A dynamic image analysis method was applied for particle characterisation to study the effect of different process conditions during twin-screw compounding of WPC. The use of distributions based on different types of quantity is discussed with respect to their sensitivity to reveal the effects of different process conditions on particle degradation. Distributions based on length proved to be most suitable to represent the initially broad length distribution of the particles before processing. Sensitivity was strong enough to show differences in particle size after processing depending on process conditions. Particle size was reduced by more than 97% compared to initial size. Degradation was stronger with increasing wood content and when the screw design contained more mixing elements. The effect of screw speed and feed rate was dependent on filler content and screw design.