Image editing software
Image processing on embedded hardware
Machine vision software provider MVTec Software presented the latest developments in its embedded vision portfolio at Embedded World: Halcon and Merlic for embedded vision as well as Halcon 17.12, the new version of the standard software for machine vision. The latter brings powerful deep learning functions to vision systems. Users can train their own neural networks for the classification of image data, using pre-trained networks as a basis. Experts will also demonstrate the use of complex deep learning algorithms on an Nvidia Jetson TX2 embedded board. The demo proves that embedded hardware is also suitable for complicated machine vision tasks.
In another demo, four embedded boards ran machine vision applications with Halcon. A linear board moves a box with barcodes, QR codes and OCR applications. Each board is connected to a camera and solves a task. Screens show the results and demonstrate the performance. Another demonstration showed how Merlic runs on an Adlink Smart Camera and uses deep learning-based OCR to accurately recognize a wide variety of fonts on packaging in fractions of a second. Visitors also learned how Merlic can be used to seamlessly integrate a PLC into vision systems and visualize the remote front end on a mobile platform. am
Embedded World, Hall 4, Stand 203








