New Image Texture Analysis, and Application to Polymer Membrane Surface Morphologies and Roughness.

Area of Science:

• Materials Science
• Image Analysis
• Surface Science

Background:

• Image texture analysis is crucial for characterizing material surfaces.
• Existing methods may be computationally intensive or limited in scope.
• Scanning electron microscopy (SEM) provides high-resolution surface imaging.

Purpose of the Study:

• To introduce a new, computationally efficient image texture analysis method.
• To apply the method to SEM images of reverse osmosis membranes.
• To correlate texture analysis results with surface morphology and roughness.

Main Methods:

• Utilized mean and standard deviation of gray levels within image domains.
• Employed recursive calculations on domains of varying sizes.
• Input gray level data into principal component analysis (PCA).
• Analyzed images holistically, not for segmentation, across all scales and directions.

Main Results:

• The method successfully identified surface morphology in SEM images.
• Texture analysis accurately predicted surface roughness, as validated by atomic force microscopy (AFM).
• The technique demonstrated robustness across domain sizes from 23 nm to 3 µm.

Conclusions:

• The proposed image texture analysis is a powerful tool for characterizing material surfaces.
• It offers a computationally efficient and comprehensive approach to texture analysis.
• The method shows promise for predicting surface properties like roughness from SEM data.