Fully Convolutional Neural Networks for Mapping Oil Palm Plantations in Kalimantan
This research is motivated by the global warming problem, which is likely influenced by human activity. Fast-growing oil palm plantations in the tropical belt of Africa, Southeast Asia and parts of Brazil lead to significant loss of rainforest and contribute to the global warming by the corresponding decrease of carbon dioxide absorption. We propose a novel approach to monitoring of the development of such plantations based on an application of state-of-the-art Fully Convolutional Neural Networks (FCNs) to solve Semantic Segmentation Problem for Landsat imagery.
The land area covered by freely available very high-resolution (VHR) imagery has grown dramatically over recent years, which has considerable relevance for forest observation and monitoring. For example, it is possible to recognize and extract a number of features related to forest type, forest management, degradation and disturbance using VHR imagery. Moreover, time series of medium-to-high-resolution imagery such as MODIS, Landsat or Sentinel has allowed for monitoring of parameters related to forest cover change. Although automatic classification is used regularly to monitor forests using medium-resolution imagery, VHR imagery and changes in web-based technology have opened up new possibilities for the role of visual interpretation in forest observation. Visual interpretation of VHR is typically employed to provide training and/or validation data for other remote sensing-based techniques or to derive statistics directly on forest cover/forest cover change over large regions. Hence, this paper reviews the state of the art in tools designed for visual interpretation of VHR, including Geo-Wiki, LACO-Wiki and Collect Earth as well as issues related to interpretation of VHR imagery and approaches to quality assurance. We have also listed a number of success stories where visual interpretation plays a crucial role, including a global forest mask harmonized with FAO FRA country statistics; estimation of dryland forest area; quantification of deforestation; national reporting to the UNFCCC; and drivers of forest change.
The article analyzes the near-Earth space as a future habitat for humankind. This article investigates the factors affecting the location in this environment. We estimate the boundaries of space and related space. The article highlights the main features of the near-Earth space as a human-friendly environment.
For the first time spatio-temporal characteristics of air pollution by sulfur dioxide (SO2), nitrogen dioxide (NO2), carbon dioxide (CO2), carbon monoxide (CO) and aerosol over Ukraine and Europe are established. It was shown that moderate risks of air pollution by sulfur dioxide of eastern and western parts of Ukraine relate as 2:1. It was shown that values of moderate and high risks of European areas most polluted by aerosol (except the north of Italy) and Ukraine (Kyiv, Donetsk and Odessa regions) are approximately related as 1:1. Moderate levels of risks for Kiev, Donetsk and Odessa regions relate to moderate risk levels of other Ukrainian regions as 1.8:1. The maximum risk value of moderate pollution by nitrogen dioxide of the atmosphere of Europe and Ukraine relate as 3:1. The analysis of concentration dynamics of carbon dioxide for atmosphere of the whole earth for the last 8 years (2004–2011) revealed the increase for more than 20 ppm. It is shown that the atmosphere of Ukraine exposed to the same level of carbon monoxide pollution, as the atmosphere of other European countries.
For the first time, using satellite Earth remote sensing data, the maps of air pollution risks by nitrogen dioxide (NO2) over the territory of Europe with spatial resolution of 0.25º×0.25º (approximately 27.5 km × 18 km for the 48º latitude) were created. The suggested risk calculation technique is simple yet delivers extensive understanding of typical air pollution character. It is shown that the highest risks of air pollution by nitrogen dioxide in Europe are observed over Germany, Belgium, Netherlands and southern part of the North Sea as well as over large cities.
The archives of measurements at a network of stations of Roshydromet stocks of available water capacity and satellite measurements of relative soil moisture topsoil according with the instrument ASCAT (MetOp satellites) were used. The estimation of the statistical structure of the fields of available water capacity in the upper 10- and 20-cm soil layers, correlations were found between the data of remote sensing (RS) data and agrometeorological stations. Developed procedure for automated pre-control ground-based measurements. The algorithm is statistically optimal transform of remote sensing data in estimating the amount of moisture in the upper soil layer.
To date, all remote sensing data are represented and stored as temporal sequences of separate “snapshots” – rasters or grids. This makes impossible to quickly obtain a time series of a variable values for the full available period for a region of a coordinate grid. Trend research – one of the most important topics in Earth science – becomes extremely complex and time consuming. This paper proposes an alternative data representation and corresponding storage technique. The data are represented as a collection of individual time series, one per each grid cell or raster pixel. New storage layout enables any time series to be always readily accessible. This approach considerably facilitates the application of existing time series techniques to remote sensing, climate reanalysis and similar data as well as provides new research and development opportunities not available before.