High-Quality, ROS Compatible Video Encoding and Decoding for High-Definition Datasets

In order for the compressed Datasets are an important tool for research and development dataset to be useful as a reference dataset for SLAM research in mobile robotics. For example, for evaluating Simultaneous and benchmarking, a truly excellent video quality has to be Localization and Mapping (SLAM) algorithms [1], maintained. As reported in [8], [9], compression artefacts can datasets play an important role [2], [3]. Sensor technology in fact reduce the applicability of video streams to various is advancing rapidly, such that better camera sensors with applications such as object detection. Specifically, we do higher resolutions and higher frame rates are available for not need to encode the videos in real-time, but will do use in robotic systems. While older datasets like KITTI from the compression in post-processing, because there are not 2013 [4] use 1.4MP cameras with a frame rate of 10Hz, later enough computation resources even on a 16-node cluster to datasets offer for example 5MP images with 10 Hz [5].