Automated Prediction System for Vegetation Cover Based on Modis-Ndvi Satellite Data and Neural Networks

Abujayyab, Sohaib K. M.; Karaş, İsmail Rakıp

doi:10.5194/isprs-archives-xlii-4-w19-9-2019

Cited by 7 publications

(4 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…NDVI values were calculated from the near infrared (band 5) and red (band 4) bands of Landsat 8 imagery to identify areas on a spectrum between −1.0 and 1.0 to represent a lack of green vegetation and dense green vegetation, respectively. Many studies have shown that NDVI values can be used to determine LULC classes and can be projected forward to forecast vegetation cover [54][55][56][57][58][59][60][61]. Landsat 8 bands for the month of January in 2015 and 2020 were used to ensure cloud free imagery with similar seasonality, and Equation (2) was used to calculate the NDVI.…”

Section: Land Use/land Cover Scenariosmentioning

confidence: 99%

A Framework for Calculating Peak Discharge and Flood Inundation in Ungauged Urban Watersheds Using Remotely Sensed Precipitation Data: A Case Study in Freetown, Sierra Leone

et al. 2021

View full text Add to dashboard Cite

As the human population increases, land cover is converted from vegetation to urban development, causing increased runoff from precipitation events. Additional runoff leads to more frequent and more intense floods. In urban areas, these flood events are often catastrophic due to infrastructure built along the riverbank and within the floodplains. Sufficient data allow for flood modeling used to implement proper warning signals and evacuation plans, however, in least developed countries (LDC), the lack of field data for precipitation and river flows makes hydrologic and hydraulic modeling difficult. Within the most recent data revolution, the availability of remotely sensed data for land use/land cover (LULC), flood mapping, and precipitation estimates has increased, however, flood mapping in urban areas of LDC is still limited due to low resolution of remotely sensed data (LULC, soil properties, and terrain), cloud cover, and the lack of field data for model calibration. This study utilizes remotely sensed precipitation, LULC, soil properties, and digital elevation model data to estimate peak discharge and map simulated flood extents of urban rivers in ungauged watersheds for current and future LULC scenarios. A normalized difference vegetation index (NDVI) analysis was proposed to predict a future LULC. Additionally, return period precipitation events were calculated using the theoretical extreme value distribution approach with two remotely sensed precipitation datasets. Three calculation methods for peak discharge (curve number and lag method, curve number and graphical TR-55 method, and the rational equation) were performed and compared to a separate Soil and Water Assessment Tool (SWAT) analysis to determine the method that best represents urban rivers. HEC-RAS was then used to map the simulated flood extents from the peak discharges and ArcGIS helped to determine infrastructure and population affected by the floods. Finally, the simulated flood extents from HEC-RAS were compared to historic flood event points, images of flood events, and global surface water maximum water extent data. This analysis indicates that where field data are absent, remotely sensed monthly precipitation data from Integrated Multi-satellitE Retrievals for GPM (IMERG) where GPM is the Global Precipitation Mission can be used with the curve number and lag method to approximate peak discharges and input into HEC-RAS to represent the simulated flood extents experienced. This work contains a case study for seven urban rivers in Freetown, Sierra Leone.

show abstract

Section: Land Use/land Cover Scenariosmentioning

confidence: 99%

A Framework for Calculating Peak Discharge and Flood Inundation in Ungauged Urban Watersheds Using Remotely Sensed Precipitation Data: A Case Study in Freetown, Sierra Leone

et al. 2021

View full text Add to dashboard Cite

show abstract

“…The second category focused on (ii) analyzing the data generated from the interaction between the learner and the learning system (e.g., de Vicente, 2003;Qu. and Johnson, 2005;Ramaha, 2012;Ramaha, 2017;Abujayyab and Karaş, 2019), these data are usually saved by the system in a log file, these types of researches applicable for attention detection in elearning systems. However, those research approaches still inaccurate, and are complex to design and maintain.…”

Section: Introductionmentioning

confidence: 99%

Towards Webcam-Based Face Direction Tracking to Detect Learners' Attention Within Asynchronous E-Learning Environment

Ramaha

Karaş²,

Gül

et al. 2021

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

Self Cite

View full text Add to dashboard Cite

Abstract. Recently, as a consequence of COVID-19 pandemic, the delivery of education at most of the educational institutions depended mainly on e-learning. So, the researchers give more attention for both synchronous and asynchronous e-learning. Although from an economical perspective, asynchronous e-learning seems to be the best e-learning option for institutions, still one of the biggest challenges is how to keep learners motivated for the entire learning process. One of important motivational factors that drives the success of the learning process is the learner attention. Therefore, to retain the learners' attention during the asynchronous e-learning process, we need first to detect their loss of attention. Accordingly, more studies started to focus on detecting learners’ attention. However, those studies can't be widely used for attention detection within asynchronous e-learning environments, as the used approaches tend to be inaccurate, and complex for the design and maintain. In contrast, in this study, we explore the possibility to find a simple way that can be widely used to detect learners' attention within the asynchronous e-learning environments. Therefore, we used webcams which are available in almost every laptop, and computer vision tools to detect learners' attention by tracking their faces. Thereafter, we evaluated the accuracy of our suggested method, the result of this evaluation showed that our method is efficient.

show abstract

“…Vegetation cover plays a critical role in sustaining life on earth by acting as a shelter for fauna, and minimizing air and water pollution (Abujayyab, Karaş, 2019). However, terrestrial ecosystems face environmental deforestation, destruction, and degradation, which are the main sources of greenhouse gas emissions (Poortinga et al, 2018) resulting from harmful anthropogenic activities (Šalić, A., and Zelić, 2018).…”

Section: Introductionmentioning

confidence: 99%

“…The local climate of an area varies with the changes and transition of vegetation cover which depends on several factors such as timing and location (Duveiller et al, 2018). The prediction of changes in vegetation cover will help to provide timely insights for decision-makers to effectively manage vegetation in a region (Abujayyab, Karaş, 2019).…”

Section: Introductionmentioning

confidence: 99%

Modeling and Forecasting Cumulative Evi Anomalies Using Sarima for Biophysical Monitoring: A Case Study in the Philippines

Jr.

Duran

2021

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

Abstract. Understanding changes in vegetation cover that affect the biophysical conditions of a region can help in formulating policies to address current and future problems of terrestrial ecosystems such as deforestation and environmental degradation. This study focuses on developing a model that forecasts the cumulative Enhanced Vegetation Index (EVI) anomalies as a tool for biophysical conditions monitoring in the Philippines. Satellite data from MODIS MYD13Q1 V6, which contains vegetation index per pixel at 16-day intervals with a resolution of 250 meters, were utilized. The cumulative EVI anomalies per instant were calculated in Google Earth Engine by aggregating the difference of a specific data point in 2011–2020 to a reference EVI mean computed from 2001–2010. The Error-Trend-Seasonality model shows that the cumulative EVI anomalies graph is non-stationary with an upward trend and seasonality. The upward trend of the cumulative EVI anomalies indicates the improvement of vegetation in the Philippines. To check the stationarity of the cumulative EVI anomalies data, the Augmented Dickey-Fuller test was utilized and the model was generated using Seasonal Autoregressive Integrated Moving Average model. Based on the analysis, the best-fit model for the cumulative EVI anomalies is SARIMA (1,1,0)(1,1,1)12 with a mean absolute percentage error (MAPE) of 13.26%. Thus, the proposed model can be used as a tool for biophysical assessment by monitoring and forecasting changes in vegetation and contribute to attaining the UN Sustainable Development Goals 2 and 15 – ‘Eliminating Hunger’ and ‘Life on Land’.

show abstract

Automated Prediction System for Vegetation Cover Based on Modis-Ndvi Satellite Data and Neural Networks

Cited by 7 publications

References 15 publications

A Framework for Calculating Peak Discharge and Flood Inundation in Ungauged Urban Watersheds Using Remotely Sensed Precipitation Data: A Case Study in Freetown, Sierra Leone

A Framework for Calculating Peak Discharge and Flood Inundation in Ungauged Urban Watersheds Using Remotely Sensed Precipitation Data: A Case Study in Freetown, Sierra Leone

Towards Webcam-Based Face Direction Tracking to Detect Learners' Attention Within Asynchronous E-Learning Environment

Modeling and Forecasting Cumulative Evi Anomalies Using Sarima for Biophysical Monitoring: A Case Study in the Philippines

Contact Info

Product

Resources

About