Silicon Software: Twenty Years in Industrial Vision
Silicon Software is celebrating its 20th anniversary in 2017. In September 1997, Dr. Klaus-Henning Noffz and Dr. Ralf Lay the company as a spin-off of the University of Mannheim. From the beginning, both managing directors used framegrabbers as flexible image capture cards combined with FPGA technology and their programming, for which they subsequently built up the graphic development environment VisualApplets. Intelligent solutions for complex applications with real-time image processing in all industries to create, this is the motto of the company to date.
Efficient image processing solutions were the focus right from the start to meet growing customer requirements with high data throughput, low latencies, long distances and effective relief of the host PC. As a result, powerful image preprocessing and signal processing functions as well as I / O control were already integrated into the standard scope of the frame grabber series, which support all important high-speed camera interfaces. FPGAs (Field Programmable Gate Array) have been a viable alternative for small and medium-sized businesses. An additional built-in frame grabber series allows image processing in difficult environments such as the production environment. The entire framegrabber segment recorded high growth rates over the last 20 years.
Standardized, customizable customer solutions
For FPGA programming, the company has now created the graphic development environment VisualApplets with over 250 operators. Complex image processing applications can be implemented much faster with the software, which is also aimed at software developers and application engineers who do not have a hardware programmer experience, than with HDL-based hardware development. The software was initially used to program its own frame grabbers and is now licensed to manufacturers of FPGA components such as (embedded) cameras and vision sensors. Baumer Optronic and SICK AG have opened entire product lines for programming with VisualApplets. They realized this with the extension VisualApplets Embedder, by defining the interfaces of the IP core of their devices once and integrating them into the FPGA overall design. Once developed algorithms and applications were then transferable in a few steps to different hardware designs of other devices. With the addition of VisualApplets Expert, experienced users are also able to integrate their own HDL code and existing libraries as generic applets, thus accelerating the market maturity of their products.
Machine Learning and CNN
As an important future subject, Machine Learning and CNN (Convolutional Neural Networks) have developed based on deep learning, which are particularly suitable for detecting defects with higher probability and better quality, for example in surface inspection. In the future, CNN will be integrated into the software as an additional operator, which as an open system flexibly incorporates new technologies.
"Silicon Software is changing on many levels for continuity: the company and its location have remained unchanged over the past 20 years, and existing product lines have been steadily developed and supplemented by new ones. Our commitment to associations and standardization bodies in which we are leading new standards for machine vision and embedded vision are also a constant of our company, "emphasizes Dr. Klaus-Henning Noffz, Managing Director of Silicon Software. And Managing Director Dr. Ralf Lay adds: "Standards like OPC UA Vision and embedded GenICam have a positive effect on our new hardware and software solutions. The company has always strengthened the balance between industrial and non-industrial applications, OEM customers and customers of standard products as well as the software product line and services as important standstocks. "
