CN-121977445-A - Image recognition-based abrasion degree detection system and method

CN121977445ACN 121977445 ACN121977445 ACN 121977445ACN-121977445-A

Abstract

The invention relates to the technical field of detection of the abrasion degree of an aircraft landing gear, and provides an abrasion degree detection system and method based on image recognition, which comprises an image acquisition unit for acquiring image data of a tread of a target tire in a detection state, the image acquisition unit at least comprises a structured light projection module and a high-resolution color camera, wherein the structured light projection module is used for projecting coded light patterns to the tread, and the image processing and analyzing unit is in communication connection with the image acquisition unit and is used for receiving the image data. By combining the structured light three-dimensional reconstruction and the high-resolution color imaging, the depth of the pattern groove can be accurately measured, two-dimensional texture defects such as tread cuts and cracks can be synchronously identified, positioned and quantified, and the two types of information are associated and analyzed by the fusion decision unit, so that a more comprehensive and three-dimensional representation of the health state of the tire is provided compared with single depth measurement.

Inventors

LI LIANTAO
WANG LEI
SUN XIAOXIA
MENG WEIFENG
SHI YONGJIE
QI HAORAN
Xiao Naijia
ZHANG JUNWEI
LI SIWEI
WANG YONGQI

Assignees

嘉兴职业技术学院

Dates

Publication Date: 20260505
Application Date: 20260125

Claims (10)

1. An image recognition-based wear degree detection system, comprising: the image acquisition unit is used for acquiring image data of the tread of the target tire in a detection state, and at least comprises a structured light projection module and a high-resolution color camera, wherein the structured light projection module is used for projecting a coded light pattern to the tread; An image processing and analyzing unit, which is in communication connection with the image acquisition unit, and is used for receiving the image data and executing the following processing of calculating three-dimensional point cloud data of the tire tread based on the coded light pattern and extracting two-dimensional texture characteristics from the color camera image; the abrasion measurement calculation unit is connected with the image processing and analyzing unit and is used for calculating depth distribution and residual depth of the tire tread groove based on the three-dimensional point cloud data and identifying and quantifying tread cuts, cracks and foreign matter embedding defects based on the two-dimensional texture features; And the fusion decision unit is respectively connected with the image processing and analyzing unit and the abrasion measurement computing unit and is used for fusing the three-dimensional depth information and the two-dimensional defect information to generate a comprehensive abrasion degree evaluation report and a maintenance suggestion.
2. The image recognition-based wear level detection system of claim 1, further comprising an off-line calibration unit including a standard step-shaped wear block of known accuracy for jointly calibrating the image acquisition unit in a non-detection state to correct dimensional errors and camera distortions of the three-dimensional reconstruction.
3. The image recognition-based wear level detection system according to claim 2, wherein the off-line calibration unit and the image processing and analysis unit are integrated, and the calibration process is performed by automatically recognizing specific identification points on the standard step-shaped wear blocks and calculating the re-projection errors, and the calibration parameters are stored and used for on-line correction of all subsequent on-machine detection tasks.
4. The image recognition-based wear detection system according to claim 1, wherein the algorithm adopted by the image processing and analysis unit to perform three-dimensional reconstruction is a multi-frequency heterodyne method combining a phase shift method and gray codes, and is used for solving the problem of phase unwrapping of a tire tread high-reflection area and a pattern edge, and the core formula relates to absolute phase Is calculated by (1): wherein For the phase-shifted fringe image gray scale values, Is the number of stripe stages resulting from the gray code decoding.
5. The image recognition-based wear degree detection system according to claim 1, further comprising an intelligent sensing triggering unit, wherein the intelligent sensing triggering unit comprises a proximity sensor and a hub rotation speed sensor, and the image acquisition unit is automatically triggered to operate when the proximity sensor detects that the tire enters a preset detection area and the hub rotation speed sensor judges that the tire is in a static state or a low-speed rotation state.
6. An image recognition-based abrasion degree detection method using the image recognition-based abrasion degree detection system according to any one of claims 1 to 5, characterized by comprising the steps of: Step one, system initialization and calibration, namely calling the off-line calibration unit, and performing joint calibration on a structural light module and a color camera by using the standard step-shaped abrasion block to acquire and store high-precision internal and external parameters of the system; Monitoring the position and the state of the tire through the intelligent sensing triggering unit, and automatically sending out an image acquisition instruction when the preset condition is met; Synchronously acquiring multi-mode images, namely, the structured light projection module projects coded light patterns to the tire tread which is static or rotates slowly, and synchronously acquiring modulated tread images by the color camera; Decoding the acquired image, calculating the absolute phases of each point of the tread, converting the absolute phases into three-dimensional point clouds, simultaneously, dividing the tread area from the color image, and extracting texture features; Fifthly, carrying out abrasion characteristic quantification calculation, namely positioning pattern grooves along the circumferential direction and the axial direction of the tire in a three-dimensional point cloud, calculating a depth curve and the minimum residual depth of the cross section of the grooves, identifying and framing various surface defects by using a trained deep learning model in a two-dimensional image, and calculating area and length quantification indexes of the surface defects; and step six, information fusion and report generation, namely establishing a mapping relation between the three-dimensional depth data and the two-dimensional defect information, and outputting a comprehensive evaluation report containing quantitative wear data, a defect distribution diagram and specific maintenance suggestions according to a preset weight fusion algorithm.
7. The method for detecting the abrasion degree based on image recognition according to claim 6, wherein in the first step, the joint calibration process comprises the steps of obtaining multiple groups of images of the standard abrasion blocks from different angles, and simultaneously optimizing parameters of the structured light projection module and the color camera through a beam adjustment method to ensure that a three-dimensional measurement result is consistent with a physical scale.
8. The method of claim 6, wherein in the fourth step, the calculation of the three-dimensional point cloud further includes preprocessing the point cloud to remove noise, filter and repair holes, and registering the current point cloud with a point cloud template of a standard new tire using an iterative nearest point algorithm to eliminate measurement deviation caused by the tire installation posture.
9. The method of claim 6, wherein in step five, the deep learning model is a modified U-Net network that introduces a attentive mechanism in the encoder portion to enhance the extraction of micro-cuts and micro-cracks features.
10. The method of claim 6, wherein in the sixth step, the fusion decision algorithm is to set a safety threshold pattern depth and a defect comprehensive index, report priority as "immediately replace" when the minimum residual depth is measured < pattern depth, report priority as "closely paying attention to, shorten the inspection cycle" when the minimum residual depth is measured not less than pattern depth but the defect index > defect comprehensive index, otherwise report as "normal state".

Description

Image recognition-based abrasion degree detection system and method Technical Field The invention relates to the technical field of detection of abrasion degree of an aircraft landing gear, in particular to an abrasion degree detection system and method based on image recognition. Background The aircraft tire is used as a key component of the landing gear, directly bears landing impact, sliding load and braking force, and the abrasion state directly influences the flight safety and the operation economy. Currently, the inspection of the wear level of aircraft tires is primarily dependent on visual inspection of the crew's ground and pattern depth measurements using simple gauges (e.g., depth gauges). This traditional approach relies heavily on the experience and responsibility of the personnel. With the continuous improvement of the requirements of the aviation industry on the operation efficiency and safety margin, the existing manual detection mode faces some objective challenges. First, inspection efficiency and consistency are to be improved. In a short time window of a flight fast stop, a fine check of a plurality of tires is performed for the existing time pressure, which may cause a miss risk. Secondly, for surface defects such as fine cuts and partial cracks existing in the tire tread, standardized and quantitative evaluation is difficult to carry out by artificial vision, and the subjectivity of the judgment standard is strong. Moreover, the existing measurement means mainly focuses on the single dimension of the depth of the pattern groove, lacks the capability of comprehensively correlating and analyzing the depth abrasion and the surface texture damage, and cannot comprehensively evaluate the overall health condition of the tire, so that a technical means capable of realizing rapid, objective and quantitative detection is needed. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a wear degree detection system and method based on image recognition, which solve the problems of dependence on manpower, low efficiency and single evaluation dimension in the existing aircraft tire wear detection. In order to achieve the above purpose, the invention is realized by the following technical scheme that the abrasion degree detection system based on image recognition comprises: the image acquisition unit is used for acquiring image data of the tread of the target tire in a detection state, and at least comprises a structured light projection module and a high-resolution color camera, wherein the structured light projection module is used for projecting a coded light pattern to the tread; An image processing and analyzing unit, which is in communication connection with the image acquisition unit, and is used for receiving the image data and executing the following processing of calculating three-dimensional point cloud data of the tire tread based on the coded light pattern and extracting two-dimensional texture characteristics from the color camera image; the abrasion measurement calculation unit is connected with the image processing and analyzing unit and is used for calculating depth distribution and residual depth of the tire tread groove based on the three-dimensional point cloud data and identifying and quantifying tread cuts, cracks and foreign matter embedding defects based on the two-dimensional texture features; And the fusion decision unit is respectively connected with the image processing and analyzing unit and the abrasion measurement computing unit and is used for fusing the three-dimensional depth information and the two-dimensional defect information to generate a comprehensive abrasion degree evaluation report and a maintenance suggestion. Preferably, the system further comprises an off-line calibration unit, wherein the off-line calibration unit comprises a standard step-shaped abrasion block with known precision and is used for jointly calibrating the image acquisition unit in a non-detection state so as to correct dimensional errors and camera distortion of three-dimensional reconstruction. Preferably, the off-line calibration unit and the image processing and analyzing unit are integrated, the calibration process is completed by automatically identifying specific identification points on the standard step-shaped abrasion blocks and calculating the reprojection errors, and the calibration parameters are stored and used for on-line correction of all subsequent on-machine detection tasks. Preferably, the algorithm adopted by the image processing and analyzing unit to execute three-dimensional reconstruction is a multi-frequency heterodyne method combining a phase shift method and Gray codes, and is used for solving the phase unwrapping problem of the tire tread highly reflective area and pattern edges, and the core formula relates to absolute phaseIs calculated by (1): wherein For the phase-shifted fringe image gray scale values,Is the number of stripe