The rapid expansion of tourism, transportation, energy, and agricultural infrastructure in mountain areas raises concerns about landscape fragmentation and impacts on aesthetic values. Effective delineation of these areas relies on negotiating various qualities that define them. In our study, we developed a collaborative consensus-building process with experts to map open spaces. Rather than collecting information on the factors characterizing open spaces, we first obtained a consensus on their delineation using a Delphi survey, followed by machine learning to extract variables explaining the spatial extent of the open spaces. Results show that the Delphi survey allowed experts to get a collective understanding on the delineation of open spaces through a process of knowledge (de)construction. By applying machine learning on the consolidated outcomes, we were then able to predict open spaces not only defined by physical aspects, but also characterized by subjective elements related to experts' perceptions of the landscape. Such an approach cannot only serve as a decision-support tool for more sustainable management of mountain areas, but as a tool to produce legitimized maps integrating knowledge and perception of various stakeholders. By incorporating these diverse perspectives, this participative process also fosters understanding and acceptance for future spatial planning decisions.