Dependence Comparison of the Effectivness of Software Tools for Splitting Video into Frames on Format and Resolution Using a Road Survey as an Example
Alexander A. Zhuravlev, Konstantin A. Aksyonov · 2024
The importance of roads in the life of almost every modern person is quite huge. One of the key characteristics of the road surface is its quality. To assess the quality of roads, there are many systems, most of which work with a stream of video information. This problem is solved mainly by analyzing images that are obtained from the video stream. To split video into frames, there are special software tools with varying effectiveness. In this research, the average time to retrieve one frame from a video is selected as the main performance indicator. FFMpeg, MoviePy and OpenCV are selected as analyzed programs for splitting videos into frames. The study will help determine the most effective software tool and format for postprocessing (splitting into frames) videos received from any source: a phone camera, drone, etc. The research material is a minute-long video containing a road surface (any other video can be used, it is just example) with a frequency of 25 frames/sec. in 4 formats (avi, mov, mp4, wmv) and 5 resolutions of 16:9 (240p, 320p, 480p, 720p, 1080p) – a total of 20 different videos for each of the selected software tools (total number of experiments are 60). All selected programs have linear dependence on the resolution for each format. However, FFMpeg software tool and wmv format have the lowest absolute time values; FFMpeg also has a significantly better effectiveness compared to MoviePy and OpenCV.