A Method of Comparison by Identifying the Curves of Images

The research work is devoted to the comparison of pictures taken at different times.As a comparison method, it is proposed to identify objects from images as curves and compare the parameters characteristic of curves.As a comparison parameter, the length of the curves and their curvatures at different points are chosen.Objects from pre-taken images are identified as curves by B-Spline model, selected comparison parameters are found and stored in txt files.The parameters detected from the captured images are then compared with the stored values.The algorithm mainly designed for the comparison of UAV images is simulated in the Python programming language, and an algorithm for the identification of winding roads as curves based on the images is developed.OpenCV, SciPy, NumPy libraries are used for this.The proposed algorithm can be used to identify various objects, parts, and curves from images.Although the UAV is mainly designed for navigation by comparison of images, it is possible to apply it to issues such as face recognition, area monitoring, and comparison of medical images.