Texture digital analysis for corrosion monitoring

• Use of texture analysis is feasible for non-destructive surface corrosion monitoring.
• Changes in image characteristics show how corrosion evolves over time.
• We use entropy, Hurst coefficient, contrast, correlation, energy and homogeneity.
• Curves present behaviour that can be compared to the corrosion process evolution.
• The technique is feasible as a new method to check the surface corrosion state.

This article proposes the use of texture analysis, an automated optical inspection technique, for non-destructive surface corrosion monitoring. It presents an ASTM A36 steel atmospheric corrosion test conducted over 44 days to obtain a photo sequence of corrosion evolution. These images have been processed for six textural characteristics, and the values are presented as “feature variation versus time” graphs. The results show that the technique is feasible as a new method to check the surface corrosion state.