Systems design of a high-resolution, large-data set, ultrasonic tire inspection machine.

摘要

Inspection of semi-truck tires helps ensure the safety of all highway travelers. Additionally, product transport costs can be reduced by improving tire reliability. While tire inspection can take many forms, non-destructive evaluation by ultrasound has been used for many years. In particular, a clear progression of ultrasound inspection techniques has lead from simple point testing devices to more complex systems that can automatically inspect a tire. However, the traditional methods of inspection have speed and resolution limitations.; A functionally complete, high-resolution, large data set, ultrasonic tire inspection machine was developed and represents an advancement in the way it inspects tires. The system development is motivated by the desire of a large tire retreading company to perform high-resolution ultrasound inspection of semi-truck tires. The inspection device for truck tires generates complete tire flaw maps in a timely fashion where an inspection can be performed at 1/8&inches; resolution in under 10 minutes. This represents an improvement greater than one order of magnitude over previous techniques that required 12 hours for a partial tire scan. The system inspection capability is demonstrated and validated.; A challenge of ultrasonic inspection methods is the large system point spread causing blurring of the signal. To satisfy resolution requirements and enhance flaw imaging, an application of double Wiener filtering to recover the system point spread function in ultrasonic tire inspection imaging is also developed. The double Wiener deconvolution is performed using an algorithm developed for this application. The algorithm has the capability of directly enhancing the blurry image produced by a given tire inspection when a known flaw is present on the tire. Improvements to the tire image by the filtering provide an enhancement recovering true flaw data from an indicated flaw data over twice the spatial size of the actual flaw. So a 4 x 4 pixel flaw that is detected as 8 x 8 pixels is adequately recovered after filtering as a 4–5 x 4–5 sized flaw.; Contributions resulting from this work are the development of an advanced high speed tire inspection methodology and apparatus. Additionally, to demonstrate system effectiveness, a deconvolution technique was applied that recovered true flaw size information from blurry flaw images. In conclusion, this work shows that tire inspection techniques can be dramatically improved with an additional improvement possible in final image quality by proper filtering. Future work possible in this area includes the streamlining of the hardware apparatus that was developed; an in depth testing and analysis to provide ultimate flaw recognition capabilities; and certainly homomorphic deconvolution techniques could be applied to eliminate the need to include a known flaw in an image prior to filtering.