Quantitative Microstructural Analysis of State-of-the-art Lithium-ion Battery Cathodes

In this application note two different segmentation methods – classic thresholding and machine learning-based – are evaluated in the context of quantitative analysis of constituents of state-of-the-art lithium-ion batteries.

Both methods are compared against reference measurements using a laboratory balance. A detailed description of the fundamental functionality of the machine learning based approach is given. It has the potential to be more robust against image variations and thus can provide a more accurate segmentation result.