CN-122023467-A - Method, device, system and storage medium for adjusting probe spacing

CN122023467ACN 122023467 ACN122023467 ACN 122023467ACN-122023467-A

Abstract

The invention discloses a probe interval adjusting method, a device, a system and a storage medium, wherein the method comprises the steps of acquiring multi-frame laser tomographic images shot by an image collector at any moment in a welding line area; the multi-frame laser fault image is input into an improved convolutional neural network model, a single-frame high-dimensional characteristic image is output, a plurality of single-frame high-dimensional characteristic images at a plurality of moments are obtained, fault displacement is calculated based on the plurality of single-frame high-dimensional characteristic images, and the distance between a probe and a welding seam central line is adjusted based on the fault displacement. The invention solves the technical problem that the distance between the probe and the center line of the welding line cannot be automatically adjusted in the prior art, and realizes the automatic adjustment of the distance between the probe and the center line of the welding line.

Inventors

ZHAO LEI
LIU JUNLI
CHEN GUODONG
LI SIHENG
ZHANG SHUISHENG
YAO ZHIHAN

Assignees

西安量子智能科技有限公司

Dates

Publication Date: 20260512
Application Date: 20251225

Claims (10)

1. A probe spacing adjustment method, comprising: acquiring multi-frame laser tomographic images shot by an image acquisition device at any moment in a welding line area; inputting the multi-frame laser tomographic image into an improved convolutional neural network model, and outputting a single-frame high-dimensional characteristic image; Acquiring a plurality of single-frame high-dimensional characteristic images at a plurality of moments, and calculating fault displacement based on the plurality of single-frame high-dimensional characteristic images; and adjusting the distance between the probe and the center line of the welding line based on the fault displacement.
2. The probe spacing adjustment method according to claim 1, wherein the improved convolutional neural network model comprises a shallow layer feature extraction module, a deep layer feature extraction and propagation module and a reconstruction module, wherein the deep layer feature extraction and propagation module adopts a feature fusion module in the propagation process, and the feature fusion module is used for aligning and weighting and summing features of different frames to obtain a fusion feature comprising multi-frame information; Wherein the improved convolutional neural network model employs a lightweight architecture design.
3. The method for adjusting the distance between probes according to claim 1, wherein the acquiring a plurality of frames of laser tomographic images taken by the image collector at any one time in the weld region includes: Responding to a first operation of a user, starting a laser to project laser stripes to the welding seam area at a preset angle, and starting the image collector to shoot the laser stripes for a plurality of times, wherein the image collector collects multi-frame laser fault images, and the laser stripes in the images generate faults due to the fact that the welding seam area is raised; and acquiring the multi-frame laser tomographic image from the image acquisition unit.
4. The probe spacing adjustment method according to claim 1, wherein the acquiring a plurality of single-frame high-dimensional feature images at a plurality of times and calculating a tomographic displacement amount based on the plurality of single-frame high-dimensional feature images includes: acquiring a first single-frame high-dimensional characteristic image at a first moment and a second single-frame high-dimensional characteristic image at a second moment; Calculating two-dimensional displacement fields of the first single-frame high-dimensional characteristic image and the second single-frame high-dimensional characteristic image by adopting an optical flow algorithm; and calculating fault displacement based on the two-dimensional displacement field.
5. The probe spacing adjustment method according to claim 4, wherein the calculating the tomographic displacement amount based on the two-dimensional displacement field includes: Identifying a fault line on the two-dimensional displacement field; respectively selecting a series of points at two sides of the fault line, and calculating displacement vector differences of the point pairs, wherein the displacement vector differences are horizontal dislocation quantity parallel to the fault line and vertical dislocation quantity vertical to the fault line; carrying out statistical analysis on the results of all the point pairs to finally generate pixel-fault displacement; correcting the camera model, and establishing a conversion relation between pixel coordinates and world coordinates; And obtaining the fault displacement based on the pixel-fault displacement and the conversion relation.
6. A probe spacing adjustment device, comprising: The laser tomographic image acquisition module is used for acquiring multi-frame laser tomographic images shot by the image acquisition device at any moment in the welding line area; The single-frame high-dimensional characteristic image generation module is used for inputting the multi-frame laser fault images into an improved convolutional neural network model and outputting single-frame high-dimensional characteristic images; the fault displacement solving module is used for acquiring a plurality of single-frame high-dimensional characteristic images at a plurality of moments and calculating fault displacement based on the plurality of single-frame high-dimensional characteristic images; and the probe interval adjusting module is used for adjusting the interval between the probe and the central line of the welding line based on the fault displacement.
7. The probe spacing adjustment device of claim 6, wherein the laser tomographic image acquisition module comprises: The laser tomography image shooting unit is used for responding to a first operation of a user, starting a laser to project laser stripes to a welding line area at a preset angle, and starting the image collector to shoot the laser stripes for a plurality of times, wherein the image collector collects a plurality of frames of laser tomography images, and the welding line area is raised, so that the laser stripes in the images generate faults; and the laser tomographic image communication unit is used for acquiring the multi-frame laser tomographic images from the image acquisition device.
8. The probe spacing adjustment device according to claim 6, wherein the tomographic displacement amount solving module includes: the multi-frame high-dimensional characteristic image acquisition unit is used for acquiring a first single-frame high-dimensional characteristic image at a first moment and a second single-frame high-dimensional characteristic image at a second moment; the two-dimensional displacement field acquisition unit is used for calculating the two-dimensional displacement fields of the first single-frame high-dimensional characteristic image and the second single-frame high-dimensional characteristic image by adopting an optical flow algorithm; And a fault displacement amount calculation unit for calculating a fault displacement amount based on the two-dimensional displacement field.
9. A probe spacing adjustment system, comprising: A probe for emitting ultrasonic waves to a workpiece to detect whether or not a crack occurs in the workpiece; The laser is used for projecting laser stripes to a welding line area of the workpiece at a preset angle; the image collector is used for shooting the laser stripes for a plurality of times so as to collect multi-frame laser fault images; A computer comprising a processor for performing the method of any one of claims 1 to 5.
10. A computer-readable storage medium storing computer-executable instructions for causing a computer to perform the method of any one of claims 1 to 5.

Description

Method, device, system and storage medium for adjusting probe spacing Technical Field The present invention relates to the field of robotics, and in particular, to a method, apparatus, system, and storage medium for adjusting probe spacing. Background Currently, welding quality detection of workpieces such as wind power towers mainly depends on inspection of robots, and the working principle of the welding quality detection is that ultrasonic waves are sent to the workpieces such as wind power towers through ultrasonic probes, and whether defects exist inside the workpieces is judged through ultrasonic wave echoes. In the robot inspection process, as the welding line area can generate bulges, the distance between the probe and the central line of the welding line can deviate from a standard value, and the standard value is generally plus or minus 2mm. Therefore, there is a need to develop a method for automatically adjusting the probe, which keeps the distance between the probe and the center line of the weld joint from exceeding the standard value. Disclosure of Invention The present invention aims to solve at least one of the technical problems in the related art to some extent. Therefore, an object of the present invention is to provide a probe spacing adjustment method, apparatus, system and storage medium, which can automatically adjust the distance between a probe and a weld center line when a robot performs inspection on the welding quality of a workpiece. The technical scheme adopted by the invention is as follows: The invention provides a probe spacing adjustment method, which comprises the steps of obtaining multi-frame laser tomographic images shot by an image collector at any moment in a welding line area, inputting the multi-frame laser tomographic images into an improved convolutional neural network model, outputting single-frame high-dimensional characteristic images, obtaining a plurality of single-frame high-dimensional characteristic images at a plurality of moments, calculating tomographic displacement based on the single-frame high-dimensional characteristic images, and adjusting the spacing between a probe and a welding line central line based on the tomographic displacement. The improved convolutional neural network model comprises a shallow layer feature extraction module, a deep layer feature extraction and propagation module and a reconstruction module, wherein the deep layer feature extraction and propagation module adopts a feature fusion module in the propagation process, the feature fusion module is used for aligning and weighting and summing features of different frames to obtain a fusion feature comprising multi-frame information, the reconstruction module comprises an up-sampling unit, a residual error learning unit and a convolution layer and is used for generating a single-frame high-dimensional feature image based on the fusion feature, and the improved convolutional neural network model adopts a lightweight architecture design. The method comprises the steps of responding to a first operation of a user, starting a laser to project laser stripes to a welding line area at a preset angle, starting the image collector to shoot the laser stripes for a plurality of times, wherein the image collector collects a plurality of frames of laser tomographic images, the welding line area is bulged, the laser stripes in the images generate faults, and the plurality of frames of laser tomographic images are obtained from the image collector. The method comprises the steps of obtaining a plurality of single-frame high-dimensional characteristic images at a plurality of moments, calculating fault displacement based on the plurality of single-frame high-dimensional characteristic images, obtaining a first single-frame high-dimensional characteristic image at a first moment and a second single-frame high-dimensional characteristic image at a second moment, calculating two-dimensional displacement fields of the first single-frame high-dimensional characteristic image and the second single-frame high-dimensional characteristic image by adopting an optical flow algorithm, and calculating fault displacement based on the two-dimensional displacement fields. The method comprises the steps of identifying a fault line on the two-dimensional displacement field, selecting a series of points on two sides of the fault line, calculating displacement vector differences of the point pairs, wherein the displacement vector differences are horizontal dislocation quantities parallel to the fault line and vertical dislocation quantities perpendicular to the fault line, carrying out statistical analysis on results of all the point pairs to finally generate pixel-fault displacement quantities, correcting a camera model, establishing a conversion relation between pixel coordinates and world coordinates, and obtaining the fault displacement quantities based on the pixel-fault displacement quantities and the convers