Video Capture and Compression on Embedded Devices

Industrial cameras are often used in conjunction with the application of machine learning. However, these cameras produce large bitrate and it needs to be reduced when processing video further away from the camera. This thesis tries to solve the problem by design of system suitable for grabbing video from industrial cameras, its compression and distribution to machine learning application. The thesis describes technologies applicable to the development of the multimedia application. For example frameworks FFmpeg and GStreamer were used implementation of the system