The dissertation is devoted to the development of information technology to expand the functionality of non-specialized unmanned aerial vehicles for the early detection of forest fires.
On the basis of the analysis performed, it can be stated that one of the main requirements for the protection of forests from fire is effective monitoring, ensuring the earliest possible detection of hot spots. Compliance with this requirement is impossible without the use of information technologies that ensure the early detection of the danger of a forest fire and the timely notification of the relevant services. Their introduction into the practice of the forest fire fighting system is the key to the rapid response of emergency crews and the prevention of the growth of small hot spots of forest fire into a large-scale environmental catastrophe.
The developed information technology is designed to eliminate the existing contradiction between the high level of costs for the development of high-tech fire-fighting equipment and their insufficient efficiency, as evidenced by the analysis of statistical data.
The essence of the proposed information technology is to attract additional nontechnical resources, namely, unmanned aerial vehicles, whose owners are individual users, to the fire monitoring process, and detection of forest fires in the early stages of fire. Involving individual operators of unmanned aerial vehicles – participants of mass events held outdoors to the cooperation, will increase the effectiveness of fire monitoring. During scheduled flights, an unmanned aerial vehicle, in parallel with its scheduled tasks, performs an additional function – detects forest fires and notifies the appropriate services about the danger of forest fires, which makes it possible to improve a modern forest fire protection system without additional costs for high-tech equipment.
To solve the problems stated in the study, there has been developed a method for processing video streams that allows making automated search of forest fires in the background using parallel computing, which ensures distribution and optimization of the use of unmanned aerial vehicle resources and control devices. The solution of the information technology problem of ensuring the efficiency of obtaining information on the presence of forest fires is provided by improving the methods of processing digital images and video sequences, which allowed them to be used in the course of monitoring and evaluating extreme situations on-line. In the course of the study, the automatic adaptation of mathematical parameters to the technical characteristics of control devices improved the wavelet transform method and the method of frame-by-frame video processing depending on the frame rate, which allows monitoring in the background using standard control devices such as tablets, smartphones, etc. . The method of automatic detection of forest fires has been improved on the basis of automatic determination of the relative number of image points belonging to a fuzzy cluster of unnatural colors, which makes it possible to increase the efficiency of determining fire sources on digital images. The developed information technology for the early detection of forest fires, suggesting the expansion of the functional information and technological capabilities of modern unmanned aerial vehicles due to the multiprocessor and multi-machine architecture, provides the possibility of timely notification of relevant services about the presence of a forest fire source, preventing the occurrence of a large-scale forest fire. In the course of the study, software was developed that allows implementing the created information technology. Field experiments were carried out, which allowed setting the optimal settings for clustering and wavelet transform parameters for processing digital images transmitted from an unmanned aircraft. The comparative analysis of the results of the functioning of the information system of automatic detection of digital sources of fire on digital images and the results of a similar task performed by a human operator, confirmed the effectiveness of the proposed information technology. In the future the developed information technology will make it possible to create decision-taking support systems based on expanding the functional information and technological capabilities of non-specialized information and technical resources that can be used in systems for assessing the current operational fire situation, predicting fire danger, and classifying forests according to fire risk, calculation of optimization of time, financial and technical costs for the organization of work on extinguishing and elimination of the consequences of forest fires, will ensure increased environmental sustainability of forestry.
