Cropland Classification Using Sentinel-1 Time Series: Methodological Performance and Prediction Uncertainty Assessment

Abstract

Methods based on Sentinel-1 data were developed to monitor crops and fields to facilitate the distribution of subsidies. The objectives were to (1) develop a methodology to predict individual crop species or or management regimes; (2) investigate the earliest time point in the growing season when the species predictions are satisfactory; and (3) to present a method to assess the uncertainty of the predictions at an individual field level. Seventeen Sentinel-1 synthetic aperture radar (SAR) scenes (VV and VH polarizations) acquired in interferometric wide swath mode from 14 May through to 30 August 2017 in the same geometry, and selected based on the weather conditions, were used in the study. The improved k nearest neighbour estimation, ik-NN, with a genetic algorithm feature optimization was tailored for classification with optional Sentinel-1 data sets, species groupings, and thresholds for the minimum parcel area. The number of species groups varied from 7 to as large as 41. Multinomial logistic regression was tested as an optional method. The Overall Accuracies (OA) varied depending on the number of species included in the classification, and whether all or not field parcels were included. OA with nine species groups was 72% when all parcels were included, 81% when the parcels area threshold (for incorporating parcels into classification) was 0.5 ha, and around 90% when the threshold was 4 ha. The OA gradually increased when adding extra Sentinel-1 scenes up until the early August, and the initial scenes were acquired in early June or mid-May. After that, only minor improvements in the crop recognition accuracy were noted. The ik-NN method gave greater overall accuracies than the logistic regression analysis with all data combinations tested. The width of the 95% confidence intervals with ik-NN for the estimate of the probability of the species with the largest probability on an individual parcel varied depending on the species, the area threshold of the parcel and the number of the Sentinel-1 scenes used. The results ranged between 0.06-0.08 units (6-8% points) for the most common species when the Sentinel-1 scenes were between 1 June and 12 August. The results were well-received by the authorities and encourage further research to continue the study towards an operational method in which the space-borne SAR data are a part of the information chain.

Time-Series of Sentinel-1 Interferometric Coherence and Backscatter for Crop-Type Mapping

“…The results obtained in this work can be compared with those obtained in previous experiments with time series of SAR data, and more specifically with Sentinel-1 data, e.g. [3], [4], but always taking into account that each study about crop-type mapping has unique characteristics, including the number and specificity of crop species present in the study site, the cultivation practices, the agro-climatic conditions, which usually depend on the geographical location, and the dates and time span of the acquisitions. This complicates comparison in absolute terms and has to be considered to draw right conclusions.…”

“…Conventional SAR-based mapping methods employ radiometric information, that is, time series of backscatter coefficients (intensities) and/or indices derived from them (such as the ratios) are used as input to classification algorithms. In this regard, it is worth mentioning that time series of Sentinel-1 radiometric data have proven effective and valuable in this context, as shown with some previous works concerning croptype mapping in different areas [1], [2], [3], [4].…”

“…This complicates comparison in absolute terms and has to be considered to draw right conclusions. For instance, the time span covered by the satellite images in this work is a whole year, with data from a single orbit, whereas only 4 months were used in [3], and more than one year and three orbits were exploited in [4].…”

“…The selection of the multilook filter employed in this work (boxcar) was done on the basis of the simplest processing, at the expenses of degraded resolution and, hence, potential problems in edges between fields and/or when parcels are very small. In this regard, recent works exploiting Sentinel-1 backscatter data have observed worse performances for small parcels than for large ones [3], [4]. Therefore, the use of more sophisticated filtering techniques, like non-local approaches [24], should be tested for the computation of both coherence and intensity.…”

Exploring TanDEM-X Interferometric Products for Crop-Type Mapping

“…As a result, the absence of vegetation, especially at the beginning of the season and after harvest, is expected to be detected by repeat-pass interferometry over agricultural areas. This detection would provide information related to the crop calendar, which normally depends on the crop type, hence being useful for crop-type mapping [11,12,29,30]. The influence of the wavelength or frequency band on temporal decorrelation is also an important aspect to be considered in this context [31,32].…”