Conferences and supporting programme
Embedded Vision Systems Used as Sensors in IoT Applications
The combination of cameras and processing boards can be used in several professional B2B applications to collect data that is generated out of the pictures that were captured with the camera. In the IoT environment sometimes it is not useful to stream videos and analyze it in the cloud, because of the huge amount of data that is generated and the high bandwidth resources that are needed. Sometimes also real time reactions are needed and this means that the data processing has to be done on the edge. In retail for example the companies want to know who is in the shop or who likes the advertisement and need only the age, gender or time of attention of the customer. Therefore special algorithms are used on the processing board to create the data. No image has to be transmitted or stored. The “IoT camera” works as a sensor and sends only data to the cloud. In factory automation an optical inspection can be done by embedded vision systems and these systems can be trained with neural networks to improve the results or to reduce the setup time of a system. In a special case the visual inspection system should detect defect parts or inspect assembled components. Therefore the system has to be trained and pictures are transmitted to a cloud where people analyze the pictures. After the system was trained, it can autonomously improve the results by using machine learning algorithms. For monitoring the consumption of energy smart meters are used today, but how can old displays be used as well? Embedded vision systems can be used for retro-fitting of old meters. Here a picture of the meter is taken and the software transfers the number into data that can be sent in a digital way to the cloud. In all cases the software is the relevant tool to get the data out of the pictures that are taken by a camera. This data can be used for analytics and even to start actions or stop machines. The embedded vision system is not used to capture or stream videos it only get the needed data and information out of the images. This presentation will show, what is possible with embedded vision systems in the IoT environment and where are potentials for these systems in the future.
--- Date: 28.02.2018 Time: 2:30 PM - 3:00 PM Location: Conference Counter NCC Ost