FPGA Image Processing - MATLAB & Simulink

Target image processing algorithms to FPGA or ASIC hardware

Image and video processing, especially at higher resolutions, is compute-intensive. FPGA and ASIC hardware, which delivers higher performance per watt than software on a general-purpose CPU, can accelerate this process.

Such hardware must also be able to operate on a stream of pixels rather than a full frame at a time as in Image Processing Toolbox™ or Computer Vision Toolbox™. Therefore, image or video processing systems are often split across the FPGA fabric and the ARM^® processor of a Xilinx^® Zynq^® or Intel^® SoC.

Example FPGA image processing applications include:

Lane boundary detection in an ADAS system
Thermal imaging in an infrared camera
Corner detection in a surveillance system

For details, see Vision HDL Toolbox™. FPGA image processing and computer vision consulting services are also available.

Examples and How To

Vision Processing for FPGA - Video Series
Multiple-Pixel-Per-Clock Processing in Vision HDL Toolbox (3:02) - Video
Design Multiple-Pixel-Per-Clock FPGA Applications (5:15) - Video
What Is Vision HDL Toolbox? (1:43) - Video
Edge Detection and Image Overlay - Example
Gamma Correction - Example
Histogram Equalization - Example
Corner Detection - Example
Multi-Zone Metering - Example

Software Reference

Frame To Pixels and Pixels To Frame conversion - Blocks
Hardware-ready image and video processing algorithms - Documentation

See also: HDL Coder™, HDL Verifier™

Vision Processing for FPGA

Vision Processing for FPGA (5 Videos)

As computer vision algorithms become more complex, the transition from algorithm development to real-time implementation becomes critical. This presentation explores how to facilitate this transition.

Simplifying the Development of Computer Vision Systems