Vegetation Types Mapping Using Multi-Temporal Landsat Images in the Google Earth Engine Platform

Aghababaei, Masoumeh; Ebrahimi‬, Ataollah; Naghipour, Ali Asghar; Asadi, Ebrahim; Verrelst, Jochem

doi:10.3390/rs13224683

Cited by 17 publications

(8 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Before 1990, Marjan was used as an agricultural area with intensive tilling and cultivation. From the year 2000, agricultural activities were reduced due to land abandonment, initiating a nature conservation and rehabilitation program [ 72 ]. The vegetation of Marjan went through natural succession, recovering to full canopy cover, with shrubs and perennial grasses becoming dominant.…”

Section: Methodsmentioning

confidence: 99%

Introducing ARTMO’s Machine-Learning Classification Algorithms Toolbox: Application to Plant-Type Detection in a Semi-Steppe Iranian Landscape

et al. 2022

Self Cite

View full text Add to dashboard Cite

Accurate plant-type (PT) detection forms an important basis for sustainable land management maintaining biodiversity and ecosystem services. In this sense, Sentinel-2 satellite images of the Copernicus program offer spatial, spectral, temporal, and radiometric characteristics with great potential for mapping and monitoring PTs. In addition, the selection of a best-performing algorithm needs to be considered for obtaining PT classification as accurate as possible . To date, no freely downloadable toolbox exists that brings the diversity of the latest supervised machine-learning classification algorithms (MLCAs) together into a single intuitive user-friendly graphical user interface (GUI). To fill this gap and to facilitate and automate the usage of MLCAs, here we present a novel GUI software package that allows systematically training, validating, and applying pixel-based MLCA models to remote sensing imagery. The so-called MLCA toolbox has been integrated within ARTMO’s software framework developed in Matlab which implements most of the state-of-the-art methods in the machine learning community. To demonstrate its utility, we chose a heterogeneous case study scene, a landscape in Southwest Iran to map PTs. In this area, four main PTs were identified, consisting of shrub land, grass land, semi-shrub land, and shrub land–grass land vegetation. Having developed 21 MLCAs using the same training and validation, datasets led to varying accuracy results. Gaussian process classifier (GPC) was validated as the top-performing classifier, with an overall accuracy (OA) of 90%. GPC follows a Laplace approximation to the Gaussian likelihood under the supervised classification framework, emerging as a very competitive alternative to common MLCAs. Random forests resulted in the second-best performance with an OA of 86%. Two other types of ensemble-learning algorithms, i.e., tree-ensemble learning (bagging) and decision tree (with error-correcting output codes), yielded an OA of 83% and 82%, respectively. Following, thirteen classifiers reported OA between 70% and 80%, and the remaining four classifiers reported an OA below 70%. We conclude that GPC substantially outperformed all classifiers, and thus, provides enormous potential for the classification of a diversity of land-cover types. In addition, its probabilistic formulation provides valuable band ranking information, as well as associated predictive variance at a pixel level. Nevertheless, as these are supervised (data-driven) classifiers, performances depend on the entered training data, meaning that an assessment of all MLCAs is crucial for any application. Our analysis demonstrated the efficacy of ARTMO’s MLCA toolbox for an automated evaluation of the classifiers and subsequent thematic mapping.

show abstract

Section: Methodsmentioning

confidence: 99%

Introducing ARTMO’s Machine-Learning Classification Algorithms Toolbox: Application to Plant-Type Detection in a Semi-Steppe Iranian Landscape

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…PEUs as a sub-class of rangeland cover are involved. Sub-classes of vegetation cover are more similar in terms of their spectral reflectance than that of a higher hierarchical land cover classification [7]. Thereby, PEUs classification process using an optimal time-series dataset is required to accurately identify and discriminate the past and current trends as well as to predict future trends of PEUs.…”

Section: Peus Multi-temporal Classification Maps For Three Periodsmentioning

confidence: 99%

“…Because of differences in elevation, soil, historic background, and land abandonment processes, a variety of plant communities with distinctive amounts and types of vegetation can be found in an area. Thereby, the identification of PEUs provides a reference for the interpretation of land cover data and research, monitoring, and land management [7]. PEUs mapping through classification techniques is extremely important, and their accuracy will directly affect the extraction of the final prediction results and change detection maps [8].…”

Section: Introductionmentioning

confidence: 99%

“…Therefore, these PEUs impose challenges to classification and predictions of future changes. In a related study about improving land cover map accuracy, Aghababaei et al [7] found that the development of an accurate land cover map is feasible in a semiarid rangeland when a dataset of time-series images is entered into the classification. So far, publicly available satellite data have been used in change detection studies, such as MODIS [16], SPOT [3], and Landsat [17,18].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Monitoring of Plant Ecological Units Cover Dynamics in a Semiarid Landscape from Past to Future Using Multi-Layer Perceptron and Markov Chain Model

Aghababaei,

Ebrahimi,

Naghipour

et al. 2024

Remote Sensing

Self Cite

View full text Add to dashboard Cite

Anthropogenic activities and natural disturbances cause changes in natural ecosystems, leading to altered Plant Ecological Units (PEUs). Despite a long history of land use and land cover change detection, the creation of change detection maps of PEUs remains problematic, especially in arid and semiarid landscape. This study aimed to determine and describe the changes in PEUs patterns in the past and present, and also predict and monitor future PEUs dynamics using the multi-layer perceptron-Markov chain (MLP-MC) model in a semiarid landscape in Central Zagros, Iran. Analysis of PEUs classification maps formed the basis for the identification of the main drivers in PEUs changes. First, an optimal time-series dataset of Landsat images were selected to derive PEUs classification maps in three periods, each separated by 16 years. Then, PEUs multi-temporal maps classified for period 1 (years 1986–1988) period 2 (years 2002–2004), and period 3 (years 2018–2020) were employed to analyze and predict PEUs dynamics. The dominant transitions were identified, and the transition potential was determined by developing twelve sub-models in the final change prediction process. Transitions were modeled using a Multi-Layer Perceptron (MLP) algorithm. To predict the PEU map for period 3, two PEUs classification maps of period 1 and period 2 were used using the MLP-MC method. The classified map and the predicted map of period 3 were used to evaluate and validate the predicted results. Finally, based on the results, transitions of future PEUs were predicted for the year 2036. The MLP-MC model proved to be a powerful model that can predict future PEUs dynamics that are the result of current human and managerial activities. The findings of this study demonstrate that the impact of anthropogenic processes and management activities will become visible in the natural environment and ecosystem in less than a decade.

show abstract

“…On account of the comparatively finer spatial resolution with a long-time sequence, the Landsat series has become a popular data source for characterizing forest dynamics [12][13][14]. Furthermore, Google Earth Engine (GEE) is a powerful geo-big data computing platform that combines spatiotemporally spectral features for large-scale forest-type classification and produces a series of multi-scale maps [15][16][17]. Zhang et al [18] developed a Landsat-based global 30 m land-use map with a detailed classification system of forest types using the metric composite method on GEE and a multi-temporal random forest model, of which the overall accuracy is 82.5%.…”

Section: Introductionmentioning

confidence: 99%

Annually Spatial Pattern Dynamics of Forest Types under a Rapid Expansion of Impervious Surfaces: A Case Study of Hangzhou City

Zhu,

Zhou,

Liu

et al. 2023

Forests

View full text Add to dashboard Cite

Dramatic forest dynamics strongly influence pressure mitigation from the increasing population and climate changes of an urban landscape. Effectively monitoring landscapes in a spatiotemporally consistent manner, satellite remote sensing has emerged as the first analytical tool to help us understand the changes in urban forests. At present, most studies focus on classification algorithms, spatial analysis methods, and ecosystem models, and the literature lacks spatiotemporally explicit research on the responses of different types of forests to urbanization. Thus, in this study, Hangzhou was selected as a typical metropolitan area to determine the annual spatial patterns of urban forests at a forest-type level. To illustrate the spatial pattern dynamics of different forest types resulting from rapid urbanization, this study characterized the Landsat-based spatial patterns of different forest types, as well as their annual changes from 2000 to 2022 using object-based backdating classification, land-use transfer matrix, area-weighted centroids, and landscape pattern indexes. The spatiotemporal effects of impervious surface expansion on forest pattern changes at a type scale were discussed. The results demonstrated that forests, mainly located in the southwest, decreased from 11,660.69 to 11,516.15 km2. Moreover, evergreen broadleaved forests occupied the largest area and had the most decreased ratio among the three forest types over 23 years, followed by evergreen needle-leaved and deciduous broadleaved forests. In total, 103.37 km2 of forest areas transformed to impervious surfaces, with the highest annual transformation of area occurring among evergreen broadleaved forests widely across Hangzhou City and the lowest occurring among deciduous broadleaved forests. Forests lost adjacency due to the development of Hangzhou City, while this southwestward shrinkage slowed down over 23 years, resulting in the highest increase in the degree of evergreen broadleaved forest fragmentation. Therefore, measures of city planning according to the deep effects of adjustments of administrative divisions to forest suitability should be implemented, such as green ecological corridor construction. This research provides a Landsat-based methodology at a spatiotemporally explicit-scale perspective for better understanding forest changes under high-speed urbanization.

show abstract

Vegetation Types Mapping Using Multi-Temporal Landsat Images in the Google Earth Engine Platform

Cited by 17 publications

References 38 publications

Introducing ARTMO’s Machine-Learning Classification Algorithms Toolbox: Application to Plant-Type Detection in a Semi-Steppe Iranian Landscape

Introducing ARTMO’s Machine-Learning Classification Algorithms Toolbox: Application to Plant-Type Detection in a Semi-Steppe Iranian Landscape

Monitoring of Plant Ecological Units Cover Dynamics in a Semiarid Landscape from Past to Future Using Multi-Layer Perceptron and Markov Chain Model

Annually Spatial Pattern Dynamics of Forest Types under a Rapid Expansion of Impervious Surfaces: A Case Study of Hangzhou City

Contact Info

Product

Resources

About