ABSTRACT
This chapter analyzes the correspondence between human conceptualizations of landscapes and spectrally derived land cover classifications. It focuses on how the abundance of freely available high-resolution imagery of the earth’s surface and the maturing of crowd science offers new opportunities for an unprecedented access to environmental information. The chapter examines how crowd science and human perceptions can be used for the purposes of improving overall quality of land cover datasets. It discusses variation between landcover classifications and explores why they exist. The chapter also discusses the issues and debates surrounding the accuracy of land cover datasets. To reduce variability in interpretation of classification, land cover classifications must be concerned with users’ natural concepts and perceptions of the land cover, and be aware of formal cognitive models about the commonsense geographic world.
