MATLAB and Simulink for Embedded Vision

Design and deploy image processing and computer vision applications to embedded systems

What Is Embedded Vision?

Embedded vision involves the application of image processing and computer vision to embedded systems. Key components of the embedded vision development workflow include algorithm design, system modeling, collaboration, and deployment of vision algorithms. 

Engineers use MATLAB® and Simulink® to develop image processing and computer vision systems and deploy them to embedded target hardware. With MATLAB and Simulink you can:

Real-World Embedded Vision Applications

Learn how MATLAB and Simulink users have developed and deployed real-world embedded vision systems

Automated Driving

Continental uses MATLAB to: automate learning different traffic sign types, access databases, generate synthetic sign samples, generate code, and monitor and evaluate classifier training using interactive apps.

Robotics

Clearpath Robotics engineers use MATLAB to prototype algorithms and to analyze and visualize data for industrial robotics research and development.

Medical Imaging

Infraredx uses MATLAB and Simulink to accelerate FPGA development for intravascular imaging systems.

Develop Algorithms and Model Systems

Design algorithms and system models for embedded vision systems using MATLAB and Simulink tools, which provide reference-standard functions and blocks. Automate common workflow steps with apps for acquiring live image and video data from cameras and other sensors as well as apps for processing, analyzing, simulating, and visualizing that data.

Incorporate Third-Party Software in a Collaborative Workflow

Incorporate third-party software tools, libraries, frameworks, and languages like Python®, OpenCV, and TensorFlow™ into your MATLAB and Simulink based workflow to support collaboration, integration with existing projects, and reusability of code.

Generating Code for Target Hardware Platforms

Targeting CPUs

Use MATLAB Coder™ to generate C and C++ code for vision algorithms developed in MATLAB. Integrate optimized libraries such as the ARM® Compute Library for ARM architectures and MKL-DNN library for Intel® CPUs.

Code Generation Targeting GPUs

Use GPU Coder™ to achieve optimized generation of CUDA code from MATLAB that can be used for prototyping on GPU-based hardware platforms including NVIDIA® Jetson™ and DRIVE™.

Targeting FPGAs and ASICs

Use HDL Coder™ to generate Verilog and VHDL code from vision algorithms that you design using Simulink and Vision HDL Toolbox for FPGA- and ASIC-based platforms.

Testing and Verification

Perform rapid prototyping, processor-in-the-loop (PIL) simulations, and hardware-in-the-loop (HIL) simulations with HDL Verifier™, Simulink Real-Time™, Embedded Coder®, and Simulink Desktop Real-Time™ to efficiently test and verify your generated code.

Connecting to Embedded Hardware and Deploying

Choose from a variety of hardware support packages for popular embedded hardware to jumpstart receiving and sending real-world data between MATLAB and Simulink, and automatically generate executables from your algorithms to run on embedded hardware platforms.