Modeling land cover change based on an artificial neural network for a semiarid river basin in northeastern Brazil

Silva, Leonardo Pereira e; Xavier, Ana Paula Campos; Silva, Richarde Marques da; Santos, Celso Augusto Guimarães

doi:10.1016/j.gecco.2019.e00811

Cited by 65 publications

(37 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The spatial resolution of the Landsat satellite images was 30 m. The data were obtained free from the United States Geological Survey (USGS) platform (US Geological Survey, 2019). Because clouds are an obstacle for the interpretation and classification of satellite images (Silva et al, 2019), the chosen images from Landsat-5 (Thematic Mapper [TM] sensor) and Landsat-8 (Operational Land Imager [OLI] sensor) do not contain clouds over the study area. Supervised and unsupervised classification are the two main classification methods used for spectral images.…”

Section: Processing Of the Satellite Images (Classification)mentioning

confidence: 99%

See 1 more Smart Citation

Evaluation of Change in Land Usage and Land Cover in Karaj, Iran

Mohammadyari

Mirsanjari

Visockienė

et al. 2020

The 11th International Conference ENVIRONMENTAL ENGINEERING 11th ICEE SELECTED PAPERS

View full text Add to dashboard Cite

In this study, classification results were derived from remote sensing data and the Support Vector Machine (SVM) algorithm used in this process, which classifies Landsat land-cover images. The accuracy of image classifications was evaluated by calculation of the Kappa coefficient. The area of study is Karaj, the capital of Alborz province, in north-central Iran. It is situated in the foothills of the Alborz Mountains and occupies a fertile agricultural plain. Landsat data used in the classification of land cover were collected from USGS websites, and multi-temporal images from the data were geometrically corrected. After this process, we calculated 11 metrics at the landscape and class-level scales: five metrics of class level and six metrics of landscape. The results showed that the landscape patterns in Karaj were changed due to the process of urbanization over an 11-year period. At the class level, for all classifications, the AI metric increased and the PD and NP metrics decreased. At the landscape level, the PD, ED, NP, and SHDI metrics decreased, and LPI and AI increased. These results provide insights about urban development policies and about whether the expansion of urban areas is beneficial for environmental sustainability in Iran and elsewhere in the world.

show abstract

Section: Processing Of the Satellite Images (Classification)mentioning

confidence: 99%

“…The accuracy (quality) of classification comes from the calculation of the kappa coefficient (Silva et al, 2019). This was obtained from the following equation (Cohen, 1960;Silva et al, 2019):…”

Section: Processing Of the Satellite Images (Classification)mentioning

confidence: 99%

Evaluation of Change in Land Usage and Land Cover in Karaj, Iran

Mohammadyari

Mirsanjari

Visockienė

et al. 2020

The 11th International Conference ENVIRONMENTAL ENGINEERING 11th ICEE SELECTED PAPERS

View full text Add to dashboard Cite

show abstract

“…The evolution in the use of RS and GIS the past two decades, has led to the application of new spatial modelling approaches such as artificial neural networks (ANN) and cellular automata [30]. The highly sophisticated "collective behaviour" of ANN has made it possible for them to be employed in complex domains such as land use changes [32], as many studies have used them to model land use/cover changes [3,26,[33][34][35].…”

Section: Introductionmentioning

confidence: 99%

Modelling of Expansion Changes of Vilnius City Area and Impacts on Landscape Patterns Using an Artificial Neural Network

Mirsanjari

Visockienė

Mohammadyari

et al. 2021

Ecological Chemistry and Engineering S

View full text Add to dashboard Cite

The present study aimed to analyse changes in the land cover of Vilnius city and its surrounding areas and propose a scenario for their future changes using an Artificial Neural Network. The land cover dynamics modelling was based on a multilayer perceptron neural network. Landscape metrics at a class and landscape level were evaluated to determine the amount of changes in the land uses. As the results showed, the Built-up area class increased, while the forest (Semi forest and Dense forest) classes decreased during the period from 1999 to 2019. The predicted scenario showed a considerable increase of about 60 % in the Built-up area until 2039. The vegetation plant areas consist about 47 % of all the area in 2019, but it will be 36 % in 2039, if this trend (urban expansion) continues in the further. The findings further indicated the major urban expansion in the vegetation areas. However, Built-up area would expand over Semi forest land and Dense forest land, with a large part of them changed into built- up areas.

show abstract

“…ANN classifiers have been widely used in satellite image classification, due to the capability to adapt and generalise different input data structures [26]. The successful application of ANN has been well demonstrated in different remote sensing contexts, including classification of endangered tree species [34] and dynamic modelling of land cover changes in semi-arid landscapes [35].…”

Section: Introductionmentioning

confidence: 99%

Probabilistic Mapping and Spatial Pattern Analysis of Grazing Lawns in Southern African Savannahs Using WorldView-3 Imagery and Machine Learning Techniques

et al. 2020

View full text Add to dashboard Cite

Savannah grazing lawns are a key food resource for large herbivores such as blue wildebeest (Connochaetes taurinus), hippopotamus (Hippopotamus amphibius) and white rhino (Ceratotherium simum), and impact herbivore densities, movement and recruitment rates. They also exert a strong influence on fire behaviour including frequency, intensity and spread. Thus, variation in grazing lawn cover can have a profound impact on broader savannah ecosystem dynamics. However, knowledge of their present cover and distribution is limited. Importantly, we lack a robust, broad-scale approach for detecting and monitoring grazing lawns, which is critical to enhancing understanding of the ecology of these vital grassland systems. We selected two sites in the Lower Sabie and Satara regions of Kruger National Park, South Africa with mesic and semiarid conditions, respectively. Using spectral and texture features derived from WorldView-3 imagery, we (i) parameterised and assessed the quality of Random Forest (RF), Support Vector Machines (SVM), Classification and Regression Trees (CART) and Multilayer Perceptron (MLP) models for general discrimination of plant functional types (PFTs) within a sub-area of the Lower Sabie landscape, and (ii) compared model performance for probabilistic mapping of grazing lawns in the broader Lower Sabie and Satara landscapes. Further, we used spatial metrics to analyse spatial patterns in grazing lawn distribution in both landscapes along a gradient of distance from waterbodies. All machine learning models achieved high F-scores (F1) and overall accuracy (OA) scores in general savannah PFTs classification, with RF (F1 = 95.73±0.004%, OA = 94.16±0.004%), SVM (F1 = 95.64±0.002%, OA = 94.02±0.002%) and MLP (F1 = 95.71±0.003%, OA = 94.27±0.003%) forming a cluster of the better performing models and marginally outperforming CART (F1 = 92.74±0.006%, OA = 90.93±0.003%). Grazing lawn detection accuracy followed a similar trend within the Lower Sabie landscape, with RF, SVM, MLP and CART achieving F-scores of 0.89, 0.93, 0.94 and 0.81, respectively. Transferring models to the Satara landscape however resulted in relatively lower but high grazing lawn detection accuracies across models (RF = 0.87, SVM = 0.88, MLP = 0.85 and CART = 0.75). Results from spatial pattern analysis revealed a relatively higher proportion of grazing lawn cover under semiarid savannah conditions (Satara) compared to the mesic savannah landscape (Lower Sabie). Additionally, the results show strong negative correlation between grazing lawn spatial structure (fractional cover, patch size and connectivity) and distance from waterbodies, with larger and contiguous grazing lawn patches occurring in close proximity to waterbodies in both landscapes. The proposed machine learning approach provides a novel and robust workflow for accurate and consistent landscape-scale monitoring of grazing lawns, while our findings and research outputs provide timely information critical for understanding habitat heterogeneity in southern African savannahs.

show abstract

Modeling land cover change based on an artificial neural network for a semiarid river basin in northeastern Brazil

Cited by 65 publications

References 48 publications

Evaluation of Change in Land Usage and Land Cover in Karaj, Iran

Evaluation of Change in Land Usage and Land Cover in Karaj, Iran

Modelling of Expansion Changes of Vilnius City Area and Impacts on Landscape Patterns Using an Artificial Neural Network

Probabilistic Mapping and Spatial Pattern Analysis of Grazing Lawns in Southern African Savannahs Using WorldView-3 Imagery and Machine Learning Techniques

Contact Info

Product

Resources

About