Density Estimates as Representations of Agricultural Fields for Remote Sensing-Based Monitoring of Tillage and Vegetation Cover

Abstract

We consider the use of remote sensing for large-scale monitoring of agricultural land use, focusing on classification of tillage and vegetation cover for individual field parcels across large spatial areas. From the perspective of remote sensing and modelling, field parcels are challenging as objects of interest due to highly varying shape and size but relatively uniform pixel content and texture. To model such areas we need representations that can be reliably estimated already for small parcels and that are invariant to the size of the parcel. We propose representing the parcels using density estimates of remote imaging pixels and provide a computational pipeline that combines the representation with arbitrary supervised learning algorithms, while allowing easy integration of multiple imaging sources. We demonstrate the method in the task of the automatic monitoring of autumn tillage method and vegetation cover of Finnish crop fields, based on the integrated analysis of intensity of Synthetic Aperture Radar (SAR) polarity bands of the Sentinel-1 satellite and spectral indices calculated from Sentinel-2 multispectral image data. We use a collection of 127,757 field parcels monitored in April 2018 and annotated to six tillage method and vegetation cover classes, reaching 70% classification accuracy for test parcels when using both SAR and multispectral data. Besides this task, the method could also directly be applied for other agricultural monitoring tasks, such as crop yield prediction.

Monitoring autumn agriculture activities using Synthetic Aperture Radar (SAR) and coherence change detection

“…The revisit period of satellites varies, but these platforms typically provide a re-look at agriculture fields with periods between image acquisitions ranging from days to weeks. Satellites are able to repeatedly capture changes in field conditions throughout the busy harvesting and seeding periods [ [9] , [10] , [11] , [12] ].…”

“…Most researchers who have studied the application of radar to tillage mapping have focused on the retrieval of rms from backscatter intensity [ 9 ], detecting changes in backscatter due to tillage [ 13 ], or integrating SAR images in classifiers [ 11 , 12 ]. [ 9 ] tilled research plots with different implements and then measured the backscatter (in HH (horizontal send, horizontal receive), VV (vertical send, vertical receive), and VH polarizations, (vertical send, horizontal receive)) with a C-band scatterometer.…”

“…[ 12 ] used a combination of Sentinel-1 and Sentinel-2 data to classify six agriculture classes including conventional plowing (mostly moldboard tillage), conservation tillage, autumn crop, grass, stubble, and stubble with a companion crop. Sentinel-1 VV and VH intensity, and two optical indices from Sentinel-2, were input into several classification techniques including Support Vector Machine (SVM), Multi-Layer Perceptron (MLP), and Random Forest (RF)).…”

Scalable Crop Yield Prediction with Sentinel-2 Time Series and Temporal Convolutional Network

“…However, a point estimate of distribution, such as mean or median, is inclined to loose important discriminating information about the samples. Histograms provide a straightforward means for utilizing more information from the reflectance distribution [15][16][17], and histograms with more than one dimension can be used if the prediction method can handle a larger feature space.…”