Abstract

Although optical remote sensing can capture the Earth's environment with visible and infrared sensors, it is limited by weather conditions. Often, only a few sets of cloud-free optical imagery are available in cloudy regions, where many agricultural towns are located. On the other hand, radar remote sensing can capture imagery under cloudy conditions. In this study, we examined the capability of Sentinel-1 multitemporal dual-polarized synthetic aperture radar (SAR) imagery in a whole year from Google Earth Engine in crop mapping in two study sites in Chongqing, China, and Landivisiau, France. Results show that it is possible to produce better crop classification maps using multitemporal SAR imagery, but the performance is limited by local terrain. Flat agricultural regions, such as Western Europe, are expected to benefit from the multitemporal SAR information. Mountain agricultural regions, such as Southwestern China, will encounter difficulties due to the undulate terrain. We also tested two sampling strategies, i.e., random sampling and regional sampling, and observed high variation in overall accuracy: the former led to a higher accuracy. The gap is caused by the diversity of training sets examined using tSNE visualization. The importance of SAR channels in each month are correlated with their entropy. Data from the growing season are important in distinguishing crop types. The 3-D convolutional neural network (CNN) achieved similar results under a huge computation cost compared with 2-D CNNs. Based on the experiments, we recommend to use a lightweight 2-D CNN that can run on the CPU for real-world crop mapping with SAR data.

Highlights

  • W ITH more than one billion people at the brink of starvation, the United Nations has announced zero hunger as one of the sustainable development goals [1]

  • The wide contextual residual network (WCRN) achieved the best result, but the OA dropped dramatically compared to random sampling, from 93.07% to 59.57%

  • We aim at using multitemporal synthetic aperture radar (SAR) data to produce crop type classification maps

Read more

Summary

Introduction

W ITH more than one billion people at the brink of starvation, the United Nations has announced zero hunger as one of the sustainable development goals [1]. Thanks to the development of remote sensing technologies, we can estimate food production from satellite imagery with minimal laborious works. Optical remote sensing relies on the visible and infrared information to distinguish crop types. This technique has shown its great potential in crop mapping. Wardlow and Egbert [2] used time-series MODIS NDVI data to create crop type maps with a hierarchical classification approach. Wang et al [8] used Fourier transform of Landsat time-series images to distinguish crop types and achieved over 80% overall accuracy without in-season field data in some regions using a random forest transfer technique. With the literature going deep, crop type classification is generated from low-resolution satellite images to from high-resolution satellite and even UAV images, but most of the studies focus on the usage of random forest because of its robustness [10]

Methods
Results
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call