CN-122016797-A - Online quality inspection system of thermos cup welding station based on machine vision

CN122016797ACN 122016797 ACN122016797 ACN 122016797ACN-122016797-A

Abstract

The invention relates to the technical field of online quality inspection, in particular to an online quality inspection system of a vacuum cup welding station based on machine vision, which comprises a positioning acquisition module, a multi-mode imaging module and an image processing module, wherein the positioning acquisition module is used for generating a trigger signal when a cup body reaches a preset position, the multi-mode imaging module is used for acquiring a multi-angle two-dimensional image and a three-dimensional contour image of a welding area according to the trigger signal, and the image processing module is used for fusing the multi-angle two-dimensional image and the three-dimensional contour image of the welding area through the image processing module and a defect analysis module to inhibit bright spots, so that small gaps are not covered by reflection, the false alarm is reduced, meanwhile, the hidden danger of small but air leakage is discovered in advance, namely, the amplification effect of fine line light on convex-concave is overlapped, the early risk is convenient to discover, and the air leakage risk which is exposed only to the rear section is moved forward to the welding station, so that the waste of reworking, scrapping and invested man-hour of the rear section is reduced.

Inventors

Mao Yunqiang
Cui Qinzhu
ZHONG YIWEN
Deng Huangfei
LI XIANGDONG

Assignees

浙江金维克家庭用品科技有限公司

Dates

Publication Date: 20260512
Application Date: 20251223

Claims (10)

1. The vacuum cup welding station online quality inspection system based on machine vision is characterized by comprising a positioning acquisition module, a positioning acquisition module and a control module, wherein the positioning acquisition module is used for generating a trigger signal when a cup body reaches a preset position; The multi-mode imaging module is used for acquiring images according to the trigger signals and acquiring multi-angle two-dimensional images and three-dimensional contour images of the welding area; the image processing module is used for generating a synthetic image by adopting a fusion algorithm for taking a median value according to the multi-angle two-dimensional image; the defect analysis module is used for extracting a plurality of characteristic indexes from the synthetic image and the three-dimensional contour image, automatically adjusting a judgment threshold, and presetting a risk evaluation model to calculate a comprehensive risk score; And the control and tracing module outputs a production line control signal according to the comprehensive risk score and records the whole flow data.
2. The machine vision-based vacuum cup welding station online quality inspection system of claim 1, wherein the operation flow of the positioning acquisition module comprises: the positioning acquisition module comprises an in-place sensor, an industrial camera and a trigger circuit, wherein the in-place sensor is used for detecting whether a cup body enters a preset imaging station or not and generating a trigger signal, and the trigger signal needs to be subjected to pretreatment based on wavelet threshold denoising, and the expression is as follows: , In the formula, For the trigger signal generated after the preprocessing, And The wavelet transform and the inverse transform are respectively carried out, For a soft threshold function set for high frequency noise, The trigger signal is synchronized with the exposure time of the industrial camera as the original binary signal.
3. The machine vision-based vacuum cup welding station online quality inspection system of claim 2, wherein the operation flow of the multi-modality imaging module comprises: The multi-angle two-dimensional image comprises at least three LED light sources with adjustable independent brightness and angle, which are arranged around an imaging station and used for irradiating welding areas from three different directions, and the three independent LED light sources are controlled according to the trigger signals The industrial camera is synchronously controlled to shoot according to the preset time sequence to obtain three welding area images with different illumination angles, and when each light source is lighted, the camera synchronously exposes, and the direction vector of the light source is set as A normal vector of a certain point on the surface of the welding area is Its brightness A simplified model conforming to a bi-directional reflection distribution function: , In the formula, Represent the first The camera sensor receives the image brightness of the point under the irradiation of the LED light sources, As a function of the reflectivity of the surface of the solder joint material, To express the first The intensity of illumination produced by the individual LED light sources at that point, Is the ambient light brightness, and further obtains 。
4. A machine vision based vacuum cup welding station online quality inspection system as claimed in claim 3, wherein the operation flow of the multi-modality imaging module further comprises: The three-dimensional contour image comprises that after the multi-angle two-dimensional image is obtained, all LEDs are turned off, a laser line is projected onto a welding bead by using a line laser projector, the high-speed camera collects the three-dimensional contour image formed by the laser line on the surface of the welding bead, and a laser plane equation calibrated in advance is used for obtaining the three-dimensional contour image And camera reference matrix , wherein, Representing the plane equation of the light projected by the laser, Is the normal vector of the plane surface and, Is a constant term, is a coordinate variable of a three-dimensional space point, and is used for image points Back-projected onto the laser plane, wherein, The transpose is represented by the number, , Representing pixel coordinates of a three-dimensional point to be solved on a two-dimensional image plane to obtain the three-dimensional point The expression: In which, in the process, Representing a non-zero scaling factor, The extrinsic matrix representing the camera is a 3x3 rotation matrix, Is a translation vector of 3x1, and the equation set is solved to obtain the laser center line three-dimensional point cloud 。
5. The machine vision-based vacuum cup welding station online quality inspection system of claim 4, wherein the operation flow of the image processing module comprises: According to the multi-angle two-dimensional image, adopting a fusion algorithm for taking the median value, and for each pixel position Calculation of Local contrast at this location And local information entropy As a quality evaluation index, expression: , In the formula, Is shown in the first A local contrast index in the image of the light source, Is that , And To take the following measures Is the standard deviation and mean in the neighborhood of the center, Is shown in the first The local information entropy in the image of the light source, Represent the first Probability of occurrence of a gray level in the current local area.
6. The machine vision-based vacuum cup welding station online quality inspection system of claim 5, wherein the operation flow of the image processing module comprises: assigning a fusion weight to each pixel of each graph The expression: , In the formula, And Is an index after normalization, and the index is a normalized index, 、 To adjust parameters, the image is synthesized The pixel values of (2) are: , In the formula, Representing the final generated composite image at pixel coordinates The brightness value at which the brightness value is to be obtained, Is shown in the first Position in an original input image The proportion of the contribution of the pixel value at that point to the final fusion result, Represent the first The original input image is displayed in pixel coordinates At the original luminance value.
7. The machine vision-based vacuum cup welding station online quality inspection system of claim 6, wherein the operation flow of the defect analysis module further comprises: the multiple characteristic indexes comprise extracting weld bead contour by edge detection, calculating circumference to evaluate continuity, counting whether the welding ring/welding spot is broken and the breaking length, obtaining information about whether the welding is broken, and setting continuity index Is in front of The ratio of the energy of each low-frequency descriptor to the total energy is expressed as follows: , In the formula, In order to make a fourier description of the object, The closer to 1, the better the continuity, the sampling width along the normal direction of the outline of the weld bead, the standard deviation of the sampling width is calculated to evaluate the uniformity of the width, the fluctuation of the width of the welding spot is counted to obtain the information whether the welding spot is wide or narrow, the micro suspected leakage points and splash adhesion are detected by morphological opening operation and spot analysis, and the area and the quantity of the micro suspected leakage points and splash adhesion are counted.
8. The machine vision-based vacuum cup welding station online quality inspection system of claim 7, wherein the operation flow of the defect analysis module further comprises: Extracting a laser center line from the three-dimensional contour image, comparing the laser center line with an ideal straight line, calculating the maximum bending distance and the average offset of the laser center line, taking the maximum bending distance and the average offset as quantization indexes of flatness and welding offset, and taking the laser center line three-dimensional point cloud as a reference point Projection onto an ideal plane fitted by least squares On the plane, calculate the directed distance of each point to the plane The flatness index is quantized to the standard deviation of the distance, the expression: , In the formula, Is an index of the flatness, is used for the flatness, Represent the first Three-dimensional points to a fitting plane Is used for the directional distance of the (c), Representing directed distance Is quantized to the average lateral offset of the laser centerline three-dimensional point cloud from an ideal straight line 。
9. The machine vision-based vacuum cup welding station online quality inspection system of claim 8, wherein the operation flow of the defect analysis module further comprises: By calculating the reflection intensity index, the expression: , In the formula, Is an index of the intensity of the light reflected, Is the average of the first 10% pixels of brightness, As global average, defect determination threshold Along with it Dynamic adjustment, expression: , In the formula, Represents the defect decision threshold after the adaptation, Represents a reference value determined for the commissioning of a good product under standard lighting conditions, Is a function of the hyperbolic tangent, Is the reflective intensity index calculated in real time, In order to adjust the coefficient of the coefficient, The method comprises the steps of establishing a second-level fuzzy comprehensive evaluation model, taking the first level as an index layer, converting the extracted six indexes of continuity, width uniformity, suspected leakage points, splash adhesion, flatness and welding bias into a fuzzy comment set through a membership function Membership vector on , wherein, Represent the first The second level is a factor layer, and weight vectors are given to each group of indexes , wherein, Represents the first Importance of each index in final evaluation, and comprehensive judgment vectors are calculated through fuzzy synthesis operation , wherein, Is a matrix of fuzzy relationships that is a fuzzy relationship, Is a fuzzy synthesis operator, which is used for synthesizing the data, Then, the comprehensive judgment vector is carried out Defuzzifying, and mapping to 0-100 points of comprehensive risk score by weighted average The lower the score, the higher the risk.
10. The machine vision-based vacuum cup welding station online quality inspection system of claim 9, wherein the control and traceability module operation flow further comprises: outputting a production line control signal according to the comprehensive risk score and recording full-flow data, wherein the method comprises the steps of presetting a release threshold value And a minimum risk tolerance threshold In the following When the control signal of the production line is outputted as release, when Outputting the line control signal as repair, and outputting the defect type and the circumferential angle position calculated based on the image coordinates, in Outputting a production line control signal as rejection; by generating a globally unique traceback code And packaging the multi-angle two-dimensional image, the three-dimensional contour image, the composite image, the continuity, the width uniformity, the suspected leakage point, the splash adhesion, the flatness, the welding bias, the comprehensive risk score, the production line control signal, the product ID and the timestamp into a data packet in a key value pair mode, and writing the data packet into a time sequence database and a relational database through asynchronous transactions.

Description

Online quality inspection system of thermos cup welding station based on machine vision Technical Field The invention relates to the technical field of online quality inspection, in particular to an online quality inspection system of a vacuum cup welding station based on machine vision. Background The machine vision is a technical system which integrates the technologies of multiple fields such as optics, computer science, image processing, mode recognition and the like, the core aim is to enable a machine to 'understand' image or video information like human beings and finish tasks such as measurement, detection, recognition, positioning and the like based on visual data, the online quality inspection of a vacuum cup welding station is a technical system for carrying out real-time, automatic and omnibearing quality monitoring on welding procedures on a vacuum cup production line, and the quality of welding seams is ensured to meet standard requirements by various detection means so as to prevent defective products from flowing into subsequent procedures. The patent publication No. CN201611129833.3 is recorded in the specification as an online production quality inspection system and method, wherein the system comprises a parameter acquisition unit, a standard parameter storage unit and a data analysis unit, wherein the parameter acquisition unit is used for respectively acquiring technological parameters of all online production equipment in the production process of products, the standard parameter storage unit is used for storing quality indexes of qualified products, the data analysis unit is used for comparing and analyzing all the technological parameters and the quality indexes to judge whether the products are qualified or not, generating product qualification notification information, searching corresponding online production equipment according to the technological parameters which do not meet the quality indexes and generating alarm information, and the notification unit is used for sending the product qualification notification information or the alarm information to a preset notification object. The invention realizes the quality detection of online production by collecting each technological parameter and comparing and analyzing with the pre-stored quality index, searches the corresponding online production equipment according to the technological parameter which does not accord with the quality index, generates and sends alarm information, and realizes the timely searching of the production equipment with problems. The prior art scheme has the advantages that in the production of the prior art, the welding position related to sealing and air leakage prevention is exposed by means of subsequent air leakage or heat preservation verification, so that hysteresis is found, high material and labor hour investment are generated, meanwhile, the cup body is a metal cylindrical surface, the reflection intensity and the bright spots are large, the traditional camera is easy to locally excessively lighten or shade and shade when photographing after welding, so that tiny abnormalities such as unwelded, welding bias, burning-through, splashing adhesion and the like are difficult to stably identify, missed detection or misjudgment occurs, the heat preservation performance of the product is reduced, and the goods returning and brand risk are caused. In summary, developing an online quality inspection system for a vacuum cup welding station based on machine vision is still a key problem to be solved in the technical field of online quality inspection. Disclosure of Invention The invention aims to solve the problems that in the prior art, a welding position related to sealing and preventing air leakage is exposed only by a follow-up air leakage or heat preservation verification stage, so that hysteresis is found, high material and working hour investment are generated, meanwhile, a cup body is a metal cylindrical surface, the reflection intensity and the bright spots are large, and the traditional camera is easy to generate local too bright or shadow shielding when photographing after welding, so that tiny abnormalities such as unwelded, welding bias, burning-through, splashing adhesion and the like are difficult to stably identify, missed detection or misjudgment occurs, the heat preservation performance of a product is reduced, and the replacement and brand risk are caused. In order to achieve the aim, the invention provides an online quality inspection system of a vacuum cup welding station based on machine vision, which comprises a positioning acquisition module, a positioning acquisition module and a control module, wherein the positioning acquisition module is used for generating a trigger signal when a cup body reaches a preset position; The multi-mode imaging module is used for acquiring images according to the trigger signals and acquiring multi-angle two-dimensional images and three-dimensional cont