Abstract
The CSSL does not hold image data, but it does hold statistical characterizations derived from analyzing both multispectral and panchromatic images through a fully automated workflow. Consequently, we obtained a decent number of vegetation indices and their in-field variability, a number of other spectral characteristics, as well as a number of GLCM-based textural characteristics (different lags, different angles). We continue to enrich that list with other image-based analytics.
Thus, on the imaging side, our analysis produced various tens of characteristics of farm fields that are either spectral or textural in nature, while fieldwork produced a number of in situ agronomic measurements, characterizing crop growth and field maintenance. All such data was semi-synchronously collected throughout the crop season at regular two-week intervals. Our philosophy is that a collection of this nature can support studies in crop identification, farm field delineation, farm practice detection and other crop-related phenomena in smallholder contexts.
We thus also present an online exploration tool that allows inspection of characteristics and their correspondences, and invite the larger scientific community to start using this resource, which accommodate time series comparisons, for instance, between different vegetation indices and textural or in situ measurements. We invite the scientific audience to use the tool, and those conducting image-based projects on smallholder farming, to contribute to its baseline through collaboration with us to enrich it with more crops, more years, and a wider geographic coverage.
| Original language | English |
|---|---|
| Pages | s1-s22 |
| Publication status | Published - 27 Sep 2017 |
| Event | Earth Observation Open Science 2017 Conference - Frascati, Italy Duration: 25 Sep 2017 → 28 Sep 2017 https://livestream.com/ESA/OpenScience2017 |
Conference
| Conference | Earth Observation Open Science 2017 Conference |
|---|---|
| Country | Italy |
| City | Frascati |
| Period | 25/09/17 → 28/09/17 |
| Internet address |
Fingerprint
Cite this
, Frascati, Italy.
}
, Frascati, Italy, 25/09/17 - 28/09/17, pp. s1-s22.
A Tool to Explore Spectral, Spatial and Temporal Features of Smallholder Crops : powerpoint. / de By, R.A.; Zurita-Milla, R.; Pasha Zadeh M., P.; Calisto, L.F.
2017. s1-s22 Earth Observation Open Science 2017 Conference, Frascati, Italy.
Research output: Contribution to conference › Other › Other research output
TY - CONF
T1 - A Tool to Explore Spectral, Spatial and Temporal Features of Smallholder Crops : powerpoint
AU - de By, R.A.
AU - Zurita-Milla, R.
AU - Pasha Zadeh M., P.
AU - Calisto, L.F.
PY - 2017/9/27
Y1 - 2017/9/27
N2 - We present a crop characteristics database plus web-based open data exploration tool as one of the results produced by the STARS project (www.stars-project.org). STARS aims to address the information scarcity around smallholder farming in Africa and Asia through the use of high-resolution satellite images. We conducted a number of studies in sites in W and E Africa as well as S Asia, which brought together fieldwork-derived and image-derived characteristics of farm fields into a central database, which we call the Crop Spectrotemporal Signature Library (CSSL). We present its structure and contents.The CSSL does not hold image data, but it does hold statistical characterizations derived from analyzing both multispectral and panchromatic images through a fully automated workflow. Consequently, we obtained a decent number of vegetation indices and their in-field variability, a number of other spectral characteristics, as well as a number of GLCM-based textural characteristics (different lags, different angles). We continue to enrich that list with other image-based analytics.Thus, on the imaging side, our analysis produced various tens of characteristics of farm fields that are either spectral or textural in nature, while fieldwork produced a number of in situ agronomic measurements, characterizing crop growth and field maintenance. All such data was semi-synchronously collected throughout the crop season at regular two-week intervals. Our philosophy is that a collection of this nature can support studies in crop identification, farm field delineation, farm practice detection and other crop-related phenomena in smallholder contexts.We thus also present an online exploration tool that allows inspection of characteristics and their correspondences, and invite the larger scientific community to start using this resource, which accommodate time series comparisons, for instance, between different vegetation indices and textural or in situ measurements. We invite the scientific audience to use the tool, and those conducting image-based projects on smallholder farming, to contribute to its baseline through collaboration with us to enrich it with more crops, more years, and a wider geographic coverage.
AB - We present a crop characteristics database plus web-based open data exploration tool as one of the results produced by the STARS project (www.stars-project.org). STARS aims to address the information scarcity around smallholder farming in Africa and Asia through the use of high-resolution satellite images. We conducted a number of studies in sites in W and E Africa as well as S Asia, which brought together fieldwork-derived and image-derived characteristics of farm fields into a central database, which we call the Crop Spectrotemporal Signature Library (CSSL). We present its structure and contents.The CSSL does not hold image data, but it does hold statistical characterizations derived from analyzing both multispectral and panchromatic images through a fully automated workflow. Consequently, we obtained a decent number of vegetation indices and their in-field variability, a number of other spectral characteristics, as well as a number of GLCM-based textural characteristics (different lags, different angles). We continue to enrich that list with other image-based analytics.Thus, on the imaging side, our analysis produced various tens of characteristics of farm fields that are either spectral or textural in nature, while fieldwork produced a number of in situ agronomic measurements, characterizing crop growth and field maintenance. All such data was semi-synchronously collected throughout the crop season at regular two-week intervals. Our philosophy is that a collection of this nature can support studies in crop identification, farm field delineation, farm practice detection and other crop-related phenomena in smallholder contexts.We thus also present an online exploration tool that allows inspection of characteristics and their correspondences, and invite the larger scientific community to start using this resource, which accommodate time series comparisons, for instance, between different vegetation indices and textural or in situ measurements. We invite the scientific audience to use the tool, and those conducting image-based projects on smallholder farming, to contribute to its baseline through collaboration with us to enrich it with more crops, more years, and a wider geographic coverage.
M3 - Other
SP - s1-s22
ER -
, Frascati, Italy.