Algorithm Pre-processing Images of the Underground Surface to Reduce the Background Noise

The article presents an algorithm for pre-processing images of the map of the underlying surface to reduce background noise. The aim of the development is to create an algorithm for a fast and highly efficient system for processing noisy areas of the underlying surface map. Methods for averaging images using filters are considered. The object of the research is the methods of preprocessing and filtering of map images. Preprocessing is an important step for analyzing the map of the underlying surface in order to eliminate image defects that prevent obtaining complete information about objects, areas of interest taken from the UAVs. An image preprocessing algorithm based on a median filter using a deep learning neural network has been studied and its software implementation has been developed. The efficiency of the developed algorithm was evaluated using two quality criteria, such as the peak signal-to-noise ratio and the standard error criterion.