Classification of Land Use and Land Cover in the Brazilian Amazon using Fuzzy Multilayer Perceptrons

Here the authors propose the use of Fuzzy Multilayer Perceptrons for classification of land use and land cover patterns in the Brazilian Amazon, using time series of vegetation index, taken from NASA's MODIS (Moderate Resolution Imaging Spectroradiometer) sensor. In addition to the traditional Multilayer Perceptron (MLP), three fuzzy implementations were investigated. These methods were applied to a study area of approximately 10.5 km2 on the east of the state of Mato Grosso in the Brazilian Amazon. For validation purposes, the authors compared the best implementation results with the ones given for the same region by the TerraClass 2010 project. The authors observed that our fuzzy MLP correctly classified 81% of the pixels analyzed.

Download Full-text

Framework for Mapping Integrated Crop-Livestock Systems in Mato Grosso, Brazil

Remote Sensing ◽

10.3390/rs10091322 ◽

2018 ◽

Vol 10 (9) ◽

pp. 1322 ◽

Cited By ~ 4

Author(s):

Víctor Manabe ◽

Marcio Melo ◽

Jansle Rocha

Keyword(s):

Land Use ◽

Crop Production ◽

Vegetation Index ◽

Phase 1 ◽

Plant Diseases ◽

Mato Grosso ◽

Moderate Resolution ◽

Moderate Resolution Imaging Spectroradiometer ◽

Dynamic Time

Integrated crop-livestock (ICL) systems combine livestock and crop production in the same area, increasing the efficiency of land use and machinery, while mitigating greenhouse gas emissions, and reducing production risks, plant diseases and pests. ICL systems are primarily divided into annual (ICLa) and multi-annual (ICLm) systems. Projects such as the “Integrated crop-livestock-forest Network” and the “Livestock Rally” have estimated the ICL areas for Brazil on a state or regional basis. However, it remains necessary to create methods for spatial identification of ICL areas. Thus, we developed a framework for mapping ICL areas in Mato Grosso, Brazil using the Enhanced Vegetation Index time-series of Moderate Resolution Imaging Spectroradiometer and a Time-Weighted Dynamic Time Warping (TWDTW) classification method. The classification of ICL areas occurred in three phases. Phase 1 corresponded to the classification of land use from 2008 to 2016. In Phase 2, the ICLa areas were identified. Finally, Phase 3 corresponded to the ICLm identification. The framework showed overall accuracies of 86% and 92% for ICL areas. ICLm accounted for 87% of the ICL areas. Considering only agricultural areas or only pasture areas, ICL systems represented 5% and 15%, respectively.

Download Full-text

A multinomial logistic model-based land use and land cover classification for the South Asian Association for Regional Cooperation nations using Moderate Resolution Imaging Spectroradiometer product

Environment Development and Sustainability ◽

10.1007/s10668-020-00864-1 ◽

2020 ◽

Cited By ~ 1

Author(s):

Ram Kumar Singh ◽

Vinay Shankar Prasad Sinha ◽

Pawan Kumar Joshi ◽

Manoj Kumar

Keyword(s):

Land Use ◽

Land Cover ◽

South Asian ◽

Logistic Model ◽

Regional Cooperation ◽

Land Cover Classification ◽

Multinomial Logistic Model ◽

Moderate Resolution ◽

Resolution Imaging ◽

Moderate Resolution Imaging Spectroradiometer

Download Full-text

Estimating soybean crop areas using spectral-temporal surfaces derived from MODIS images in Mato Grosso, Brazil

Pesquisa Agropecuária Brasileira ◽

10.1590/s0100-204x2010000100010 ◽

2010 ◽

Vol 45 (1) ◽

pp. 72-80 ◽

Cited By ~ 25

Author(s):

Rui Dalla Valle Epiphanio ◽

Antonio Roberto Formaggio ◽

Bernardo Friedrich Theodor Rudorff ◽

Eduardo Eiji Maeda ◽

Alfredo José Barreto Luiz

Keyword(s):

Bare Soil ◽

Temporal Response ◽

Mato Grosso ◽

Moderate Resolution ◽

Resolution Imaging ◽

Moderate Resolution Imaging Spectroradiometer ◽

Using Data ◽

Agricultural Regions ◽

Large Farms

The objective of this work was to evaluate the application of the spectral-temporal response surface (STRS) classification method on Moderate Resolution Imaging Spectroradiometer (MODIS, 250 m) sensor images in order to estimate soybean areas in Mato Grosso state, Brazil. The classification was carried out using the maximum likelihood algorithm (MLA) adapted to the STRS method. Thirty segments of 30x30 km were chosen along the main agricultural regions of Mato Grosso state, using data from the summer season of 2005/2006 (from October to March), and were mapped based on fieldwork data, TM/Landsat-5 and CCD/CBERS-2 images. Five thematic classes were considered: Soybean, Forest, Cerrado, Pasture and Bare Soil. The classification by the STRS method was done over an area intersected with a subset of 30x30-km segments. In regions with soybean predominance, STRS classification overestimated in 21.31% of the reference values. In regions where soybean fields were less prevalent, the classifier overestimated 132.37% in the acreage of the reference. The overall classification accuracy was 80%. MODIS sensor images and the STRS algorithm showed to be promising for the classification of soybean areas in regions with the predominance of large farms. However, the results for fragmented areas and smaller farms were less efficient, overestimating soybean areas.

Download Full-text

An Effective Method for Generating Spatiotemporally Continuous 30 m Vegetation Products

Remote Sensing ◽

10.3390/rs13040719 ◽

2021 ◽

Vol 13 (4) ◽

pp. 719

Author(s):

Xiuxia Li ◽

Shunlin Liang ◽

Huaan Jin

Keyword(s):

Vegetation Index ◽

Normalized Difference Vegetation Index ◽

Continuous Data ◽

Landsat Data ◽

Area Index ◽

Free Data ◽

Moderate Resolution ◽

Resolution Imaging ◽

Moderate Resolution Imaging Spectroradiometer ◽

Resolution Data

Leaf area index (LAI) and normalized difference vegetation index (NDVI) are key parameters for various applications. However, due to sensor tradeoff and cloud contaminations, these data are often temporally intermittent and spatially discontinuous. To address the discontinuities, this study proposed a method based on spectral matching of 30 m discontinuous values from Landsat data and 500 m temporally continuous values from Moderate-resolution Imaging Spectroradiometer (MODIS) data. Experiments have proven that the proposed method can effectively yield spatiotemporally continuous vegetation products at 30 m spatial resolution. The results for three different study areas with NDVI and LAI showed that the method performs well in restoring the time series, fills in the missing data, and reasonably predicts the images. Remarkably, the proposed method could address the issue when no cloud-free data pairs are available close to the prediction date, because of the temporal information “borrowed” from coarser resolution data. Hence, the proposed method can make better use of partially obscured images. The reconstructed spatiotemporally continuous data have great potential for monitoring vegetation, agriculture, and environmental dynamics.

Download Full-text

Vegetative growth of grasslands based on hyper-temporal NDVI data from the Modis sensor

Pesquisa Agropecuária Brasileira ◽

10.1590/s0100-204x2016000700009 ◽

2016 ◽

Vol 51 (7) ◽

pp. 858-868

Author(s):

Marcos Cicarini Hott ◽

Luis Marcelo Tavares de Carvalho ◽

Mauro Antonio Homem Antunes ◽

Polyanne Aguiar dos Santos ◽

Tássia Borges Arantes ◽

...

Keyword(s):

Vegetative Growth ◽

Vegetation Index ◽

Normalized Difference Vegetation Index ◽

High Growth ◽

Growth Index ◽

Data Series ◽

Moderate Resolution ◽

Resolution Imaging ◽

Moderate Resolution Imaging Spectroradiometer ◽

Very High

Abstract: The objective of this work was to analyze the development of grasslands in Zona da Mata, in the state of Minas Gerais, Brazil, between 2000 and 2013, using a parameter based on the growth index of the normalized difference vegetation index (NDVI) from the moderate resolution imaging spectroradiometer (Modis) data series. Based on temporal NDVI profiles, which were used as indicators of edaphoclimatic conditions, the growth index (GI) was estimated for 16-day periods throughout the spring season of 2012 to early 2013, being compared with the average GI from 2000 to 2011, used as the reference period. Currently, the grassland areas in Zona da Mata occupy approximately 1.2 million hectares. According to the used methods, 177,322 ha (14.61%) of these grassland areas have very low vegetative growth; 577,698 ha (45.96%) have low growth; 433,475 ha (35.72%) have balanced growth; 39,980 ha (3.29%) have high growth; and 5,032 ha (0.41%) have very high vegetative growth. The grasslands had predominantly low vegetative growth during the studied period, and the NDVI/Modis series is a useful source of data for regional assessments.

Download Full-text

Improved fuel moisture prediction in non-native tropical Megathyrsus maximus grasslands using Moderate-Resolution Imaging Spectroradiometer (MODIS)-derived vegetation indices

International Journal of Wildland Fire ◽

10.1071/wf16131 ◽

2017 ◽

Vol 26 (5) ◽

pp. 384

Author(s):

L. M. Ellsworth ◽

A. P. Dale ◽

C. M. Litton ◽

T. Miura

Keyword(s):

Vegetation Index ◽

Vegetation Indices ◽

Fuel Moisture ◽

Fire Behaviour ◽

Native Grass ◽

Tropical Ecosystems ◽

Moderate Resolution ◽

Resolution Imaging ◽

Live Fuel Moisture ◽

Moderate Resolution Imaging Spectroradiometer

The synergistic impacts of non-native grass invasion and frequent human-derived wildfires threaten endangered species, native ecosystems and developed land throughout the tropics. Fire behaviour models assist in fire prevention and management, but current models do not accurately predict fire in tropical ecosystems. Specifically, current models poorly predict fuel moisture, a key driver of fire behaviour. To address this limitation, we developed empirical models to predict fuel moisture in non-native tropical grasslands dominated by Megathyrsus maximus in Hawaii from Terra Moderate-Resolution Imaging Spectroradiometer (MODIS)-based vegetation indices. Best-performing MODIS-based predictive models for live fuel moisture included the two-band Enhanced Vegetation Index (EVI2) and Normalized Difference Vegetation Index (NDVI). Live fuel moisture models had modest (R2=0.46) predictive relationships, and outperformed the commonly used National Fire Danger Rating System (R2=0.37) and the Keetch–Byram Drought Index (R2=0.06). Dead fuel moisture was also best predicted by a model including EVI2 and NDVI, but predictive capacity was low (R2=0.19). Site-specific models improved model fit for live fuel moisture (R2=0.61), but limited extrapolation. Better predictions of fuel moisture will improve fire management in tropical ecosystems dominated by this widespread and problematic non-native grass.

Download Full-text

Development of a remote sensing-based rice yield forecasting model

Spanish Journal of Agricultural Research ◽

10.5424/sjar/2016143-8347 ◽

2016 ◽

Vol 14 (3) ◽

pp. e0907 ◽

Cited By ~ 14

Author(s):

Mostafa K. Mosleh ◽

Quazi K. Hassan ◽

Ehsan H. Chowdhury

Keyword(s):

Remote Sensing ◽

Vegetation Index ◽

Normalized Difference Vegetation Index ◽

Rice Yield ◽

Decision Makers ◽

Coefficient Of Determination ◽

Moderate Resolution ◽

Resolution Imaging ◽

Moderate Resolution Imaging Spectroradiometer ◽

Boro Rice

This study aimed to develop a remote sensing-based method for forecasting rice yield by considering vegetation greenness conditions during initial and peak greenness stages of the crop; and implemented for “boro” rice in Bangladeshi context. In this research, we used Moderate Resolution Imaging Spectroradiometer (MODIS)-derived two 16-day composite of normalized difference vegetation index (NDVI) images at 250 m spatial resolution acquired during the initial (January 1 to January 16) and peak greenness (March 23/24 to April 6/7 depending on leap year) stages in conjunction with secondary datasets (i.e., boro suitability map, and ground-based information) during 2007-2012 period. The method consisted of two components: (i) developing a model for delineating area under rice cultivation before harvesting; and (ii) forecasting rice yield as a function of NDVI. Our results demonstrated strong agreements between the model (i.e., MODIS-based) and ground-based area estimates during 2010-2012 period, i.e., coefficient of determination (R2); root mean square error (RMSE); and relative error (RE) in between 0.93 to 0.95; 30,519 to 37,451 ha; and ±10% respectively at the 23 district-levels. We also found good agreements between forecasted (i.e., MODIS-based) and ground-based yields during 2010-2012 period (R2 between 0.76 and 0.86; RMSE between 0.21 and 0.29 Mton/ha, and RE between -5.45% and 6.65%) at the 23 district-levels. We believe that our developments of forecasting the boro rice yield would be useful for the decision makers in addressing food security in Bangladesh.

Download Full-text

Índices de vegetação Modis aplicados na discriminação de áreas de soja

Pesquisa Agropecuária Brasileira ◽

10.1590/s0100-204x2012000900017 ◽

2012 ◽

Vol 47 (9) ◽

pp. 1317-1326 ◽

Cited By ~ 18

Author(s):

Joel Risso ◽

Rodrigo Rizzi ◽

Bernardo Friedrich Theodor Rudorff ◽

Marcos Adami ◽

Yosio Edemir Shimabukuro ◽

...

Keyword(s):

Mato Grosso ◽

Moderate Resolution ◽

Resolution Imaging ◽

Moderate Resolution Imaging Spectroradiometer

O objetivo deste trabalho foi avaliar o desempenho do índice de vegetação realçado (EVI) e do índice de vegetação da diferença normalizada (NDVI) - ambos do sensor "moderate resolution imaging spectroradiometer" (Modis) -, para discriminar áreas de soja das áreas de cana‑de‑açúcar, pastagem, cerrado e floresta, no Estado do Mato Grosso. Foram utilizadas imagens adquiridas em dois períodos: durante a entressafra e por ocasião do pleno desenvolvimento da cultura da soja. Para cada classe analisada, foram selecionadas 31 amostras de mapas de referência e avaliadas as diferenças nos valores de cada índice de vegetação, para a classe soja, foram avaliadas frente às demais classes, por meio do teste de Tukey‑Kramer. Em seguida, foram avaliadas as diferenças entre os índices de vegetação, por meio do teste de Wilcoxon pareado. O NDVI apresentou melhor desempenho na discriminação das áreas de soja na entressafra, particularmente com uso das imagens do dia do ano (DA) 161 a 273, enquanto o EVI apresentou melhor desempenho no período de pleno desenvolvimento da cultura, especificamente com uso das imagens de DA 353 a 33. Portanto, o melhor resultado para classificação da soja, no Estado do Mato Grosso, via séries temporais do sensor Modis, pode ser obtida por meio do uso combinado do NDVI na entresssafra e do EVI no pleno desenvolvimento da soja.

Download Full-text