CTTI CT for Tire Industry

CT for Objects of Regular Textures: Simulation and Detection

Motivation:

The most common technique for computerized tire inspection is the analysis of two dimensional X-ray images. The absence of any information regarding the third dimension makes the process of detecting some anomalies in the tire construction very hard, impossible at times. Computer
Tomography whose purpose is the generation of three dimensional images of the inspected object is still not effectively used in tire industry so far. This work aims at investigating the advantages and the limitations of CT, and whether this technique can be used in the tire inspection process.

Approach:

The idea of this project is to develop an effective way for using CT setup and image analysis in tire industry, and thus the most suitable machine design. Until now, there are no machines for tires' CT that can deliver 3D data or 3D images. Therefore, we will model the tire mathematically. This will enable us to generate a database of different tires of known constructions. Additionally, errors can be simulated and included in the model.

Tahe modeled 3D object will be used with different setups of X-ray detectors and tubes. For each setup, data acquiring and detector readings are simulated. The acquired data can be used to reconstruct a 3D model of the tire. Different CT scan reconstruction filters will be applied and verified to optimize the object detection and segmentation of the tire parts.

The approach can be divided into the following stages: Tire Modeling, Machine Modeling & Scan Simulation, Tire Reconstruction and finally the Tire inspection. These phases are shown in the next diagram.

picture7

 

Tire Modeling

The tire is normally constructed of rubber and steel wires. The rubber has a certain pattern and inside the rubber is a mesh of steel wires with a well known arrangement. This can be considered as a revolutionary object around an axis. A detailed model of the tire construction will be developed. The relative attenuation factor of the different materials has to be taken into consideration. The VRML (Virtual reality Modeling Language) can be used to view the constructed model. In summary

  • 3D X-ray tire scans are not available.
  • Rigorous modeling is needed.
  • ?The Models cover the all tire parts.
  • Four tire anomalies are also modeled.

Machine Modeling and Tire Scan Simulation

The X-ray tube has several emission patterns like central and parallel projection, and can have a wide range for its power values. The X-ray detectors have also different types and arrangements, these could be linear or circular etc. The best arrangement of the X-ray tube and the detector will be decided according to the other steps and to minimum cost criterion. For the several X-ray setups the CT-scan can be simulated. The determining parameters for that are the angle step and the slice width. Since the beam attenuation can be determined according to the substance that the X-ray passes through, we can estimate the detector reading.

  • Many choices for the X-ray machine setup.
  • X-ray detector readings are to be estimated.
  • Mathematical intersection between each beam and the tire’s components.
  • Two factors: distance and material.

Tire Reconstruction

The theory of the CT-scan is well established, but different filters and reconstruction methods can be applied. For our specific application we might need a certain specialized filter which must be investigated in the project.
Beside the time latency of the algorithm, quality measure metrics must be defined to compare different reconstruction filters. Those metrics can include the robustness of error detection and other metrics which will be determined in the project.

  • The theory of the computed tomography is well established .
  • An object can be totally reconstructed from the X-ray detectors readings.
  • Reconstruction is done per cross section.
  • Efficient MATLAB implementation is used.

3D Image Analysis & Tire Inspection

The simulated reconstructed 3D images will be analyzed taking into consideration different error sources. These sources include added artificial irregularities, the reconstruction noise and added simulated noise from the X-ray setup.
The analysis processes includes 3D image segmentation, tire parts detection and anomaly detection of different known errors.
  • Many image processing techniques were applied including: filtering, segmentation, sharpening, and edge detections.
  • Clustering techniques to distinguesh between tire layers.

 

 



picture1 picture2
picture3 picture4
picture5 picture6

 

CTTI Machine

More Info

Our Motto

 

Commitment to Innovation

Our Mission

To provide the customer with a highly accurate, efficient, reliable, and cost-effective solution. Raising Product quality and production efficiency is the center of our attention.

Newsletter

Subscribe