Hardware and Software Concepts for Image Analysis
AI-based image analysis
FPGAs allow image data to be processed directly on a frame grabber or embedded vision device, from image acquisition to analysis results. This is done without loading the CPU, which is important for compute-intensive applications using CNNs. This allows smaller PCs to be used without a GPU, reducing the cost of the overall system.
AI with a Basler frame grabber and VisualApplets
Graphical FPGA programming with VisualApplets can be used to integrate suitable network architectures of varying size and complexity. Network and parameter information can also be imported from third-party software and training tools such as TensorFlow.
Our range of services include:
Implementation: If you have already created and trained a mesh, we transfer and implement it together with the weights (parameters) on the frame grabber's FPGA.
Full development: We take care of the setup and training of the network as well as the entire FPGA implementation.
AI applications in embedded environments
The most compact and cost-effective AI vision systems are developed using embedded technology: Smart edge devices offer fast runtimes, low latency, and enhanced privacy and security.
Our range of services include:
Development of neural networks with different architectures, e.g. U-Net architecture, MobileNet backbone.
Adaptation and training of existing neural networks
Integration and command of AI accelerators
Integration with target systems
Classic Image Analysis
With Basler pylon vTools, we offer tools that are already integrated into the system software, allowing you to build a complete image processing system yourself. pylon vTools are available for Windows and Linux / Linux ARM.
pylon vTools for Image Processing
Use pylon vTools to create powerful image processing functions and flexibly integrate them into your own application. Integrated with camera control and image acquisition, they are perfectly matched to Basler cameras.
To the pylon vTools
