Image recognition in unmanned aviation using modern programming languages

The paper considers the different methods of image recognition in unmanned aviation using modern programming languages. It is shown that the new era in aviation is characterized by new challenges and threats, as well as uncertainty, and it is not always possible to identify a threat through standard means of control. The authors summarize the various methodologies of analysis and justify the algorithm for recognition zones of video observation of possible icing of the surface of the aircraft. The tested methods, in general, were divided into three groups: the preliminary filtering and image preparation, the logical processing of the results of the filtering and—machine learning which in general are divided into three groups. The filtering that allows highlighting of images in the recognition area, linearization, the transformation of “Hafa” and filtering contours as a separate class of filters were selected as the main methods of filtering images. The authors propose to use a device that can determine possible areas of icing of aircraft using airborne meteorological radar. The problem is the ratio of the image, which was before the icing, and the changes in this image in the presence of ice.