Efficient averaging for enhanced resolution with acceptable quality in laser speckle imaging

Laser Speckle Imaging (LSI) identifies objects areas with different dynamics or properties for non-destructive testing. LSI raw data comprise temporally correlated images. The LSI output consists of activity maps representing 2D distributions of statistical estimates. Averaging is essential to mitigate the speckle nature of the raw data and enhance the sensitivity by achieving a higher contrast of the maps. Averaging over time at a single pixel or over multiple pixels improves spatial or temporal resolution. Given the strong temporal correlation and low spatial correlation of the raw LSI data, this study aims to develop a balanced averaging method that optimally exhibits good spatial and temporal resolutions while maintaining acceptable contrast in the activity map. The approach is validated by simulation and is applied to speckle data collected under tensile loading conditions, when dynamic speckle excitation is due to the sample extension.