DEVELOPMENT OF INFORMATION TECHNOLOGY FOR EMERGENCY MONITORING USING COMPUTER VISION METHODS

Abstract

The paper presents the results of IT monitoring of natural emergencies. The types of monitoring studies that are carried out within IT in order to prevent (reduce the risk of occurrence), as well as reduce material losses from emergencies (mitigation), in particular monitoring of the state of natural hazards and processes, monitoring of emergency sources, environmental monitoring, monitoring of fires in geosystems of a natural nature are described. The modes of conducting monitoring studies are described: before the occurrence of an emergency in the mode of daily functioning, at the time of the threat of an emergency in high alert mode, after the occurrence of an emergency in the emergency mode. Information processes within IT emergency monitoring are described. IT consists of interrelated information processes - a set of procedures for collecting and registering satellite information, preparing information arrays, processing, accumulating and storing data, transferring data from sources of occurrence to the place of processing, and results - to consumers of information for making management decisions.  The roles of users of IT monitoring of emergencies are defined: Head of emergency response, Head of the State Emergency Service, Commander of the State Emergency Service, Head of the fire and rescue unit and his deputies. Assigning user roles allows you to control the actions that each user can perform within IT emergency monitoring, optimizing the workflow. A use diagram was built for the role of "Commander of the emergency rescue department of the State Emergency Service". IT emergency monitoring database was developed using MySQL Workbench - a graphical tool for working with MySQL servers and databases. The structural and logical diagram of the database is constructed. Operational solution of emergency monitoring tasks in geosystems of a natural nature (forest fires) is carried out according to the materials of multispectral space surveys obtained from the Landsat 8 satellite. Landsat 8 takes data using two different sensors - the Operational Land Imager (OLI) and the Thermal Infrared Sensor (TIRS), which collect data in nine shortwave bands and two longwave thermal bands.

To identify the outline of a forest fire, a segmentation method was applied by highlighting the boundaries of objects. Examples of segmentation of forest fires are given.