This paper presents a comparative analysis of the effectiveness of object detection algorithms within the framework of a simulation system for modelling unmanned aerial vehicles (UAVs) visual guidance. Specific challenges associated with flying object detection, such as variability in target size and shape, dynamic changes in background, lighting, and weather conditions, as well as possible object overlap with other scene elements, are considered. Three variations of the modified detection algorithm based on binary image contour extraction are described. The performance of the algorithms is tested on a set of video data with different lighting conditions, target parameters and their total number. The experiments conducted show that the algorithms using preprocessing show better target detection accuracy results. The best performance was obtained with the algorithm using adaptive binarization. However, the disadvantage of increasing the number of false positives was found.