Ecological Systems and Devices Annotation << Back ON THE ISSUE OF USING A NEURAL NETWORK ALGORITHM TO MONITOR THE STATE OF THE URBAN ENVIRONMENT R.A. D’yachenko, D.A. Gura, D.A. Bespyatchuk, S.V. Samarin, D.A. Drazheckij Currently, the issue of the ecological state of urban areas is quite acute and relevant, since the rapid increase in the urban population over the past 100 years has led to a negative ecological state of many modern cities. In this scientific work, an algorithm has been created to combat one of the main sources of pollution of urbanized territories – unauthorized solid waste storage sites. The developed method is based on fully-convoluted artificial neural networks, since, firstly, they are capable of learning, and secondly, they can be used for pattern recognition. The input data for the neural network developed in the presented article were photographic images obtained from unmanned aerial vehicles. The created neural network is based on the "Faster-R-CNN" architecture. The main advantage of this architecture is that neural networks built on "Faster-R-CNN" can quickly and efficiently search for objects in photo images, while not requiring large computing resources from a computer. As a result, the neural net-work detected solid household waste in photographic images and the accuracy of gar-bage detection was 75 %. The conclusion of the article notes that it is necessary to conduct further research on this topic to improve the quality of the algorithm. Keywords: Neural networks, urbanization, urban environment, solid household waste, photo images, SDGs. DOI: 10.25791/esip.8.2023.1392 Pp. 42-50.