Conferences and supporting programme
Bridging Embedded Vision and Machine Vision with a Hybrid Driver
Driven by rising embedded processor power that are both cost effective and energy efficient, it is nowadays possible to execute advanced image processing on embedded systems. Embedded designers can choose from various sensor modules , embedded boards and software frameworks to build new vision applications. While the latter two are certainly more in focus, the sensor module, its image quality and possibilities to pre-process undoubtedly must be considered for the success of a vision application as well. To be easily integrated into an embedded system, sensor modules e.g. with MIPI CSI-2 interface should support Video4Linux2 (V4L2) for image transmission. This also enables directly the use of Gstreamer or OpenCV for image processing. On the other side, the machine vision market established a set of standards, called GenICam. GenICam represents a generic software interface for camera manufacturers and third-party image processing libraries. The possibilities for the customer controlling and setting up the camera exceed the capabilities of the V4L2 standard control set by far. We have developed a CSI-2 hybrid driver to provide easy control of our camera module to embedded customers and enable machine vision customers to re-use their code. The driver supports two modes of operation. The first one provides a V4L2 user interface and access to common functionalities with lean register set and the second one makes use of the more advanced GenICam approach with an extended feature set.
--- Date: 26.02.2019 Time: 5:00 PM - 5:30 PM Location: Conference Counter NCC Ost