A GEOBIA Methodology for Fragmented Agricultural Landscapes

Angel Garcia-Pedrero,Consuelo Gonzalo-Martin,Mario Lillo-Saavedra,David Fonseca-Luengo

doi:10.3390/rs70100767

Angel Garcia-Pedrero, Consuelo Gonzalo-Martin + Show 2 more

Open Access

https://doi.org/10.3390/rs70100767

Copy DOI

Abstract

Very high resolution remotely sensed images are an important tool for monitoring fragmented agricultural landscapes, which allows farmers and policy makers to make better decisions regarding management practices. An object-based methodology is proposed for automatic generation of thematic maps of the available classes in the scene, which combines edge-based and superpixel processing for small agricultural parcels. The methodology employs superpixels instead of pixels as minimal processing units, and provides a link between them and meaningful objects (obtained by the edge-based method) in order to facilitate the analysis of parcels. Performance analysis on a scene dominated by agricultural small parcels indicates that the combination of both superpixel and edge-based methods achieves a classification accuracy slightly better than when those methods are performed separately and comparable to the accuracy of traditional object-based analysis, with automatic approach.

Highlights

Over the last three decades, remote sensing has played an important role in agricultural monitoring and land-cover/land-use classification [1,2]
We propose an automatic GEOBIA methodology for fragmented agricultural landscapes focusing on the generation of a thematic map of the available classes in the scene
Edge-based segments are well suited to most of the small agricultural parcels. From these segments, using a compactness factor of 80, a total of 297,257 superpixels covering the whole image were automatically generated, where each superpixel is composed of 25 pixels on average

Summary

Introduction

Over the last three decades, remote sensing has played an important role in agricultural monitoring and land-cover/land-use classification [1,2]. At a global and regional scale, mapping extensive crops is carried out effectively through the analysis of remote-sensing images with a spatial resolution of 15 m or greater (e.g., images coming from sensors such as ETM+ or ASTER) Images of this type are not suitable for modeling fragmented agricultural landscapes (with parcels of 5 ha or less), because a precise delimitation of the parcels is required for its analysis; and this resolution does not allow a representation of the spatial behavior of the parcels of intensive cultivation [3]. This higher resolution represents a challenge for available data exploitation and dissemination approaches

Methods

Results

Conclusion