| Summary: | The article explores the application of the approach using lattice models and the theory of cellular automata in identifying extended objects in images obtained during monitoring of urbanized areas. Such images have a certain degree of "blurriness" caused not only by the limitations of the images themselves, but also by the incompleteness of the accepted object model, processing algorithms, thermodynamic and quantum effects. For the created methodology and its software implementation, a study was carried out in order to assess the efficiency of obtaining results and the quality of work. Parameters of algorithms for segmentation and identification of objects on the earth's surface were selected as evaluation criteria. A previously developed binary image filtering device is considered as a variant of a cellular automaton. The purpose of developing a device of such a filtration device is to increase the speed by parallelizing the procedures performed, which is characteristic of a cellular automaton having a parallel (not "von Neumann") architecture. A scheme of the memory matrix element of the device in question is presented. It is shown that the set of identification features can be expanded due to the elements of triangulation using. The injection of triangulation elements and supplementation of additional reference points during the construction of the triangulation grid can be used in the monitoring process to identify potentially vulnerable objects. In addition, the proposed technique allows extracting new information from images about such objects. An example of such information is also presented in the article. The results obtained make it possible to perceive with optimism the ongoing developments and to recommend the use of the developed technique for the operational identification of extended objects during remote sensing of the Earth.
|
|---|