Urbanisation is usually modelled to account for the trade-off between the rent from an agricultural and urban land-use in a location. In this article, we propose a model that includes a characterisation of the land in respect of not only its economic and physical aspects, but also using variables in relation to landscape perception. To that end, we develop an original two-stage approach consisting of estimating a probability of urbanisation and then taking the uncertainty of urbanisation into account using an internal meta-regression method. The landscape descriptors, constructed based on a textual analysis of the Landscape Atlases, are introduced in this second stage. The application of this method to the urban area of Angers shows the importance of these elements in analysing urbanisation.