Lem Benchmark Database for Tropical Agricultural Remote Sensing Application

Abstract

<p><strong>Abstract.</strong> The monitoring of agricultural activities at a regular basis is crucial to assure that the food production meets the world population demands, which is increasing yearly. Such information can be derived from remote sensing data. In spite of topic’s relevance, not enough efforts have been invested to exploit modern pattern recognition and machine learning methods for agricultural land-cover mapping from multi-temporal, multi-sensor earth observation data. Furthermore, only a small proportion of the works published on this topic relates to tropical/subtropical regions, where crop dynamics is more complicated and difficult to model than in temperate regions. A major hindrance has been the lack of accurate public databases for the comparison of different classification methods. In this context, the aim of the present paper is to share a multi-temporal and multi-sensor benchmark database that can be used by the remote sensing community for agricultural land-cover mapping. Information about crops in situ was collected in Luís Eduardo Magalhães (LEM) municipality, which is an important Brazilian agricultural area, to create field reference data including information about first and second crop harvests. Moreover, a series of remote sensing images was acquired and pre-processed, from both active and passive orbital sensors (Sentinel-1, Sentinel-2/MSI, Landsat-8/OLI), correspondent to the LEM area, along the development of the main annual crops. In this paper, we describe the LEM database (crop field boundaries, land use reference data and pre-processed images) and present the results of an experiment conducted using the Sentinel-1 and Sentinel-2 data.</p>

A Many-to-Many Fully Convolutional Recurrent Network for Multitemporal Crop Recognition

“…The second region is located in Luis Eduardo Magalhães (LEM) municipality, also in Brazil, with an area of 3,940 km 2 (Sanches et al, 2018a). A set of 13 pre-processed Sentinel-1 SAR images acquired between June 2017 and June 2018 was used in our experiments.…”

Understanding the Dynamic of Tropical Agriculture for Remote Sensing Applications: A Case Study of Southeastern Brazil

“…The second issue is more difficult to address, since field information is time-consuming and expensive to be acquired and demand specialists. However, researchers are becoming more willing to share their own data, and the number of free available databases containing field information are increasing, for example, the databases of Sanches et al (2018a) and Sanches et al (2018b).…”

Crop Recognition in Tropical Regions Based on Spatio-Temporal Conditional Random Fields From Multi-Temporal and Multi-Resolution Sequences of Remote Sensing Images

“…Field works to record in-situ data from each region were performed by experts from INPE and EMBRAPA, while the acquisition and pre-processing of radar images have been carried out during the development of this thesis. A detailed description and more information about each dataset can be found in [11] and in [12].…”

“…They were in charge of the field works to collect data from the study areas to assign a label to each plot, while the acquisition and pre-processing of Sentinel-1A scenes was performed at PUC-Rio. These datasets received financial support from the Brazilian agencies CAPES 1 and CNPq 2 for Campo Verde [11], and from the ISPRS 3 , for Luis Eduardo Magalhães [12].…”

“…Level-2 Landsat 8 data products were acquired from the USGS Earth Resources Observation and Science Center in surface reflectance, mosaicked and clipped. Further details about this dataset can be found in [12]. The reference data was built by two field campaigns conducted between June 26th-30th, 2017 (second harvest, dry season) and March 14th-19th, 2018 (first harvest, wet season) to record the geographic coordinates, land-use classes of each field and its phenology phase.…”