Development of an Environmental Monitoring System Based on Spatial Marking and Machine Vision Technologies

Abstract

Introduction. The use of available satellite images and aerial photography by unmanned aerial vehicles (UAVs) in the tasks of environmental monitoring is challenged by the imperfection of existing tools. Geographic information systems are characterized by insufficient flexibility to automatically work with heterogeneous sources. The latest models based on artificial intelligence in ecology require preliminary data preparation. The article presents the results of designing a software system for environmental monitoring based on machine vision sensor data, which provides data unification while being flexible both in terms of data sources and methods of their analysis. Aim . Creation of a generalized software system for coordinated spatial marking of heterogeneous machine vision data for environmental monitoring tasks. Materials and methods . Software engineering methods, database theory methods, spatial markup methods, image processing methods. Results . A generalized method for unifying data was developed. The method is based on the analysis of existing open data from remote sensing of the Earth, as well as UAV aerial photography and approaches to environmental monitoring. To implement the method, a flexible architecture of the software system was designed, and a data model for a document-oriented DBMS was developed, which allows storing data and scaling the data analysis procedure. Conclusion . The existing sources of data and tools for environmental monitoring were analyzed. A generalized method for unifying machine vision data, an architecture, and a data model was created. The method, architecture, and model were successfully implemented as a software system with a web interface