Many deep-learning-based, domain adaptation methods for remote sensing applications rely on adversarial training strategies to align features extracted from images of different domains in a shared latent space. However, the performance of such representation matching techniques is negatively impacted when class occurrences in the target domain, for which no labelled data is available during training, are highly imbalanced. In this work, we propose a deep-learning-based representation matching approach for domain adaptation in the context of change detection tasks. We further evaluate the approach in a deforestation mapping application, characterized by a high-class imbalance between the deforestation and no-deforestation classes. The domains represent different sites in the Amazon and Brazilian Cerrado biomes. To mitigate the class imbalance problem, we devised an unsupervised pseudo-labeling scheme based on Change Vector Analysis that prevents the feature alignment to be biased towards the over-represented class. The experimental results indicate that the proposed approach can improve the accuracy of cross-domain deforestation detection.