GPUCorrel: A GPU accelerated Digital Image Correlation software written in Python

Victor Couty (1) , Jean-François Witz (1) , Pauline Lecomte-Grosbras (1) , Julien Berthe (2, 1) , Eric Deletombe (2, 1) , Mathias Brieu (3)

1 LaMcube - Laboratoire de Mécanique, Multiphysique, Multiéchelle - UMR 9013
2 DMAS, ONERA [Lille]
3 CAL STATE LA - California State University [Los Angeles]

This article presents an open-source Integrated Digital Image Correlation (I-DIC) software written in Python using CUDA-enabled GPUs designed to run at high (1-100 Hz) frequency. The field computation is performed using a global approach and the result is a projection of the real field in a user-defined base of fields. This software can be used in many applications and one use in experimental mechanics is demonstrated by driving a bi-axial tensile test on a cruciform specimen.

Mots clés: DIC; Digital Image Correlation; GPU; Mechanics; CUDA; Python; Strain; Measurement

Domaines: Sciences de l'ingénieur [physics]; Physique [physics]; Chimie

Victor Couty, Jean-François Witz, Pauline Lecomte-Grosbras, Julien Berthe, Eric Deletombe, et al.. GPUCorrel: A GPU accelerated Digital Image Correlation software written in Python. SoftwareX, 2021, 16, pp.100815. ⟨10.1016/j.softx.2021.100815⟩. ⟨hal-03380269⟩