Conferences and supporting programme
Embedded Computer Vision Applications with Qt
This presentation showcases how to build complex HMIs for computer vision applications using Qt. Computer vision applications often feature several video streams, video manipulation, overlays, as well as interactive HMI elements which makes them non-trivial to implement, especially with the limited resources of an embedded device. Qt is a popular C++ framework used for all kinds of embedded applications. With its Qt Quick graphics stack it allows to build beautiful, fluid, animated HMIs, utilizing the GPU for accelerated rendering. Based on our real-world experience we demonstrate how to apply the Qt Quick architecture to the challenges of computer vision applications. Qt, which is focused on visualization, makes it easy to integrate with analytical framework such as OpenCV. For simple applications, which only need to place markers such as rectangles, cross hairs or points on top of the video stream, Qt provides everything that is needed out of the box. After an architecture walk through of such an application, we have a look at the implementation of a few more complex use cases. Here we show how to output feature points, visualize the result of line detection, highlight areas of interest or even generate point clouds via custom Qt scene graph elements which are rendered using the GPU.
--- Date: 26.02.2019 Time: 4:00 PM - 4:30 PM Location: Conference Counter NCC Ost