Corrosion pits in aluminium alloys nucleate and grow in the proximity of constituent particles. Descriptions of the spatial statistics of the particles and pits are critical for stochastic modelling of corrosion. Statistical models, which are generalizations of a hard-core sequential inhibition process, are proposed. The models include the effects of randomness in the number, the locations, and the sizes of the features. The applicability of the models are based on experimental data taken from 2024-T3 aluminium alloy specimens subjected to a 0.5 M NaCl environment. The observed constituent particles tend to be clustered, whereas the subsequent corrosion pits tend towards regularity as measured by nearest neighbour spacing.