Deriving ocean color products using neural networks

Ioannou, Ioannis; Gilerson, Alexander; Gross, Barry; Moshary, Fred; Ahmed, Samir

doi:10.1016/j.rse.2013.02.015

Cited by 75 publications

(40 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…2016, 8,377 4 of 25 (NOMAD) and (c) data from our own field campaign in the Chesapeake Bay during 2013, which represents well the typical range of water optical properties and chlorophyll-a concentrations in coastal regions. These tests confirmed good retrievals of a ph443 by the NN [13,14] and provided strong motivation for its use in the current work. The results of the present work have further confirmed its viability for effective retrievals of a ph443 in the coastal waters of the WFS from VIIRS measurements at the 486, 551 and 671 nm bands.…”

Section: Algorithm Trainingsupporting

confidence: 72%

“…In the work presented, we applied a synthetically trained NN algorithm previously developed and reported by us [12][13][14][15] to solve the inverse problem [19,20] of retrieving physical variables, including a ph443 from, VIIRS observations of Rrs 486, 551 and 671 nm in the WFS. A background summary is given here.…”

Section: A11 Backgroundmentioning

confidence: 99%

“…This model, Figure 1a, with four components: water, phytoplankton, colored dissolved organic matter (CDOM) and non-algal particulates (NAP) as inputs, was constructed to create a synthetic database consisting of 20,000 IOP combinations covering the entire observable range of IOP parameters. These IOPs were then used with a HydroLight based [36], parameterized forward model described in Lee et al, 2002 [23], to generate corresponding sets of Rrs values at 486, 551 and 671 nm for VIIRS, and at 488, 555 and 667 nm for MODIS-A, associated with each of the 20,000 IOPs (and the related different combinations of the four components) [12][13][14][15]. The range and variability of IOPs used in our bio-optical modeling is well represented in the literature [22][23][24][25][26][27][28][29][30][31][32][33][34][35] and covers both case 1 and case 2 waters.…”

Section: A12 Synthetic Datasetmentioning

confidence: 99%

“…NN Training 50% (10,000) of these sets of Rrs and related IOPs were then used [12][13][14][15], to train a VIIRS NN. Figure 1b, to model the relationship between input Rrs values at the 486, 551 and 671 nm bands for VIIRS and the related IOPs a ph , a g , a d , bb p coefficients, all at 443 nm, which is at the peak of phytoplankton absorption spectrum, and thus exhibits most variation in natural waters, (and where a g and a d are mutually constrained through an empirically derived relationship based on a study of NOMAD values) [12][13][14][15]21]. Figure 1b shows the VIIRS NN Architecture used.…”

Section: A12 Synthetic Datasetmentioning

confidence: 99%

See 3 more Smart Citations

Satellite Retrievals of Karenia brevis Harmful Algal Blooms in the West Florida Shelf Using Neural Networks and Comparisons with Other Techniques

et al. 2016

Self Cite

View full text Add to dashboard Cite

Abstract:We describe the application of a Neural Network (NN) previously developed by us, to the detection and tracking, of Karenia brevis Harmful Algal Blooms (KB HABs) that plague the coasts of the West Florida Shelf (WFS) using Visible Infrared Imaging Radiometer Suite (VIIRS) satellite observations. Previous approaches for the detection of KB HABs in the WFS primarily used observations from the Moderate Resolution Imaging Spectroradiometer Aqua (MODIS-A) satellite. They depended on the remote sensing reflectance signal at the 678 nm chlorophyll fluorescence band (Rrs678) needed for both the normalized fluorescence height (nFLH) and Red Band Difference algorithms (RBD) currently used. VIIRS which has replaced MODIS-A, unfortunately does not have a 678 nm fluorescence channel so we customized the NN approach to retrieve phytoplankton absorption at 443 nm (a ph443 ) using only Rrs measurements from existing VIIRS channels at 486, 551 and 671 nm. The a ph443 values in these retrieved VIIRS images, can in turn be correlated to chlorophyll-a concentrations [Chla] and KB cell counts. To retrieve KB values, the VIIRS NN retrieved a ph443 images are filtered by applying limiting constraints, defined by (i) low backscatter at Rrs 551 nm and (ii) a minimum a ph443 value known to be associated with KB HABs in the WFS. The resulting filtered residual images, are then used to delineate and quantify the existing KB HABs. Comparisons with KB HABs satellite retrievals obtained using other techniques, including nFLH, as well as with in situ measurements reported over a four year period, confirm the viability of the NN technique, when combined with the filtering constraints devised, for effective detection of KB HABs.

show abstract

Section: Algorithm Trainingsupporting

confidence: 72%

Section: A11 Backgroundmentioning

confidence: 99%

Section: A12 Synthetic Datasetmentioning

confidence: 99%

Section: A12 Synthetic Datasetmentioning

confidence: 99%

See 2 more Smart Citations

Satellite Retrievals of Karenia brevis Harmful Algal Blooms in the West Florida Shelf Using Neural Networks and Comparisons with Other Techniques

et al. 2016

Self Cite

View full text Add to dashboard Cite

show abstract

“…In general, very diverse data such as classification of biological objects, chemical kinetic data, or even clinical parameters can be handled in essentially the same way. Selected examples of ANNs applications in remote sensing are: water quality monitoring [14], estimation of evapotranspiration [15], derivation of ocean color products [16], mapping fractional snow cover [17], prediction of soil organic matter [18], spatial assessment of air temperature [19], mapping contrasting tillage practices [20], classification of soil texture [21] prediction of productive fossil localities [22], sub-pixel mapping and sub-pixel sharpening [23], etc. Often, ANNs are used even when some basic conditions for their use are not fulfilled.…”

Section: Introductionmentioning

confidence: 99%

Remotely Sensed Soil Data Analysis Using Artificial Neural Networks: A Case Study of El-Fayoum Depression, Egypt

Amato

Havel

Gad

et al. 2015

IJGI

View full text Add to dashboard Cite

Earth observation and monitoring of soil quality, long term changes of soil characteristics and deterioration processes such as degradation or desertification are among the most important objectives of remote sensing. The georeferenciation of such information contributes to the development and progress of the Digital Earth project in the framework of the information globalization process. Earth observation and soil quality monitoring via remote sensing are mostly based on the use of satellite spectral data. Advanced techniques are available to predict the soil or land use/cover categories from satellite imagery data. Artificial Neural Networks (ANNs) are among the most widely used tools for modeling and prediction purposes in various fields of science. The assessment of satellite image quality and suitability for analysing the soil conditions (e.g., soil classification, land use/cover estimation, etc.) is fundamental. In this paper, methodology for data screening and subsequent application of ANNs in remote sensing is presented. The first stage is achieved

show abstract

A neural network‐based four‐band model for estimating the total absorption coefficients from the global oceanic and coastal waters

Cui

Quan²

2015

JGR Oceans

View full text Add to dashboard Cite

In this study, a neural network-based four-band model (NNFM) for the global oceanic and coastal waters has been developed in order to retrieve the total absorption coefficients a(k). The applicability of the quasi-analytical algorithm (QAA) and NNFM models is evaluated by five independent data sets. Based on the comparison of a(k) predicted by these two models with the field measurements taken from the global oceanic and coastal waters, it was found that both the QAA and NNFM models had good performances in deriving a(k), but that the NNFM model works better than the QAA model. The results of the QAA model-derived a(k), especially in highly turbid waters with strong backscattering properties of optical activity, was found to be lower than the field measurements. The QAA and NNFM models-derived a(k) could be obtained from the MODIS data after atmospheric corrections. When compared with the field measurements, the NNFM model decreased by a 0.86-24.15% uncertainty (root-mean-square relative error) of the estimation from the QAA model in deriving a(k) from the Bohai, Yellow, and East China seas. Finally, the NNFM model was applied to map the global climatological seasonal mean a(443) for the time range of July 2002 to May 2014. As expected, the a(443) value around the coastal regions was always larger than the open ocean around the equator. Viewed on a global scale, the oceans at a high latitude exhibited higher a(443) values than those at a low latitude.

show abstract

Deriving ocean color products using neural networks

Cited by 75 publications

References 47 publications

Satellite Retrievals of Karenia brevis Harmful Algal Blooms in the West Florida Shelf Using Neural Networks and Comparisons with Other Techniques

Satellite Retrievals of Karenia brevis Harmful Algal Blooms in the West Florida Shelf Using Neural Networks and Comparisons with Other Techniques

Remotely Sensed Soil Data Analysis Using Artificial Neural Networks: A Case Study of El-Fayoum Depression, Egypt

A neural network‐based four‐band model for estimating the total absorption coefficients from the global oceanic and coastal waters

Contact Info

Product

Resources

About