A new method is proposed for the characterization of the atmospheric turbulence strength parameter () based on the principal component analysis (PCA) of the texture of the speckle intensity pattern obtained on long-range propagation of a focused charge +1 vortex beam. Employing the split-step propagation method, datasets containing instantaneous intensity images of the focused vortex beam are generated for three distinct values, specifically , , and , representing medium to high turbulence levels for a 2 km propagation distance. The gray level co-occurrence matrix (GLCM) methodology is employed to extract key texture attributes, like contrast, correlation, homogeneity, and energy from the intensity images. These extracted texture parameters serve as inputs for training a PCA model, enabling the identification of associated values. The PCA analysis exhibits distinct clustering of the first three principal components for each of the three values, forming individual clusters on the PCA plot. Standard deviational ellipses are drawn to clearly demarcate these clusters on the PCA plot. The texture-based PCA classification of atmospheric turbulence was also performed for a focused Gaussian beam. The comparison of PCA plots between vortex and Gaussian beams showed that a pronounced clear separation of values is obtained for the vortex beam. This indicates that the non-zero orbital angular momentum of the vortex beam also plays an important role in achieving the distinct separation of values on the PCA plot. The proposed method can provide efficient real-time turbulence estimation solely on the basis of texture of the instantaneous intensity speckle with prior training and therefore may simplify the estimation of turbulence strength parameter.
Read full abstract