In the domain of planetary science, novelty detection is gaining attention because of the operational opportunities it offers, including annotated data products and downlink prioritization. Using a variational autoencoder (VAE), this work improves upon state-of-the-art novelty detection performance in the context of Martian exploration by $>7\%$ (measured by the area under the receiver operating characteristic curve (ROC AUC)). Autoencoders, especially VAEs, perform well across all classes of novelties defined for Martian exploration. VAEs are shown to have high recall in the Martian context, making them particularly useful for on-ground processing. Convolutional autoencoders (CAEs), on the other hand, demonstrate high precision making them good candidates for onboard downlink prioritization. In our implementation adversarial autoencoders (AAEs) are also shown to perform on par with state-of-the-art. Dimensionality reduction is a key feature of autoencoders for novelty detection. In this study the impact of dimensionality reduction on detection quality is explored, showing that both VAEs and AAEs achieve comparable ROC AUCs to CAEs despite observably poorer (blurred) image reconstructions; this is observed both in Martian data and in lunar analogue data.