Gamification is a widely used resource to engage and retain users. It is about the use of game elements and mechanics in systems and domains that are not naturally games. Nevertheless, the usage of gamification does not always achieve the expected results due to the too much generalized approach that makes invisible the different motivations, characteristics and playing styles among the players. Currently, research on adaptive gamification deals with the gamification that each particular user needs at a particular moment, adapting gamification to users and contexts. Collaborative location-based collecting systems (CLCS) are a particular case of collaborative systems where a community of users collaboratively collect geo-referenced data. This article proposes an adapted gamification approach for CLCS, through the automatic game challenge generation. Particularly a model of user profile considering the spacetime behavior and challenge completion, a model for the different types of challenges applicable in CLCS, a model for the CLCS objectives and coverage, and a strategy for the application of Machine Learning techniques for adaptation.