performing the signal interception operation, first perform the tilt correction operation;

[0018] When performing the tilt correction operation, generate a reflection scan TOF array based on the surface ultrasonic reflection source signal group, and calculate and generate the flight time correction amount corresponding to each surface ultrasonic reflection source signal based on the reflection scan TOF array at the smooth position;

[0019] Use the calculated flight time correction amount to perform tilt correction on the corresponding surface ultrasonic reflection source signal to generate a surface ultrasonic reflection tilt correction signal after tilt correction;

[0020] Generate a surface ultrasonic reflection tilt correction signal group based on all the surface ultrasonic reflection tilt correction signals;

[0021] When performing signal interception operation, signal interception is performed on the surface ultrasonic reflection tilt correction signal group to generate a surface ultrasonic reflection intercepted signal group after signal interception.

[0022] When calculating the flight time correction amount corresponding to each weldment ultrasonic reflection scan signal, it includes:

[0023] Based on the reflection scan TOF values corresponding to the smoothing positions in the reflection scan TOF array, plane fitting is performed to generate a TOF fitting plane;

[0024] For each surface ultrasonic scan point, the TOF fitting value of the current surface ultrasonic scan point is determined using the TOF fitting plane, and the flight time correction amount is calculated based on the determined TOF fitting value and the corresponding reflection scan TOF value in the reflection scan TOF array;

[0025] During tilt correction, when the flight time correction amount is a non-integer, the surface ultrasonic reflection source signal is sequentially displacement-corrected based on the integer correction value and the decimal correction value, and a surface ultrasonic reflection tilt correction signal is generated after displacement correction, where

[0026] The integer correction value is the integer part value of the flight time correction amount;

[0027] The decimal correction value is the decimal part value of the flight time correction amount, and when performing displacement correction based on the decimal correction value, the displacement correction method includes linear interpolation.

[0028] When generating a weld seam state image, it includes:

[0029] Based on the weld seam ultrasonic reflection signal group, a weld seam area signal intensity image representing the ultrasonic reflection signal intensity is generated;

[0030] The weld seam area signal intensity image is binarized to generate a weld seam state image after binarization, where

[0031] During binarization, two signal feature intensity peaks are searched in the weld seam area signal intensity image, and the signal feature intensity valley value between the two searched signal feature intensity peaks is configured as the binarization segmentation threshold;

[0032] The weld seam area signal intensity image is binarized using the configured binarization segmentation threshold.

[0033] When performing image fusion, it includes:

[0034] Select the weld surface state image or the weld seam state image as the fusion reference image, and then configure the weld seam state image or the weld surface state image as the fusion paired image;

[0035] Mirror flip the fusion reference image to generate a fusion reference mirror-flipped image after mirror flipping;

[0036] Logically invert the fusion reference mirror-flipped image and perform rotation correction after logical inversion to generate a fusion reference inverted and corrected image, where the fusion reference inverted and corrected image includes a superimposed reference area, and the binary value state of the superimposed reference area is different from the binary value state of the weld reference area;

[0037] Extract the paired feature positions in the fusion paired image and superimpose the fusion paired image on the fusion reference inverted and corrected image based on the extracted feature positions to form a welded fusion image, and generate a welding quality identification area in the welded fusion image, where,

[0038] Form a weld connection area based on the superimposed state of the weld reference area and the superimposed reference area, and the part of the superimposed reference area other than forming the weld connection area forms a welding transition area.

[0039] When calculating the fusion ratio, it includes:

[0040] Statistically calculate the pixel area of the weld connection area and the pixel area of the welding quality identification area, and use the ratio of the pixel area of the weld connection area to the pixel area of the welding quality identification area as the fusion ratio;

[0041] The quality detection ratio threshold includes a better quality ratio threshold and a non-conforming quality ratio threshold,

[0042] When the fusion ratio is not less than the better quality ratio threshold, the welding quality state of the target overlapping weldment is good welding quality;

[0043] When the fusion ratio is less than the non-conforming quality ratio threshold, the welding quality state of the target overlapping weldment is judged as non-conforming welding quality.

[0044] When the fusion ratio is not less than the non-conforming quality ratio threshold and less than the better quality ratio threshold, the welding quality state of the target overlapping weldment is judged as qualified welding quality, and the quality grade of the welding quality state is given, where,

[0045] When determining the quality grade of the welding quality, it includes:

[0046] Extract the ultrasonic reflection signals corresponding to the welded fusion images, calculate the corresponding reflection signal feature intensity values based on the extracted ultrasonic reflection signals, and normalize the calculated reflection signal feature intensity values. Thereafter, generate the normalized reflection signal feature intensity histograms of the weld connection area, the normalized reflection signal feature intensity histograms of the welding transition area, and the normalized reflection signal feature intensity histograms of the unfused area based on the normalized reflection signal feature intensity values, where the unfused area is the area in the welded fusion image except for the welding quality identification area;

[0047] For all the above-mentioned normalized reflection signal feature intensity histograms, calculate the corresponding histogram statistical information for each normalized histogram, where the histogram statistical information includes the mean, standard deviation, and kurtosis;

[0048] Based on the corresponding histogram statistical information of the weld connection area, the welding transition area, and the unfused area, determine the quality grade of the welding quality status.

[0049] When the fusion ratio is not less than the quality unqualified ratio threshold and less than the quality better ratio threshold, then judge the welding quality status of the target overlapping weldment as qualified welding quality, and give the quality grade of the welding quality status, where,

[0050] When the number of welding quality detections of the target overlapping weldment exceeds the detection number threshold, then use the comprehensive quality coefficient to determine the quality grade of the determined welding quality status, where,

[0051] For the comprehensive quality coefficient, there is:

[0052]

[0053] In the formula, is the comprehensive quality coefficient, is the mean of the welding transition area, is the standard deviation of the welding transition area, is the kurtosis of the weld connection area, 、 、 are the weight coefficients;

[0054] Calculate the comprehensive quality coefficients corresponding to all historical target overlapping weldments, and determine the quality grade threshold based on all the comprehensive quality coefficients,

[0055] Compare the comprehensive quality coefficient of the current target overlapping weldment with the determined quality grade threshold above to determine the quality grade of the welding quality status corresponding to the current target overlapping weldment.

[0056] A welding quality detection system suitable for overlapping weldments, including an ultrasonic scanning probe and a welding quality detection terminal, where the ultrasonic scanning probe is adaptively connected to the welding quality detection terminal. Among them,

[0057] For any target overlapping weldment, the ultrasonic scanning probe is used to perform ultrasonic scanning on the target overlapping weldment. After that, the welding quality detection terminal uses the above welding quality detection method to perform quality detection and judgment to determine the welding quality status of the current target overlapping weldment.

[0058] Advantages of the present invention: Ultrasonic scanning is performed on two opposite surfaces of the target overlapping weldment to respectively obtain a weld ultrasonic reflection signal group related to the weld seam and a weld surface ultrasonic reflection signal group related to the weld surface. After that, a weld seam state image is generated based on the weld ultrasonic reflection signal group, a weld surface state image is generated based on the weld surface ultrasonic reflection signal group, the weld seam state image and the weld surface state image are subjected to image fusion, and a welding fusion image is generated;

[0059] Calculate the fusion ratio of the welding connection area within the welding quality identification area, and compare the calculated fusion ratio with the quality detection ratio threshold to determine the welding quality status of the target overlapping weldment after comparison, so as to realize ultrasonic detection of the welding quality of the overlapping weldment and improve the accuracy of ultrasonic detection. Description of the Drawings

[0060] Figure 1 It is a flowchart of an embodiment of the welding quality detection of the present invention.

[0061] Figure 2 It is a perspective view of an embodiment of the target overlapping weldment of the present invention.

[0062] Figure 3 It is a schematic structural diagram of an embodiment of the target overlapping weldment of the present invention.

[0063] Figure 4 It is a schematic diagram of an embodiment of the signal intensity image of the weld seam area of the present invention.

[0064] Figure 5 It is a schematic diagram of an embodiment of the weld seam state image of the present invention.

[0065] Figure 6 It is a schematic diagram of an embodiment of the signal intensity image of the weld surface area of the present invention.

[0066] Figure 7 It is a schematic diagram of an embodiment of the weld surface state image of the present invention.

[0067] Figure 8 It is a schematic diagram of an embodiment of the present invention in which the weld surface state image is used as a fusion reference graph and a fusion reference mirror-flipped image is formed after mirror flipping.

[0068] Figure 9 For Figure 8 An exemplary diagram showing an image obtained by flipping a fusion reference mirror image in

[0069] Figure 10 For Figure 9 An exemplary diagram showing a fusion reference inverted and corrected image generated after rotation correction of the fusion reference inverted image in

[0070] Figure 11 For Figure 10 An exemplary diagram showing a welding fusion image generated by using the fusion reference inverted and corrected image in

[0071] Figure 12 For Figure 5 An exemplary diagram showing the result after edge detection of the weld reference area in

[0072] Figure 13 An exemplary diagram showing an image of the signal intensity in the weld area of the present invention.

[0073] Explanation of reference numerals: 1 - First welding substrate, 2 - Second welding substrate, and 3 - Substrate welding stripe. Detailed implementation manner

[0074] The present invention will be further described below with reference to specific drawings and embodiments.

[0075] In order to perform ultrasonic detection on the welding quality of overlapping weldments and improve the accuracy of ultrasonic detection, the present invention provides a welding quality detection method suitable for overlapping weldments. Specifically, the welding quality detection method includes:

[0076] Providing a target overlapping weldment, and respectively performing ultrasonic scanning processing on two opposite surfaces of the target overlapping weldment to respectively obtain a weld ultrasonic reflection signal group related to the weld and a weld surface ultrasonic reflection signal group related to the weld surface;

[0077] Generating a weld state image based on the weld ultrasonic reflection signal group, and generating a weld surface state image based on the weld surface ultrasonic reflection signal group. Among them, the weld state image includes a weld reference area representing the welding state, and the weld surface state image includes a weld surface reference area representing the weld surface range;

[0078] Fusing the weld state image and the weld surface state image, and generating a welding fusion image, where

[0079] When performing image fusion, at least align the weld reference area with the weld surface reference area, and fuse the weld reference area within the weld surface reference area to form a welding quality identification area in the welded fusion image. Among them, the welding quality identification area at least includes a welding connection area and a welding transition area;

[0080] Based on the weld fusion image, calculate the fusion ratio of the welding connection area within the welding quality identification area, and compare the calculated fusion ratio with the quality detection ratio threshold to determine the welding quality status of the target overlapping weldment after comparison.

[0081] Figure 1 FIG. shows a flowchart of an embodiment for detecting the welding quality of overlapping weldments according to the present invention. It can be seen from the figure that when performing welding quality detection, a target overlapping weldment should be provided, that is, when performing welding quality detection, mainly determine the welding quality status of the target overlapping weldment; specifically, for the target overlapping weldment, the weld is located in the non-external surface area of the target overlapping weldment. At this time, the welding quality of the weld area cannot be directly observed by the naked eye. The following combines Figure 2 and Figure 3 to give an example of the situation of the target overlapping weldment.

[0082] Figure 2 and Figure 3 FIG. shows an embodiment of the target overlapping weldment. In the figure, the target overlapping weldment includes a first welding substrate 1 and a second welding substrate 2. Among them, there is an overlapping area between the first welding substrate 1 and the second welding substrate 2, and the weld is located in the overlapping area between the first welding substrate 1 and the second welding substrate 2. The first welding substrate 1 and the second welding substrate 2 can be welded and connected by existing common welding means, and the welding method used can be selected according to needs. For example, laser welding can be used. Therefore, the first welding substrate 1 and the second welding substrate 2 should be weldable connection materials. The specific welding connection method and process can be consistent with the prior art and will not be elaborated here.

[0083] From Figure 2 and Figure 3 the target overlapping weldment shown, it can be seen that the target overlapping weldment at least includes two opposite surfaces. For example, the outer surface of the first welding substrate 1 and the corresponding outer surface of the second welding substrate 2 form two corresponding surfaces of the target overlapping weldment, and the two corresponding surfaces are both corresponding to the weld. For example, the two surfaces are respectively located on both sides of the weld. The weld can be consistent with the prior art, that is, the first welding substrate 1 and the second welding substrate 2 can be connected into one body through the weld. When the target overlapping weldment is of other types, the two corresponding surfaces of the target overlapping weldment can be determined according to the position of the weld, which will not be listed one by one here.

[0084] In specific implementation, for each target overlapping weldment, after determining the two surfaces corresponding to the weld of the target overlapping weldment, ultrasonic scanning treatment should be performed on at least the two determined surfaces respectively, so that after the ultrasonic scanning treatment, a weld ultrasonic reflection signal group related to the weld and a weld surface ultrasonic reflection signal group related to the weld surface can be obtained. Among them, the weld ultrasonic reflection signal group can characterize the reflection state of the weld on the ultrasonic signal, and the weld surface ultrasonic reflection signal group can characterize the reflection state of the weld surface on the ultrasonic signal. The method of ultrasonic scanning treatment and the situation of generating the weld ultrasonic reflection signal group and the weld surface ultrasonic reflection signal group through ultrasonic scanning treatment will be specifically described below.

[0085] It can be seen from Figure 1 that after obtaining the weld ultrasonic reflection signal group, a weld state image can be generated based on the weld ultrasonic reflection signal group. Among them, the weld state image should at least include a weld reference area characterizing the welding state. The method and process of generating the weld state image will be specifically described below. Similarly, a weld surface state image can be generated based on the weld surface ultrasonic reflection signal group. The weld surface state image can include a weld surface reference area characterizing the weld surface range. Among them, the weld surface range can be the expected welding area. For example, when the welding method mentioned above is laser welding, it is the area range for carrying out laser welding. When it is other welding methods, the corresponding weld surface range can be determined according to the specific welding method. The specific situation of determining the weld surface range will not be exemplified one by one here.

[0086] In order to perform welding quality detection, image fusion should be carried out on the weld state image and the weld surface state image. Among them, during image fusion, at least the weld reference area and the weld surface reference area should be aligned. After alignment, the weld reference area should be fused into the weld surface reference area, that is, the weld reference area should be placed within the weld surface reference area. And after placing the weld reference area within the weld surface reference area, a welding fusion image can be formed, and a welding quality identification area can be generated in the welding fusion image. The method and process of image fusion will be exemplified below, and specific reference can be made to the corresponding descriptions below.

[0087] After obtaining the weld fusion image, the fusion ratio of the welded connection area within the welding quality identification area can be calculated. Thereafter, the calculated fusion ratio can be compared with the quality detection ratio threshold to determine the welding quality state of the target overlapping weldment after comparison. Among them, the determined welding quality state of the target overlapping weldment generally can include good welding quality or unqualified welding quality. It can be understood that after determining the welding quality state of the target overlapping weldment, the welding quality detection of the target overlapping weldment is completed.

[0088] In one embodiment of the present invention, among the two opposite surfaces of the target overlapping weldment where ultrasonic scanning is performed, at least the first surface and the second surface corresponding to the weld seam are involved. Among the first surface and the second surface, at least one surface is a welding bearing surface;

[0089] When performing ultrasonic scanning on each surface, it at least includes a surface scanning operation and a signal intercepting operation performed in sequence. Among them,

[0090] After performing the surface scanning operation on the current surface, a surface ultrasonic reflection source signal group is generated;

[0091] When performing the signal intercepting operation, signal interception is performed on the surface ultrasonic reflection source signal group to generate a surface ultrasonic reflection intercepted signal group;

[0092] When the current surface is a welding bearing surface, the surface ultrasonic reflection intercepted signal group is configured as a welding surface ultrasonic reflection signal group. Otherwise, the surface ultrasonic reflection intercepted signal group is configured as a weld seam ultrasonic reflection signal group.

[0093] It can be seen from the above description that the two surfaces of the target overlapping weldment where ultrasonic scanning is performed should be related to the weld seam. When the two surfaces corresponding to the weld seam are respectively called the first surface and the second surface, at least one surface of the first surface and the second surface is a welding bearing surface. For example, when the welding method of the above-mentioned target overlapping weldment is laser welding, the surface bearing the laser welding is the welding bearing surface. Figure 2 and Figure 3 In the embodiment shown in, the corresponding outer surface of the second welding substrate 2 can be a welding bearing surface. When laser welding is used, the laser is incident on the corresponding surface of the second welding substrate 2 to act on the solder between the first reference welding body 1 and the second welding substrate 2 by using the laser and achieve the purpose of welding. Specifically, the area where the laser is incident on the surface of the second welding substrate 2 should overlap with the area corresponding to the first reference welding body 1 to enable effective welding. Specifically, when the welding method is other welding, the corresponding welding bearing surface can be determined accordingly. When the welding bearing surface is determined, the surface corresponding to the welding bearing surface can be further determined, and then ultrasonic scanning can be performed on each surface.

[0094] In specific implementation, the method of performing ultrasonic scanning on each surface is the same. Generally, it can at least include a surface scanning operation and a signal intercepting operation. Among them, the surface scanning operation is to perform ultrasonic scanning on the current surface and generate a surface ultrasonic reflection source signal group after ultrasonic scanning. The method and process of performing ultrasonic scanning will be illustrated by examples below.

[0095] When performing ultrasonic scanning, a high-frequency probe with a transmitting frequency of 20 MHz can be selected. For the reflected ultrasonic signals, the sampling frequency can be set to 50 MHz. The ultrasonic coupling method for the target overlapping weldment is the immersion type, that is, the target overlapping weldment is placed in water. After that, the position of the high-frequency probe and the scanning parameters should be adjusted. The position of the high-frequency probe should be adjusted to obtain a strong reflected signal. The configured scanning parameters can include the scanning step accuracy and the scanning range. Among them, the step accuracy can include the rough scanning step accuracy and / or the fine scanning step accuracy. The rough scanning step accuracy can be 1 mm, and the fine scanning step accuracy can be 0.1 mm. For the scanning range, generally, it should effectively cover the weld and the weld surface range. For example, in the above embodiment, when the outer surface of the second welding base 2 is the welding bearing surface, Figure 2 the area between the left end of the second welding base 2 and the right end of the first reference welding body 1 can be set as the scanning range. Similarly, the area corresponding to the first reference welding body 1 can be determined as the scanning range.

[0096] During specific implementation, Figure 2 one end foot at the left end of the second welding base 2 in can be used as the scanning starting point. The width direction of the second welding base 2 is the scanning X direction, and the length direction of the second welding base 2 is the scanning Y direction. After that, the high-frequency probe is used to perform ultrasonic scanning under the above configuration to generate a surface ultrasonic reflection source signal group after scanning. It can be understood that the surface ultrasonic reflection source signal group includes several surface ultrasonic reflection source signals, and the surface ultrasonic reflection source signals in the surface ultrasonic reflection source signal group are distributed in an array.

[0097] It should be noted that surface scanning operations should be performed on the first surface and the second surface of the target overlapping weldment respectively. That is, after two surface scanning operations, two corresponding surface ultrasonic reflection source signal groups can be generated. From this, it can be seen that the target overlapping weldment of the present invention should be of a type suitable for double-sided ultrasonic scanning operation, that is, the target overlapping weldment should not be limited to Figure 2 the embodiment shown in . Specifically, it should be based on being able to meet the double-sided ultrasonic scanning operation and respectively obtain the corresponding surface ultrasonic reflection source signal groups. Those skilled in the art know that the surface ultrasonic reflection source signal is a signal similar to a sine wave, and the surface ultrasonic reflection source signal includes the sampling values of several sampling points. Different sampling points can represent the reflection of ultrasonic signals at different positions. Therefore, in order to obtain a weld ultrasonic reflection signal group and a weld surface ultrasonic reflection signal group related to the weld, a signal truncation operation should be performed on the surface ultrasonic reflection source signal group.

[0098] It can be understood that when performing signal interception operations, it should be related to the type of the surface corresponding to the generated surface ultrasonic reflection source signal group. For example, when the surface ultrasonic reflection source signal group is generated by performing a surface scanning operation on the welding bearing surface, during the signal interception operation, the sampling points corresponding to the ultrasonic signal reflection of the welding bearing surface should be mainly retained, so that a weld surface ultrasonic reflection signal group can be obtained after signal interception; while when the surface ultrasonic reflection source signal group is generated by performing a surface scanning operation on the surface corresponding to the welding bearing surface, during the signal interception operation, the sampling points corresponding to the ultrasonic signal reflection of the weld seam should be mainly retained, so that a weld seam ultrasonic reflection signal group can be obtained after signal interception.

[0099] As can be seen from the above description, the signal interception operation should be related to the type of the current surface. After determining the type of the current surface, a signal interception operation can be performed on the surface ultrasonic reflection source signal group. The way of performing the signal interception operation can be consistent with the prior art. For example, reference can be made to the description of the signal interception processing adopted in the publication number CN118362642A, which will not be elaborated here. As can be seen from the above description, after the signal interception operation, a weld seam ultrasonic reflection signal group and a weld surface ultrasonic reflection signal group can be obtained respectively. Since the depths corresponding to the weld surface and the weld seam are different, therefore, during the signal interception operation, in order to form a weld surface ultrasonic reflection signal group, the relatively earlier sampling points in the surface ultrasonic reflection source signal should be intercepted and retained to accurately represent the reflection of the weld surface to the ultrasonic signal; while in order to form a weld seam ultrasonic reflection signal group, the sampling points in the middle part of the surface ultrasonic reflection source signal need to be intercepted and retained to accurately represent the reflection of the weld seam to the ultrasonic signal.

[0100] In an embodiment of the present invention, during ultrasonic scanning processing, an inclination correction operation is further included, wherein,

[0101] Before performing the signal interception operation, the inclination correction operation is first performed;

[0102] When performing the inclination correction operation, a reflection scanning TOF array is generated based on the surface ultrasonic reflection source signal group, and a flight time correction amount corresponding to each surface ultrasonic reflection source signal is calculated and generated based on the reflection scanning TOF array at the smooth position;

[0103] The corresponding surface ultrasonic reflection source signal is inclination-corrected by using the calculated flight time correction amount, so as to generate a surface ultrasonic reflection inclination-corrected signal after inclination correction;

[0104] A surface ultrasonic reflection inclination-corrected signal group is generated based on all the surface ultrasonic reflection inclination-corrected signals;

[0105] When performing the signal truncation operation, signal truncation is performed on the surface ultrasonic reflection tilt correction signal group to generate a surface ultrasonic reflection post-truncation signal group after signal truncation.

[0106] Specifically, when performing the surface scanning operation, coupling methods such as water immersion need to be adopted. And when placing the target overlapping weldment, an inevitable tilt phenomenon will occur, such as a global tilt of the target overlapping weldment caused by the placement. When there is a global tilt, if the signal truncation operation is directly performed, it will cause interference in the weld surface ultrasonic reflection signal group and the weld seam ultrasonic reflection signal group generated after signal truncation, thereby affecting the accuracy and reliability of subsequent welding quality detection.

[0107] In order to improve the accuracy of welding quality detection, in the present invention, the tilt correction operation can be performed before the signal truncation operation. Therefore, when performing the tilt correction operation, the objects targeted are the above two surface ultrasonic reflection source signal groups. As can be seen from the above description, the surface ultrasonic reflection source signals in each surface ultrasonic reflection source signal group are distributed in an array. Therefore, the reflection scanning TOF (time of flight, TOF) value corresponding to each surface ultrasonic reflection source signal is calculated, and a reflection scanning TOF array can be generated based on the calculated reflection scanning TOF values. Specifically, when calculating the reflection scanning TOF value corresponding to each surface ultrasonic reflection source signal, the commonly used technical means in the prior art can be adopted, specifically based on being able to calculate the corresponding reflection scanning TOF value.

[0108] In order to achieve tilt correction, in an embodiment of the present invention, the reflection scanning TOF array corresponding to the smoothing position can be used as the tilt correction reference, where the smoothing position can be the middle position within the above-mentioned configured scanning range. Therefore, after selecting the smoothing position, the reflection scanning TOF array corresponding to the smoothing position can be determined. Thereafter, the flight time correction amount corresponding to each surface ultrasonic reflection source signal can be calculated and generated.

[0109] In an embodiment of the present invention, when calculating and generating the flight time correction amount corresponding to each weldment ultrasonic reflection scanning signal, it includes:

[0110] Performing plane fitting based on the reflection scanning TOF value corresponding to the smoothing position in the reflection scanning TOF array to generate a TOF fitting plane;

[0111] For each surface ultrasonic scanning point, using the TOF fitting plane to determine the TOF fitting value of the current surface ultrasonic scanning point, and calculating and generating the flight time correction amount based on the determined TOF fitting value and the corresponding reflection scanning TOF value in the reflection scanning TOF array;

[0112] During tilt correction, when the flight time correction amount is a non-integer, the surface ultrasonic reflection source signal is sequentially displacement-corrected based on the integer correction value and the decimal correction value, and a surface ultrasonic reflection tilt correction signal is generated after the displacement correction, where,

[0113] the integer correction value is the integer part value of the flight time correction amount;

[0114] the decimal correction value is the decimal part value of the flight time correction amount, and when displacement correction is performed based on the decimal correction value, the displacement correction method includes linear interpolation.

[0115] As can be seen from the above description, when selecting the smoothing position, all reflection scan TOF values corresponding to the smoothing position can be obtained. Thereafter, based on all the obtained reflection scan TOF values, plane fitting can be performed to generate a TOF fitting plane after plane fitting. The method and process of generating the TOF fitting plane will be illustrated by examples below.

[0116] Specifically, since the tilt of the surface during surface scanning operations is global, each reflection scan TOF value can be expressed as:

[0117]

[0118] where, is the position coordinate corresponding to a reflection scan TOF value, represents the change slope of the current reflection scan TOF value in the x direction, represents the change slope of the current reflection scan TOF value in the y direction, represents the intercept.

[0119] In specific implementation, all reflection scan TOF values corresponding to the smoothing position are plane-fitted using the least squares method to generate a TOF fitting plane after plane fitting. The method and process of plane fitting using the least squares method can be consistent with the prior art and will not be elaborated here. Of course, other methods can also be used for plane fitting, and the plane fitting method can be selected according to needs to meet the plane fitting requirements here.

[0120] As can be seen from the above description, each reflection scan TOF value corresponds to a surface ultrasonic scan point. Therefore, after obtaining the TOF fitting plane, for each surface ultrasonic scan point, the TOF fitting value corresponding to the current surface ultrasonic scan point can be obtained. Specifically, there is:

[0121]

[0122] where, is the position coordinate of the current surface ultrasonic scan point, is the slope of the TOF fitting plane in the x direction, represents the slope of the TOF fitting plane in the y direction, represents the intercept of the TOF fitting plane.

[0123] In specific implementation, when calculating the corresponding time-of-flight correction amount for each surface ultrasonic scanning point, there is:

[0124]

[0125] where, is the position coordinate of the time-of-flight correction amount of the surface ultrasonic scanning point, is the reflected scanning TOF value of the surface ultrasonic scanning point with the position coordinate and is the TOF fitting value of the surface ultrasonic scanning point with the position coordinate .

[0126] It should be understood that after obtaining the time-of-flight correction amount of each surface ultrasonic scanning point, the surface ultrasonic wave reflection source signal of the current surface ultrasonic scanning point can be subjected to tilt correction so that after the tilt correction, all the surface ultrasonic wave reflection tilt correction signals have no tilt in the depth direction. Specifically, during the tilt correction, the sampling points of each surface ultrasonic wave reflection source signal are corrected using the time-of-flight correction amount, such as displacing the sampling points corresponding to the surface ultrasonic wave reflection source signal.

[0127] Specifically, for each surface ultrasonic wave reflection source signal, the surface ultrasonic wave reflection source signal can be represented by the sequence where t is the sampling time corresponding to the surface ultrasonic wave reflection source signal;

[0128] After determining the time-of-flight correction amount, a time axis of the current surface ultrasonic wave reflection source signal is established according to the time-of-flight correction amount, and there is: , where when the time-of-flight correction amount is an integer value, then is also an integer value. Therefore, the corresponding sampling values in the surface ultrasonic wave reflection source signal can be subjected to corresponding time shifts to complete the displacement correction after the time shift. When the time-of-flight correction amount is a non-integer value, then is also a non-integer value. At this time, the integer part value and the decimal part value of the time-of-flight correction amount should be extracted first. Among them, the extracted integer part value can form an integer correction value, and the extracted decimal part can form a decimal correction value. For example, when the time-of-flight correction amount is 3.7, the formed integer correction value is 3, and the decimal correction value is 0.7. ​

[0129] During specific implementation, when performing displacement correction based on an integer correction value, the above-mentioned time-of-flight correction amount can be referred to. In the case of an integer value. When performing displacement correction based on a decimal correction value, a feasible displacement correction method can be:

[0130]

[0131] Wherein, is the surface ultrasonic reflection tilt correction signal generated after displacement correction based on the decimal correction value, is the integer part value of the time-of-flight correction amount, i.e., extracting the time-of-flight correction amount of the integer part value, is the decimal correction value, is the time axis established based on the integer correction value.

[0132] Specifically, after each surface ultrasonic reflection source signal is subjected to the above-mentioned tilt correction, the corresponding surface ultrasonic reflection tilt correction signal can be obtained. Thereafter, based on all the surface ultrasonic reflection tilt correction signals corresponding to the same surface ultrasonic reflection source signal group, a surface ultrasonic reflection tilt correction signal group can be generated. When performing a signal truncation operation, the surface ultrasonic reflection tilt correction signal group is subjected to signal truncation to generate a surface ultrasonic reflection truncated signal group after signal truncation.

[0133] In an embodiment of the present invention, when generating a weld state image, it includes:

[0134] Generating a weld zone signal intensity image representing the intensity of the ultrasonic reflection signal based on the weld ultrasonic reflection signal group;

[0135] Performing binarization processing on the weld zone signal intensity image to generate a weld state image after binarization processing, wherein,

[0136] During binarization processing, search for two signal feature intensity peaks in the weld zone signal intensity image, and configure the signal feature intensity valley value between the two searched signal feature intensity peaks as the binarization segmentation threshold;

[0137] Perform binarization processing on the weld zone signal intensity image using the configured binarization segmentation threshold.

[0138] In specific implementation, after obtaining the weld ultrasonic reflection signal group by the above method, a signal intensity image of the weld area can be generated based on the weld ultrasonic reflection signal group. When generating the signal intensity image of the weld area, the signal feature intensity value of each weld ultrasonic reflection signal in the weld ultrasonic reflection signal group should be calculated. The signal feature intensity value can be the PPV (Peak to Peak Voltage) value, amplitude, envelope area of the weld ultrasonic reflection signal, or other characteristic values that can reflect the energy of the weld ultrasonic reflection signal. When the signal feature intensity value is PPV, amplitude or envelope area, the corresponding signal feature intensity value can be calculated by using the commonly used technical means in the technical field. The specific calculation method and process are not described here again.

[0139] After calculating the signal feature intensity value of each weld ultrasonic reflection signal, map each signal feature intensity value onto a grayscale image, and then the signal intensity image of the weld area can be obtained. That is, the obtained signal intensity image of the weld area is a grayscale image. Figure 4 An embodiment of the signal intensity image of the weld area is shown.

[0140] After obtaining the signal intensity image of the weld area, binary processing can be performed on the signal intensity image of the weld area. In order to improve the accuracy of binary processing, two signal feature intensity peaks can be searched in the signal intensity image of the weld area. For example, when the signal feature intensity value is the PPV value, two PPV peaks can be searched on the signal intensity image of the weld area. That is, the two searched PPV peaks should be greater than other PPV values on the signal intensity image of the weld area. Therefore, based on the generation method of the signal intensity image of the weld area, the corresponding two signal feature intensity peaks can be searched by using the commonly used method in the technical field.

[0141] After searching for the two signal feature intensity peaks, configure the signal feature intensity valley value between the two signal feature intensity peaks as the binary segmentation threshold. Among them, the signal feature intensity valley value is the smallest signal feature intensity value between the two signal feature intensity peaks. As can be seen from the above description, each signal feature intensity value will be mapped to the signal intensity image of the weld area. After searching for the two signal feature intensity peaks, the corresponding signal feature intensity valley value can be searched. Thereafter, the value of the signal feature intensity corresponding to the signal feature intensity valley value can be used as the binary segmentation threshold.

[0142] Figure 13When the signal feature intensity value is the PPV value, a signal intensity image of the weld zone is generated. In the figure, the abscissa is the normalized PPV value, divided into 256 intervals, and the ordinate is the frequency, that is, the number of data points in each interval. It can be seen from the figure that there are two peak regions of the PPV value. At this time, the corresponding signal feature intensity peaks are determined within the two peak regions. Thereafter, a signal feature intensity valley is searched between the two determined signal feature intensity peaks. It should be noted that when other types of signal feature intensity values are used, the description here can be referred to in order to determine the corresponding binary segmentation threshold, and no further examples will be given here.

[0143] During binary processing, for any pixel on the signal intensity image of the weld zone, when the pixel value of the pixel is greater than the binary segmentation threshold, the binary value corresponding to the current pixel can be set to 1, otherwise, the binary value corresponding to the current pixel can be set to 0. By performing the binary processing here on all pixel values, the binary processing of the signal intensity image of the weld zone can be achieved, and a weld state image can be generated. Therefore, the weld state image is a binary image.

[0144] In specific implementation, in order to eliminate noise and isolated points, after binary processing, image morphology processing can also be adopted, and a weld state image is generated after image morphology processing. Among them, the adopted image morphology processing can include opening operation, closing operation, and / or small area removal. The specific situation of image morphology processing can be selected according to needs and will not be elaborated here. Figure 5 The weld state image generated after image morphology processing is shown. Figure 5 In it, the middle black area is the weld reference area representing the welding state.

[0145] It can be understood that when generating a weld surface state image, the above method for generating a weld state image can be adopted. Therefore, when adopting the method for generating a weld state image, the method and process for generating a weld surface state image can refer to the corresponding description above. In specific implementation, when adopting the same method as generating a weld state image, first, a signal intensity image of the weld surface area representing the ultrasonic reflection signal intensity is generated based on the weld surface ultrasonic reflection signal group. Figure 6 An embodiment of generating a signal intensity image of the weld surface area is shown. It should be noted that for Figure 6 For the signal intensity image of the weld surface area shown, the outer surface corresponding to the second welding base 2 serves as the welding bearing surface. Figure 6 The horizontal bars in it correspond to the base welding stripes 3 on the second welding base 2. Figure 7 An embodiment of generating a weld surface state image after binary processing and image morphology processing of the signal intensity image of the weld surface area is shown. Figure 7 In it, the middle black area is the weld surface reference area representing the weld surface range.

[0146] In an embodiment of the present invention, when performing image fusion, it includes:

[0147] Select the solder joint surface state image or the weld seam state image as the fusion reference image, and then configure the weld seam state image or the solder joint surface state image as the fusion paired image;

[0148] Mirror - flip the fusion reference image to generate a fusion reference mirror - flipped image after mirror - flipping;

[0149] Logically invert the fusion reference mirror - flipped image, and perform rotation correction after logical inversion to generate a fusion reference inverted and corrected image, where the fusion reference inverted and corrected image includes a superposition reference area, and the binary value state of the superposition reference area is different from the binary value state of the weld seam reference area;

[0150] Extract the paired feature positions in the fusion paired image, and superimpose the fusion paired image on the fusion reference inverted and corrected image based on the extracted feature positions to form a welding fusion image, and generate a welding quality recognition area in the welding fusion image, where,

[0151] Form a weld seam connection area based on the superposition state of the weld seam reference area and the superposition reference area, and the part of the superposition reference area except for forming the weld seam connection area forms a welding transition area.

[0152] It should be noted that when performing image fusion, one of the solder joint surface state image or the weld seam state image needs to be selected as the fusion reference image. For example, the solder joint surface state image can be selected as the fusion reference image. At this time, the weld seam state image can be used as the fusion paired image. Similarly, when the weld seam state image is used as the fusion reference image, the solder joint surface state image can be used as the fusion paired image.

[0153] It should be noted that when performing ultrasonic scanning on the target overlapping weldment in the above - mentioned manner, the obtained weld seam state image and the solder joint surface state image are in a mirror - image state in space. Therefore, after selecting the fusion reference image, in order to align the weld seam state image and the solder joint surface state image in space, the fusion reference image can be mirror - flipped to generate a fusion reference mirror - flipped image. Among them, when mirror - flipping, it can be horizontal mirror - flipping or vertical mirror - flipping. The method and process of mirror - flipping the fusion reference image can be consistent with the prior art and will not be elaborated here. Figure 8 An embodiment is shown in which the solder joint surface state image is used as the fusion reference image, and a fusion reference mirror - flipped image is generated after mirror - flipping the solder joint surface state image.

[0154] As can be seen from the above description, the fused reference mirror flipped image is a binary image. Therefore, for the convenience of subsequent fusion processing, the fused reference mirror flipped image can be logically inverted, and rotation correction can be performed after the logical inversion to generate a fused reference inverted and corrected image. Here, logical inversion means inverting the binary value of each pixel in the fused reference mirror flipped image. After inverting the binary value, the black area can be transformed into a white area, and the white area can be transformed into a black area. Figure 9 shows an embodiment after logically inverting the fused reference mirror flipped image in Figure 8 Specifically, after logical inversion, image correction can be performed through a rotation correction operation to further improve the accuracy of image alignment in image fusion, and thus the accuracy of welding quality detection can be improved.

[0155] It should be noted that during rotation correction, the corresponding upper and lower edges of the fused reference mirror flipped image and the fused paired image are mainly made to be on the same parallel line. Therefore, the commonly used rotation correction methods in this technical field can be adopted, and the rotation correction method and process can be consistent with the prior art, which will not be elaborated here. To obtain the effect of rotation correction, the required parallel line state can be determined first, and then the fused reference mirror flipped image and the fused paired image are both corrected to the corresponding parallel line state. Of course, to simplify the rotation correction, the upper and lower edges of the fused paired image can be selected as the rotation correction reference, and then the fused reference mirror flipped image is rotationally corrected. From this, it can be seen that the rotation correction method can be selected according to needs to make the fused reference mirror flipped image and the fused paired image meet the calibration alignment and not affect the subsequent superposition fusion.

[0156] Specifically, the fused reference inverted and corrected image includes a superposition reference area. Figure 10 shows an embodiment of rotating and correcting the image in Figure 9 and generating a fused reference inverted and corrected image. Figure 10 In

[0157] In order to superimpose the fused paired image on the image obtained by inverting and correcting the above-generated fusion reference, the paired feature positions in the fused paired image should be extracted so that the fused paired image can be accurately superimposed on the image obtained by inverting and correcting the fusion reference using the extracted paired feature positions, thereby forming a welded fusion image. It should be noted that the paired feature positions should be related to the formed fused paired image, and through the paired feature positions, the fused paired image can be accurately aligned with the image obtained by inverting and correcting the above-mentioned fusion reference. Hereinafter, taking the weld surface state image as the fusion reference image and the weld seam state image as the fused paired image as an example, the extraction of the paired feature positions in the fused paired image will be illustrated by way of example.

[0158] When the weld seam state image is configured as the fused paired image, the edge of the fused paired image can be used as the paired feature position. For example, Figure 2 and Figure 3 In the illustrated embodiment, the position corresponding to the left end of the second welding base 2 in the fused paired image can be used as the paired feature position, and the image obtained by inverting and correcting the fusion reference also includes the end position of the second welding base 2. Therefore, common technical means in the technical field can be used to extract the feature at the corresponding edge position in the fused paired image. Thereafter, the corresponding position is extracted on the image obtained by inverting and correcting the fusion reference as the reference alignment position. Align the paired feature position with the reference alignment position, and then the fused paired image can be superimposed on the image obtained by inverting and correcting the fusion reference to form a welded fusion image.

[0159] It should be noted that for a determined target overlapping weldment, a determined position or area can be selected as the paired feature position according to the type of the target overlapping weldment, that is, the selected paired feature position can be related to the type of the target overlapping weldment, etc. After selecting the paired feature position, common technical means in the technical field can be used to achieve image alignment, and the image alignment method and process can be consistent with the prior art, and will not be exemplified one by one here.

[0160] In an embodiment of the present invention, when the fused paired image is superimposed on the image obtained by inverting and correcting the fusion reference, the weld seam reference area and the superimposition reference area are mainly aligned and superimposed. During the alignment and superimposition, the binary values corresponding to the weld seam reference area are superimposed and calculated with the binary values corresponding to the superimposition reference area. Specifically, during the superimposition calculation, for any position of the superimposition calculation, when the binary value corresponding to the weld seam reference area or the binary value corresponding to the superimposition reference area is 1, the binary value of the pixel corresponding to the current superimposition calculation position is 1; when the binary values corresponding to both the weld seam reference area and the superimposition reference area are 0, the binary value of the pixel corresponding to the current superimposition calculation position is 0 only.

[0161] As can be seen from the above description of the superposition operation, the formed welding quality identification area should correspond to the superposition reference area, that is, the range of the welding quality identification area does not exceed the range of the superposition reference area. When the binarized value of the pixel after superposition is 1, the weld connection area can be formed, and the area other than the weld connection area in the superposition reference area can form the welding transition area. Figure 11 An embodiment of forming the welding quality identification area is shown in Figure 11 Among them, the middle black area is the weld connection area, and the white area outside the weld connection area is the welding transition area.

[0162] It can be understood that after selecting the fusion reference image and the fusion paired image, the above image fusion method is used for image fusion, and the required welding quality identification area, weld connection area and welding transition area can be obtained, and then the fusion ratio can be calculated.

[0163] In an embodiment of the present invention, when calculating the fusion ratio, it includes:

[0164] Statistical pixel area of the weld connection area and the pixel area of the welding quality identification area, and taking the ratio of the pixel area of the weld connection area to the pixel area of the welding quality identification area as the fusion ratio;

[0165] The quality detection ratio threshold includes a better quality ratio threshold and a non-conforming quality ratio threshold.

[0166] When the fusion ratio is not less than the better quality ratio threshold, the welding quality state of the target overlapping weldment is good welding quality;

[0167] When the fusion ratio is less than the non-conforming quality ratio threshold, the welding quality state of the target overlapping weldment is judged to be non-conforming welding quality.

[0168] As can be seen from the above method of forming the welding quality identification area, the pixel area of the weld connection area and the pixel area of the welding quality identification area can be statistically obtained by using common technical means in the technical field. Thereafter, the ratio of the pixel area of the weld connection area to the pixel area of the welding quality identification area can be calculated, and the calculated ratio is used as the fusion ratio.

[0169] When performing welding quality detection, a quality detection ratio threshold should also be set, such as at least setting a better quality ratio threshold and a non-conforming quality ratio threshold. Generally, the non-conforming quality ratio threshold can be 60%, and the better quality ratio threshold can be 90%. The better quality ratio threshold and the non-conforming quality ratio threshold can be specifically selected and determined according to the type of the target overlapping weldment and the welding quality detection standard, and will not be listed one by one here.

[0170] In an embodiment of the present invention, when the fusion ratio is not less than the quality unqualified ratio threshold and less than the better quality ratio threshold, the welding quality state of the target overlapping weldment is determined to be qualified, and the quality grade of the welding quality state is given, where

[0171] When determining the quality grade of the welding quality, it includes:

[0172] Extract the ultrasonic reflection signal corresponding to the welding fusion image, calculate the corresponding reflection signal feature intensity value based on the extracted ultrasonic reflection signal, and normalize the calculated reflection signal feature intensity value. Thereafter, generate the normalized reflection signal feature intensity histograms of the weld connection area, the welding transition area, and the unfused area based on the normalized reflection signal feature intensity value, where the unfused area is the area in the welding fusion image except the welding quality identification area;

[0173] For all the above-mentioned normalized reflection signal feature intensity histograms, calculate the histogram statistical information corresponding to each normalized histogram, where the histogram statistical information includes the mean, standard deviation, and kurtosis;

[0174] Determine the quality grade of the welding quality state based on the corresponding histogram statistical information of the weld connection area, the welding transition area, and the unfused area.

[0175] It can be understood that when the welding quality state of the target overlapping weldment is determined to be good, it can be regarded as the weldment detection of the target overlapping weldment passing, while when the welding quality state of the target overlapping weldment is unqualified, it can be regarded as the welding quality of the target overlapping weldment not passing. Specifically in implementation, for the calculated fusion ratio, there may be a situation where it is not less than the quality unqualified ratio threshold but less than the better quality ratio threshold. At this time, it can be regarded as the welding quality state of the target overlapping weldment being qualified. When the welding quality state is qualified, in order to further determine the welding quality, the quality grade corresponding to the welding quality state of the target overlapping weldment should also be determined to further judge the qualified state of the target overlapping weldment through the quality grade.

[0176] When determining the quality grade of the welding quality, the ultrasonic reflection signal corresponding to the welding fusion image should be extracted. In an embodiment of the present invention, when the fusion reference image is the weld surface state image, the weld ultrasonic reflection signals corresponding to the weld connection area, the welding transition area, and the unfused area in the weld state image can be extracted as the ultrasonic reflection signals corresponding to the weld connection area, the welding transition area, and the unfused area, where the unfused area is the remaining area in the weld state image other than the areas corresponding to the weld connection area and the welding transition area.

[0177] In specific implementation, after obtaining the welding fusion object, common technical means in this technical field can be used to determine the region positions corresponding to the weld connection region, the welding transition region, and the lack of fusion region in the weld state image. Thereafter, corresponding weld ultrasonic reflection signals can be extracted from the weld ultrasonic reflection signal group for generating the weld state image.

[0178] In specific implementation, after extracting the corresponding weld ultrasonic reflection signals, the corresponding reflection signal characteristic intensity values can be calculated. The reflection signal characteristic intensity values can be the same as the above-mentioned signal characteristic intensity values, such as the PPV value. After calculating the reflection signal characteristic intensity values of all the weld ultrasonic reflection signals, all the reflection signal characteristic intensity values can be normalized. The normalization method adopted can be selected according to needs and will not be elaborated here. Thereafter, using the normalized reflection signal characteristic intensity values and the corresponding states of the weld connection region, the welding transition region, and the lack of fusion region, a normalized reflection signal characteristic intensity histogram of the weld connection region, a normalized reflection signal characteristic intensity histogram of the welding transition region, and a normalized reflection signal characteristic intensity histogram of the lack of fusion region can be generated. The manner and process of generating the corresponding histograms can be consistent with the prior art and will not be elaborated here.

[0179] For all the above-mentioned normalized reflection signal characteristic intensity histograms, that is, the normalized reflection signal characteristic intensity histogram of the weld connection region, the normalized reflection signal characteristic intensity histogram of the welding transition region, and the normalized reflection signal characteristic intensity histogram of the lack of fusion region, the corresponding histogram statistical information is calculated. Among them, the histogram statistical information includes the mean value, the standard deviation, and the kurtosis. Thereafter, based on the corresponding histogram statistical information of the weld connection region, the welding transition region, and the lack of fusion region, the quality grade of the welding quality state can be determined.

[0180] In specific implementation, the quality grade can include the first grade, the second grade, and the third grade. The following is an example of how to determine the quality grade according to the histogram statistical information. Specifically:

[0181] The welding state corresponding to the first grade is: the distribution in the weld connection region is concentrated and sharp, and the distribution in the welding transition region is concentrated and close to the distribution characteristics of the weld connection region, indicating sufficient fusion and a continuous bonding interface. The corresponding judgment situations are:

[0182] , and,

[0183] Among them, is the mean value of the weld connection region, is the standard deviation of the weld connection region, is the kurtosis of the weld connection region, is the mean value of the welding transition region, is the standard deviation of the welding transition region, is the kurtosis of the welding transition region.

[0184] The welding state corresponding to the second level is: there are local lack of fusion points in the welding transition zone, but the overall transition trend is complete; the distribution of the characteristic intensity values of the reflection signals shows a wide peak characteristic, indicating the existence of microscopic inhomogeneity in the bonding state. The corresponding judgment situations are:

[0185] 、 and ,

[0186] or,

[0187] the peak value of the characteristic intensity value of the reflection signal in the welding transition region < the mean value of the histogram statistical information corresponding to the unfused region;

[0188] or,

[0189] 。

[0190] The welding state corresponding to the third level is: the distribution of the weld connection area is dispersed, the bonding state in the welding transition zone is unstable, the signals in the welding transition region and the unfused region are mixed, and there are obvious defects at the bonding interface. The corresponding judgment situations are:

[0191]

[0192] or,

[0193]

[0194] or,

[0195] the overlap degree of the distribution of the characteristic intensity values of the reflection signals in the welding transition region and the characteristic intensity values of the reflection signals in the unfused region > 30%.

[0196] In specific implementation, common technical means in this technical field can be used to calculate the distribution overlap degree between the characteristic intensity value of the reflected signal in the welding transition region and the characteristic intensity value of the reflected signal in the unfused region. For example, a feasible method is as follows: for the welding quality identification region, extract the ultrasonic reflected signal corresponding to the welding fusion image in the above manner. After that, normalize the extracted ultrasonic reflected signal, and respectively generate histograms of the characteristic intensities of the reflected signals corresponding to the welding transition region and the unfused region. Among them, the generated histogram consists of multiple intervals, each interval represents a set range of signal intensity values, and count the number of signals within this range. Generally, it is divided into 256 intervals. After that, for each interval, take the smaller value of the two histogram values for summation to obtain the overlapping area. Finally, divide this overlapping area by the total area of the histogram of the unfused region to obtain the value of the distribution overlap degree.

[0197] It should be understood that other technical means can also be used to determine the distribution overlap degree between the characteristic intensity value of the reflected signal in the welding transition region and the characteristic intensity value of the reflected signal in the unfused region. The specific determination method can be selected according to needs and will not be listed one by one here.

[0198] In an embodiment of the present invention, when the fusion ratio is not less than the quality unqualified ratio threshold and less than the quality better ratio threshold, the welding quality status of the target overlapping weldment is determined to be qualified, and the quality grade of the welding quality status is given, where

[0199] When the number of welding quality detections of the target overlapping weldment exceeds the detection number threshold, the quality grade of the determined welding quality status is determined by using the comprehensive quality coefficient, where

[0200] For the comprehensive quality coefficient, there is:

[0201]

[0202] In the formula, is the comprehensive quality coefficient, is the mean value of the welding transition region, is the standard deviation of the welding transition region, is the kurtosis of the weld connection region, 、 、 are the weight coefficients;

[0203] Calculate the comprehensive quality coefficients corresponding to all historical target overlapping weldments, and determine the quality grade threshold based on all the comprehensive quality coefficients.

[0204] Compare the comprehensive quality coefficient of the current target overlapping weldment with the determined quality grade threshold above to determine the quality grade of the welding quality status corresponding to the current target overlapping weldment.

[0205] It can be understood that in actual production, batch welding quality inspections are carried out on target overlapping weldments of the same type. However, when the number of target overlapping weldments inspected is small, the above-mentioned quality level determination method can be used. After welding quality inspections are carried out on a large number of target overlapping weldments, when making a quality level determination, the target overlapping weldments that have undergone welding quality inspections can be used for comprehensive judgment. Generally, the inspection quantity threshold can be 100, that is, when the number of target overlapping weldments of the same type undergoing quality welding inspections exceeds 100, the quality level can be determined through the comprehensive quality coefficient.

[0206] It should be noted that when determining the quality level of the welding quality state based on the comprehensive quality coefficient, the above-mentioned ultrasonic reflection signal extraction, calculation of the corresponding reflection signal characteristic intensity value, generation of a histogram, and statistics of the histogram to obtain histogram statistical information are still required. The specific method and process for obtaining the histogram statistical information can refer to the corresponding description above and will not be elaborated here. After obtaining the corresponding histogram statistical information, the comprehensive quality coefficient corresponding to the current target overlapping weldment can be determined. , specifically, the weight coefficient , the weight coefficient , the weight coefficient can be selected according to needs. For example, the sum of the weight coefficient , the weight coefficient , and the weight coefficient can be 1. Of course, it can also be other values, and no further examples will be given here.

[0207] For target overlapping weldments belonging to the same type of weldments, the comprehensive quality coefficient corresponding to the historical target overlapping weldments can be calculated, and the distribution of the comprehensive quality coefficient can be statistically analyzed to obtain the statistical value of the comprehensive quality coefficient corresponding to the historical target overlapping weldments. The statistical value of the comprehensive quality coefficient can be the mean value of the historical target overlapping weldments. Of course, it can also be other statistically obtained values. After obtaining the statistical value of the comprehensive quality coefficient, the 60% quantile of the statistical value of the comprehensive quality coefficient can be used as the second level threshold, and the 90% quantile of the statistical value of the comprehensive quality coefficient can be used as the first level threshold.

[0208] In specific implementation, if the comprehensive quality coefficient of the current target overlapping weldment is greater than the first level threshold, the quality level of the current target overlapping weldment is determined as the first level; if the comprehensive quality coefficient of the current target overlapping weldment is not less than the second level threshold and less than the first level threshold, the quality level of the current target overlapping weldment is determined as the second level; if the comprehensive quality coefficient When it is less than the second-level threshold, the quality level of the current target overlapping weldment is determined to be the third level.

[0209] It should be noted that the criteria for quality level division here are consistent with the above-mentioned quality level division criteria. Specifically, when the quality level is the first level, it indicates that the welding transition zone is closely combined and the weld connection zone is concentrated; when the quality level is the second level, it indicates that there are microscopic inhomogeneities in the welding transition zone; when the quality level is the third level, it indicates that the combination degree of the welding transition zone is poor.

[0210] In specific implementation, when judging the quality level status based on the comprehensive quality coefficient, there will be statistical deviations. Therefore, in order to accurately obtain the quality level of the welding quality status, fine-tuning processing is required, including at least the comprehensive quality coefficient fine-tuning processing and the quality level fine-tuning processing carried out in sequence. Specifically,

[0211] When carrying out the comprehensive quality coefficient fine-tuning processing, it includes:

[0212] Extract the edge contour of the weld connection area, calculate the curvature of each point on the edge contour, and statistically calculate the standard deviation of the curvature;

[0213] Compare the statistically obtained standard deviation of the curvature with the quality coefficient standard deviation threshold, and adjust the value of the comprehensive quality coefficient .

[0214] Specifically, the Canny algorithm can be used to extract the edge of the weld connection area to obtain the closed edge contour of the weld connection area. Figure 12 An embodiment of extracting the closed edge contour of the weld connection area is shown in. In the figure, the white curve is the closed edge contour of the weld connection area. Of course, other methods can also be used to extract the closed edge contour, and the method of extracting the closed edge contour can be selected according to needs, and will not be exemplified one by one here.

[0215] After extracting the closed edge contour, the curvature of each point on the closed edge contour can be calculated. The method of calculating the curvature of each point can be consistent with the prior art and will not be elaborated here. After calculating the curvature of each point, the standard deviation of the curvature can be statistically obtained by using the common technical means in the prior art. The method and process of statistically obtaining the standard deviation of the curvature can be consistent with the prior art and will not be elaborated here.

[0216] It can be understood that when the standard deviation of the curvature is low, the closed edge contour can be considered relatively smooth, while when the standard deviation of the curvature is high, the closed edge contour can be considered to have obvious concavities and convexities or serrations. In specific implementation, the quality coefficient standard deviation threshold can be set to 0.5. When the statistically obtained standard deviation of the curvature is less than 0.5, the above-mentioned calculated comprehensive quality coefficient Add the first adjustment coefficient. When the calculated standard deviation of curvature is not less than 0.5, the comprehensive quality coefficient calculated above can be subtracted by the second adjustment coefficient. Among them, the first adjustment coefficient can be 0.1, and the second adjustment coefficient can be 0.15. Of course, the first adjustment coefficient and the second adjustment coefficient can also be other values, which can be specifically selected according to actual needs to better characterize the welding quality of the target overlapping weldment.

[0217] When performing fine-tuning processing of the quality grade, it includes:

[0218] Extract the weld ultrasonic reflection signal corresponding to the welding transition area in the weld state image, and calculate the TOF value of the weld ultrasonic reflection signal corresponding to the welding transition area, so as to statistically calculate the TOF standard deviation and the TOF mean value of the welding transition area based on the calculated TOF value;

[0219] Calculate the TOF coefficient of variation of the welding transition area, then there is:

[0220]

[0221] Among them, is the TOF coefficient of variation, is the TOF standard deviation of the welding transition area, is the TOF mean value of the welding transition area;

[0222] When the TOF coefficient of variation < 10%, then confirm the current quality grade of the welding quality state. When the TOF coefficient of variation ≥ 10%, then force the quality grade of the confirmed welding quality state to be downgraded by one level to update the quality grade of the welding quality state. For example, when the previous quality grade is the first level, then when the TOF coefficient of variation ≥ 10%, the quality grade of the welding quality state should be configured as the second level.

[0223] It should be noted that, for example, when the previous quality grade is the third level, then when the TOF coefficient of variation < 10%, the quality grade of the welding quality state can be maintained as the third level, or the welding quality state of the target overlapping weldment can be judged as unqualified welding quality. Other situations can refer to the description here.

[0224] From the above description, a welding quality detection system suitable for overlapping weldments can be obtained. Specifically, it includes an ultrasonic scanning probe and a welding quality detection terminal. The ultrasonic scanning probe is adaptively connected to the welding quality detection terminal. Among them,

[0225] For any target overlapping welded part, perform ultrasonic scanning on the target overlapping welded part through an ultrasonic scanning probe. Thereafter, the welding quality detection terminal uses the above-mentioned welding quality detection method to perform quality detection and judgment to determine the welding quality status of the current target overlapping welded part.

[0226] Specifically, the welding quality detection terminal can be a commonly used computer terminal in the existing ones. The type of the welding quality detection terminal can be selected according to needs, and no further examples will be given here. The ultrasonic scanning probe can be the high-frequency probe mentioned above. The method and process of using the ultrasonic scanning probe and the welding quality detection terminal to detect the welding quality of the target overlapping welded part can refer to the above description and will not be elaborated here.

Claims

1. A welding quality detection method suitable for overlapping weldments, characterized in that, The described welding quality detection method includes: Providing a target overlapping weldment, and respectively performing ultrasonic scanning on two opposite surfaces of the target overlapping weldment to respectively obtain a weld ultrasonic reflection signal group related to the weld seam and a weld surface ultrasonic reflection signal group related to the weld surface; Generating a weld seam state image based on the weld ultrasonic reflection signal group, and generating a weld surface state image based on the weld surface ultrasonic reflection signal group. Among them, the weld seam state image includes a weld seam reference area representing the welding state, and the weld surface state image includes a weld surface reference area representing the weld surface range; Performing image fusion on the weld seam state image and the weld surface state image, and generating a welding fusion image, where During image fusion, at least align the weld seam reference area with the weld surface reference area, and fuse the weld seam reference area within the weld surface reference area to form a welding quality identification area in the welding fusion image. Among them, the welding quality identification area at least includes a welding connection area and a welding transition area; Based on the weld fusion image, calculate the fusion ratio of the welding connection area within the welding quality identification area, and compare the calculated fusion ratio with a quality detection ratio threshold to determine the welding quality state of the target overlapping weldment after comparison.

2. The welding quality detection method for overlapping weldments according to claim 1, characterized in that, Among the two opposite surfaces on which the ultrasonic scanning process is performed on the target overlapping weldment, at least the first surface and the second surface corresponding to the weld seam. Among the first surface and the second surface, at least one surface is a welding bearing surface; When performing ultrasonic scanning on each surface, it at least includes a surface scanning operation and a signal truncation operation performed in sequence. Among them, After performing the surface scanning operation on the current surface, a surface ultrasonic reflection source signal group is generated; When performing the signal truncation operation, perform signal truncation on the surface ultrasonic reflection source signal group to generate a surface ultrasonic reflection truncated signal group; When the current surface is a welding bearing surface, configure the surface ultrasonic reflection truncated signal group as the weld surface ultrasonic reflection signal group, otherwise, configure the surface ultrasonic reflection truncated signal group as the weld ultrasonic reflection signal group.

3. The welding quality detection method for overlapping weldments according to claim 2, characterized in that, in When performing ultrasonic scanning, it also includes a tilt correction operation. Among them, Before performing the signal truncation operation, first perform the tilt correction operation; When performing the tilt correction operation, generate a reflection scanning TOF array based on the surface ultrasonic reflection source signal group, and calculate and generate a flight time correction amount corresponding to each surface ultrasonic reflection source signal based on the reflection scanning TOF array at the smooth position; Use the calculated flight time correction amount to perform tilt correction on the corresponding surface ultrasonic reflection source signal to generate a surface ultrasonic reflection tilt-corrected signal after tilt correction; Generate a surface ultrasonic reflection tilt-corrected signal group based on all the surface ultrasonic reflection tilt-corrected signals; When performing the signal truncation operation, perform signal truncation on the surface ultrasonic reflection tilt-corrected signal group to generate a surface ultrasonic reflection truncated signal group after signal truncation.

4. The welding quality detection method for overlapping weldments according to claim 3, characterized in that, When calculating and generating the flight time correction amount corresponding to each weldment ultrasonic reflection scanning signal, it includes: Plane fitting is performed based on the reflected scan TOF values corresponding to the smooth positions within the reflected scan TOF array to generate a TOF fitting plane; For each surface ultrasonic scan point, the TOF fitting value of the current surface ultrasonic scan point is determined using the TOF fitting plane, and a time-of-flight correction amount is calculated and generated based on the determined TOF fitting value and the corresponding reflected scan TOF value within the reflected scan TOF array; During tilt correction, when the time-of-flight correction amount is non-integer, the surface ultrasonic reflection source signal is sequentially displacement-corrected based on the integer correction value and the decimal correction value, and a surface ultrasonic reflection tilt correction signal is generated after the displacement correction, where, The integer correction value is the integer part value of the time-of-flight correction amount; The decimal correction value is the decimal part value of the time-of-flight correction amount, and when performing displacement correction based on the decimal correction value, the displacement correction method includes linear interpolation.

5. The welding quality detection method for overlapping weldments according to any one of claims 1 to 4, characterized in that, When generating a weld state image, it includes: Generating a weld zone signal intensity image representing the intensity of the ultrasonic reflection signal based on the weld ultrasonic reflection signal group; Performing binarization processing on the weld zone signal intensity image to generate a weld state image after binarization processing, where, During binarization processing, two signal feature intensity peaks are searched in the weld zone signal intensity image, and the signal feature intensity valley value between the two searched signal feature intensity peaks is configured as the binarization segmentation threshold; The weld zone signal intensity image is binarized using the configured binarization segmentation threshold.

6. The welding quality detection method for overlapping weldments according to claim 5, characterized in that, When performing image fusion, it includes: Selecting the weld surface state image or the weld state image as the fusion reference image, and then configuring the weld state image or the weld surface state image as the fusion paired image; Mirror-flipping the fusion reference image to generate a fusion reference mirror-flipped image after mirror-flipping; Performing a logical inversion on the fusion reference mirror-flipped image and performing a rotation correction after the logical inversion to generate a fusion reference inverted and corrected image, where the fusion reference inverted and corrected image includes a superimposition reference area, and the binary value state of the superimposition reference area is different from the binary value state of the weld reference area; Extracting the paired feature positions within the fusion paired image and superimposing the fusion paired image on the fusion reference inverted and corrected image based on the extracted feature positions to form a welding fusion image, and generating a welding quality identification area within the welding fusion image, where, A weld connection area is formed based on the superimposition state of the weld reference area and the superimposition reference area, and the part of the superimposition reference area except for forming the weld connection area forms a welding transition area.

7. The welding quality detection method for overlapping weldments according to claim 6, characterized in that, When calculating the fusion ratio, it includes: Statistical the pixel area of the weld connection area and the pixel area of the welding quality identification area, and taking the ratio of the pixel area of the weld connection area to the pixel area of the welding quality identification area as the fusion ratio; The quality detection ratio threshold includes a better quality ratio threshold and a non-conforming quality ratio threshold, When the fusion ratio is not less than the better quality ratio threshold, the welding quality state of the target overlapping weldment is good welding quality; When the fusion ratio is less than the quality non - compliance ratio threshold, the welding quality status of the target overlapping weldment is determined to be unqualified.

8. The welding quality detection method for overlapping weldments according to claim 7, characterized in that, When the fusion ratio is not less than the quality non - compliance ratio threshold and less than the better quality ratio threshold, the welding quality status of the target overlapping weldment is determined to be qualified, and the quality grade of the welding quality status is given. Among them, When determining the quality grade of the welding quality, it includes: Extract the ultrasonic reflection signal corresponding to the welding fusion image, calculate the corresponding reflection signal feature intensity value based on the extracted ultrasonic reflection signal, and normalize the calculated reflection signal feature intensity value. After that, generate the normalized reflection signal feature intensity histogram of the weld connection area, the normalized reflection signal feature intensity histogram of the welding transition area, and the normalized reflection signal feature intensity histogram of the unfused area based on the normalized reflection signal feature intensity value. Among them, the unfused area is the area in the welding fusion image except the welding quality recognition area; For all the above - mentioned normalized reflection signal feature intensity histograms, calculate the histogram statistical information corresponding to each normalized histogram. Among them, the histogram statistical information includes the mean, standard deviation, and kurtosis; Based on the histogram statistical information of the weld connection area, the welding transition area, and the unfused area, determine the quality grade of the welding quality status.

9. The welding quality detection method for overlapping weldments according to claim 7, characterized in that, When the fusion ratio is not less than the quality non - compliance ratio threshold and less than the better quality ratio threshold, the welding quality status of the target overlapping weldment is determined to be qualified, and the quality grade of the welding quality status is given. Among them, When the number of welding quality detections of the target overlapping weldment exceeds the detection number threshold, the quality grade of the determined welding quality status is determined using the comprehensive quality coefficient. Among them, For the comprehensive quality coefficient, there is: In the formula, is the comprehensive quality coefficient, is the mean value of the welding transition area, is the standard deviation of the welding transition area, is the kurtosis of the weld connection area, , , are the weight coefficients; Calculate the comprehensive quality coefficients corresponding to all historical target overlapping weldments, and determine the quality grade threshold based on all the comprehensive quality coefficients. Compare the comprehensive quality coefficient of the current target overlapping weldment with the determined quality grade threshold above to determine the quality grade of the welding quality status corresponding to the current target overlapping weldment.

10. A welding quality detection system suitable for overlapping weldments, characterized in that, It includes an ultrasonic scanning probe and a welding quality detection terminal. The ultrasonic scanning probe is adaptively connected to the welding quality detection terminal. Among them, For any target overlapping weldment, perform ultrasonic scanning processing on the target overlapping weldment through the ultrasonic scanning probe. After that, the welding quality detection terminal uses the welding quality detection method in any one of claims 1 - 9 above to perform quality detection and judgment to determine the welding quality status of the current target overlapping weldment.

Images