The data of remote sensing of the Earth’s surface, obtained with the help of sensors in different spectral ranges, installed on aerospace crafts, are characterized by high informativeness, reliability, and operational accessibility, as a result, they are effectively used to solve a wide range of problems of monitoring the natural environment and anthropogenic objects.
The accumulated for a half-century fund of aerial and space imagery is one of the most informative sources of information about dynamic geographical phenomena and processes. The development of studies of the dynamics of phenomena with the help of aerospace methods, the development of methodological approaches, the theory, practical methods and organization of work made it possible to single out the dynamic aerospace sounding as a special direction, as an independent method for studying this phenomena and as a means of improving maps in dynamics. The most important directions on using this data are:
– A study of natural resources;
– An engineering assessment of the area;
– An ecological monitoring;
– Monitoring of natural and man-caused disasters.
The greatest technical and economic effect from the use of the airspace data of the Earth can be obtained through a comprehensive study and mapping of natural, both agricultural and industrial installations, as well as engineering structures. Aerospace images obtained by remote sensing of the earth’s surface, being multi-purpose, function as a single basis for conducting complex, interrelated studies of the natural environment. The results of the thematic processing of aerospace images can be presented as a series of mutually agreed thematic maps reflecting spatial location, qualitative and quantitative characteristics of natural and economic objects on a territory in question.
Aerospace images contain very valuable information about the components of natural and territorial complexes since all these components are reflected simultaneously. Landscapes are indicators for determining the properties of the various components of the natural environment. Very often, complex or sectoral case studies use the landscape indicator method of data interpretation. It is adapted the most in geological, agricultural, and hydrogeological research. The landscape indication consists in determining the hard-to-observe components with the physiognomic, easily observable components. Aerospace images provide excellent opportunities for studying trends in the dynamics of natural and agro-industrial objects, including under the influence of anthropogenic impact. Identification of long-term trends in the development of natural and artificial objects of individual regions is done by comparing different photographs with old survey-topographic maps. Thematic processing of images is the identification, recognition, and determination of the characteristics of the objects, displayed on the picture. The procedure for thematic processing is divided into successive stages, the main ones of which are segmentation and interpretation.
In the case of thematic processing of aerospace survey data, information classes are defined as a set of objects that need to be selected according to the requirements of the thematic task being solved and which are separated in the available data set. The main problem with thematic segmentation of complex images of the Earth’s surface is that the use of classification algorithms using certain homogeneity conditions and a certain space of classification characteristics often leads to the isolation of segments that do not correspond to the information classes of the thematic task being solved. To overcome the difficulties encountered when developing methods for processing remote sensing data, it is necessary to apply multilevel segmentation algorithms using various homogeneity conditions and different subsets of classification characteristics that take into account the structural and spatial characteristics of the image in order to solve problems of monitoring the natural environment and anthropogenic objects. The ways of solving this problem are defined as the development of the theory and methodology of multilevel thematic processing of aerospace images of the earth’s surface, carrying out experimental studies of the effectiveness of the developed multilevel processing methods in solving various thematic problems, and a study of the possibilities of increasing the effectiveness of thematic segmentation through the use of the parallel image processing technology. Also, mathematical statistics and digital image processing methods are used.
The theory and practical methodology of multilevel thematic processing of aerospace images of the Earth’s surface were developed last time. These methods can be widely used in the automated systems for controlling and forecasting of the surrounding environment, in forestry and agriculture, in meteorology, in dynamic mapping, and during monitoring the consequences of emergencies. The study of terrestrial natural ecosystems and anthropogenic objects requires new methods and algorithms for processing information in order to their further sustainable development.