Fourier transforms for detecting multitemporal landscape fragmentation by remote sensing

Remote sensing is a useful tool for detecting landscape fragmentation, typically by creating land-use maps from remotely sensed images acquired at different dates. Nonetheless, classification may present a number of drawbacks since it degrades the information content of images leading to the loss of continuous information about fragmentation processes. For exploratory purposes, methods to detect landscape change based on continuous information should not require any a-priori assumptions about landscape characteristics. Accordingly, Fourier transforms may represent the best algorithmic solution. In this paper, we describe a Fourier transform tool developed in a free and open-source environment to detect potential fragmentation over the landscape. We briefly introduce Fourier transforms applied to remotely sensed imagery by further showing their potential application with an empirical example. We argue that Fourier transforms represent a straightforward approach for detecting spatial fragmentation of the landscape, on the strength of their potential to detect trends in increase or decrease of complexity/heterogeneity of the landscape in an objective manner. To our knowledge, this is the first open-source tool for analysing fragmentation of the landscape in multitemporal series based on Fourier transforms, which guarantees a high robustness and reproducibility of the applied algorithms.