The Surround View Design Platform includes 4 camerasaption of image
Xilinx has delivered a trio of Spartan 6-based development kits to help create the next generation of automotive applications. These include the Surround View Design Platform (Fig. 1), the Hybrid Cluster Design Platform and the Infotainment Design Platform. The kits include not only an FPGA development board but also the auxiliary hardware such as four cameras for the Surround View kit. Software and FPGA IP necessary for the application area are also included. These allow developers to get started quickly for faster delivery of an end product.
The three automomtive areas that Xilinx has targeted are no simple examples. The Surround View application takes four cameras with fish eye lens and creates a composite, corrected, top down image in real time. This type of product is included on some high end cars now but Xilinx is providing this in a package that can be configured and augmented with features like blind spot detection. The Xylon 4 camera daughter board plugs into a PCI Express FPGA development board. The latter has video output so a complete demo system is easy to configure. The DSP framework is provided as a project along with the image analysis firmware.
The Hybrid Cluster and Infotainment Design Platforms are designed in a similar fashion. They include the necessary hardware and firmware to set up a demo system almost immediately. The design projects are included so the base system can be augmented.
The Hybrid Cluster design is based on a logiCraft-6 FPGA board from Xylon. It can generate a heads up display using minimal optics with image correction handled by the FPGA. The Infotainment Design Platform uses a Xylon logiCraft-CC FPGA board. It also comes with FPGA IP for multimedia presentation.
Xilinx is tailoring their kits for specific application areas like image processing and automotive applications. In most cases, the FPGA provides a more flexible development platform. The falling costs of FPGAs are also making them a useful deployment platform even in large production environments.