Cloud-Free Satellite Image Mosaics with Regression Trees and Histogram Matching

Abstract

Cloud-free optical satellite imagery simplifies remote sensing, but land-cover phenology limits existing solutions to persistent cloudiness to compositing temporally resolute, spatially coarser imagery. Here, a new strategy for developing cloud-free imagery at finer resolution permits simple automatic change detection. The strategy uses regression trees to predict pixel values underneath clouds and cloud shadows in reference scenes from other scene dates. It then applies improved histogram matching to adjacent scenes. In the study area, the islands of Puerto Rico, Vieques, and Culebra, Landsat image mosaics resulting from this strategy permit accurate detection of land development with only spectral data and maximum likelihood classification. Between about 1991 and 2000, urban/built-up lands increased by 7.2 percent in Puerto Rico and 49 percent in Vieques and Culebra. The regression tree modeling and histogram matching require no manual interpretation. Consequently, they can support large volume processing to distribute cloud-free imagery for simple change detections with common classifiers.

The impact of forest regeneration on streamflow in 12 mesoscale humid tropical catchments

“…Although the maps for 1951 and 1978 were rasterized at a resolution of ∼ 30 and ∼ 11 m, respectively, the actual mapping resolution used by the photo interpreters is estimated at ∼ 300 and ∼ 50 m, respectively. The 1991 and 2000 maps 1 (∼ 30 m resolution; Helmer and Ruefenacht, 2005) were derived from Landsat data using regression-tree modelling and histogram match-ing. All land-cover maps were reprojected to a common 0.0001 • (∼ 11 m) geographical grid by nearest-neighbour interpolation.…”

Forest recovery in a tropical landscape: what is the relative importance of biophysical, socioeconomic, and landscape variables?

“…The images were originally classified into eleven land cover classes, which we reduced to nine for our analyses (see below). Details of classification accuracy are provided in Helmer and Ruefenacht (2005). In brief, their method correctly classifying 85.4% of points and yielded an error matrix with a Kappa coefficient of agreement of 0.66 ± 0.12.…”

Continued Reforestation and Urban Expansion in the New Century of a Tropical Island in the Caribbean

“…To collect the ground reference data for the year of 2000, we first chose the land cover polygons greater than 2.25 hectares (5 × 5 pixels) based on the historical land cover map in 2000 (see Section 2.4) [29,40] and created a 30-m negative buffer to exclude the edges. We then randomly selected 2,000 points from the centers of these polygons and manually checked the land cover of each point based on the highresolution imagery from Google Earth.…”

“…For comparison, a forest change map for the period of 1991-2000 was also created according to the historical land cover map of 1991 and 2000 derived from the Landsat TM and ETM+ using decision-tree classifier [29,40]. To quantify the rates of forest change along the gradients of forest coverage, we divided the main island of Puerto Rico into non-overlapping grid cells according to three window sizes, i.e., 25 × 25, 50 × 50, and 100 × 100 pixels, and cross-tabulated the forest/non-forest maps of the years of 2000 and 2010 at each scale.…”