Improving the accuracy of land cover classification in cloud persistent areas using optical and radar satellite image time series

Abstract

The recent availability of high spatial and temporal resolution optical and radar satellite imagery has dramatically increased opportunities for mapping land cover at fine scales. Fusion of optical and radar images has been found useful in tropical areas affected by cloud cover because of their complementarity. However, the multitemporal dimension these data now offer is often neglected because these areas are primarily characterized by relatively low levels of seasonality and because the consideration of multitemporal data requires more processing time. Hence, land cover mapping in these regions is often based on imagery acquired for a single date or on an average of multiple dates. The aim of this work is to assess the added value brought by the temporal dimension of optical and radar time series when mapping land cover in tropical environments. Specifically, we compared the accuracies of classifications based on (a) optical time series, (b) their temporal average, (c) radar time series, (d) their temporal average, (e) a combination of optical and radar time series and (f) a combination of their temporal averages for mapping land cover in Jambi province, Indonesia, using Sentinel‐1 and Sentinel‐2 imagery. Using the full information contained in the time series resulted in significantly higher classification accuracies than using temporal averages (+14.7% for Sentinel‐1, +2.5% for Sentinel‐2 and +2% combining Sentinel‐1 and Sentinel‐2). Overall, combining Sentinel‐2 and Sentinel‐1 time series provided the highest accuracies (Kappa = 88.5%). Our study demonstrates that preserving the temporal information provided by satellite image time series can significantly improve land cover classifications in tropical biodiversity hotspots, improving our capacity to monitor ecosystems of high conservation relevance such as peatlands. The proposed method is reproducible, automated and based on open‐source tools satellite imagery.

Large-Scale High-Resolution Coastal Mangrove Forests Mapping Across West Africa With Machine Learning Ensemble and Satellite Big Data

“…Radar imagery is sensitive to target structure and can be useful when combined with optical imagery for discriminating vegetation types which may be spectrally similar to mangrove forests. Therefore, fusion of the complementary information on vegetation characteristics from the optical Sentinel-2 and radar Sentinel-1 data provided the capability to improve mangrove mapping accuracies (Hamdan et al, 2014;Aslan et al, 2016;Lopes et al, 2020). The use of higher-resolution imagery was, in part, responsible for the improved mangrove maps (Figure 4).…”

No evidence for trade‐offs between bird diversity, yield and water table depth on oil palm smallholdings: Implications for tropical peatland landscape restoration

“…Bird species presence on farms can be affected by the distance to areas of forest remaining in the wider cultivated landscape (Azhar et al, 2011), which act as population sources (Hamer et al, 2021). We therefore used land cover data and aerial imagery (Figure 1A) from existing analysis (Crowson et al, 2018; Lopes et al, 2020) to calculate the distance from plot centroids to the edge of the contiguous forest cover within the forest fragment, to use as a co‐variate in analyses of bird diversity in oil palm. Understorey vegetation of oil palm farms (27 plant species recorded across all plots; Table S1) was dominated by the fern Nephrolepis falcate and the herb Asystasia gangetica .…”

Recent advances in the remote sensing of insects

“…(2016), for example, used a combination of LiDAR‐derived canopy heights and hyperspectral reflectance data to characterise landscape composition and its influence on the foraging distances of bumblebees. The launch of the Copernicus programme satellites, in particular, has greatly facilitated the availability of co‐registered radar (Sentinel‐1) and optical (Sentinel‐2) satellite imagery, enabling improved classifications of habitat, for instance in regions where persistent cloud cover is an issue (Lopes et al ., 2020).…”