Abstract:

Deformation measurement at high temperatures is one of the hot and important issues in the study of the methods of mechanical properties experiments. In order to achieve accurate deformation measurements in a high temperature environment, this paper applies smallest univalue segment assimilating nucleus (SUSAN) corner detection and optical flow tracking matching techniques and proposes a method of high temperature deformation measurement based on SUSAN corner features. This method makes full use of the advantages of SUSAN corner features, and maintains the high accuracy, strong anti-interference ability, etc. By testing the simulated speckle images, the paper finds that the error is kept within 1%, and the algorithm has not only high accuracy, but also good stability. High temperature mechanics experiments show that this method can be applied to high temperature deformation measurements. This study extends the application of image recognition technology in the field of non-contact measurement technology.

Key words: high temperature, strain measurement, machine vision, image matching, feature extraction