Assessment of the X- and C-Band Polarimetric SAR Data for Plastic-Mulched Farmland Classification

Abstract

We present a classification of plastic-mulched farmland (PMF) and other land cover types using full polarimetric RADARSAT-2 data and dual polarimetric (HH, VV) TerraSAR-X data, acquired from a test site in Hebei, China, where the main land covers include PMF, bare soil, winter wheat, urban areas and water. The main objectives were to evaluate the outcome of using high-resolution TerraSAR-X data for classifying PMF and other land covers and to compare classification accuracies based on different synthetic aperture radar bands and polarization parameters. Initially, different polarimetric indices were calculated, while polarimetric decomposition methods were used to obtain the polarimetric decomposition components. Using these polarimetric components as input, the random forest supervised classification algorithm was applied in the classification experiments. Our results show that in this study full-polarimetric RADARSAT-2 data produced the most accurate overall classification (94.81%), indicating that full polarization is vital to distinguishing PMF from other land cover types. Dual polarimetric data had similar levels of classification error for PMF and bare soil, yielding mapping accuracies of 53.28% and 59.48% (TerraSAR-X), and 59.56% and 57.1% (RADARSAT-2), respectively. We found that Shannon entropy made the greatest contribution to accuracy in all three experiments, suggesting that it has great potential to improve agricultural land use classifications based on remote sensing.

Dryland Crop Classification Combining Multitype Features and Multitemporal Quad-Polarimetric RADARSAT-2 Imagery in Hebei Plain, China

“…Based on assessment of the relative importance of all the features, it was demonstrated that the Shannon entropy extracted using the C–P polarization decomposition method made an important contribution to the accuracy of dryland crop mapping. This finding is also in accord with the results of previous research [ 9 ]. The backscattering coefficient in different polarization modes and the polarization decomposition parameters extracted using the C–P, F–D, and Yamaguchi decomposition methods have been used widely for land-cover and land-use classification.…”

“…In addition to these features, we also used other backscattering intensity features, e.g., the main diagonal elements in the Coherency Matrix, and the polarization parameters extracted using the MCSM decomposition approach to enhance the accuracy of dryland crop classification. This novel combination of various features is markedly different from that of previous studies [ 9 , 20 , 29 , 35 ]. For optimization of the process, we used the RF algorithm to identify the most important features and the critical dates of imagery acquisition, the use of which was shown to both improve significantly the efficiency of crop classification and reduce the workload and costs associated with crop mapping.…”

“…Affected by the monsoon climate of medium latitudes, 80% of average annual precipitation falls in this region during July–September, which just covers the main growing periods of maize and cotton [ 8 ]. Unfortunately, the availability and quality of optical remotely sensed imagery covering this region are often very poor, owing to the cloudy/rainy weather that occur frequently during the critical phenological stages of this two typical dryland crops in Hebei plain [ 9 ]. Consequently, optical remote-sensing techniques are inadequate for accurate and timely monitoring of these crops.…”

Remote Sensing of Agricultural Greenhouses and Plastic-Mulched Farmland: An Analysis of Worldwide Research

“…With few exceptions focused on PMF land cover extraction [31,42,66], SAR data has been practically not used as a data source for PCG mapping. However, with the very recent arrival of very high spatio-temporal resolution SAR data provided by the new generation of small satellites (e.g., ICEYE SAR (https://www.iceye.com/)), with X-band radar imaging available even at a 1 m ground sample distance resolution [67], it is likely that there will be a need to test its promising potential in RSAGPM.…”

Remote sensing of plastic‐covered greenhouses and plastic‐mulched farmlands: Current trends and future perspectives

“…In addition, LANDSAT‐8 and SENTINEL‐2 multispectral data have been used in a number of studies published recently (Jimenez‐Lao et al, 2020). Synthetic aperture radar (SAR) data was also used in a number of studies for the classification of plastic‐mulched farmland areas (e.g., Hasituya & Chen, 2017; Liu et al, 2019). A summary of latest studies using remote sensing data for mapping and monitoring PCGs plastic‐mulched horticulture areas is given in Appendix 1.…”