Open-source Machine Vision

Statement
All the topic articles of robotsky are exclusively published on the Internet. The article page shows that the materials of "Robot Sky Original" are original files of this site, make sure that the source is "www.robotsky.com" or "Robot sky ". If you have any questions, please contact us. Thank you for your cooperation!

Keywords
Machine Vision, visual recognition, visual recognition module, Object Recognition Based on Color Features, open source code (Open Source ).

Topic navigation
Robotsky's open-source machine vision topic is currently the most comprehensive Chinese column on the Internet about open-source machine vision materials. Like all other robotsky topics, open-source machine vision topics will be constantly updated and enriched. We welcome your suggestions and suggestions.
[Phase 1] provides the basic knowledge of robot visual recognition and four of the best open-source code visual recognition modules in the world. In the upcoming [Phase II], robotsky will provide you with a large number of excellent open-source code visual recognition software projects on the Internet.

Basic knowledge

Robot Visual Recognition Technology Overview
Abstract: This article introduces the basic concepts of robot visual recognition technology, especially for the "color-based object recognition" necessary in various robot competitions (such as RoboCup he FIRA ". All color-based object recognition is based on the computer's representation of color, which is what we call color space. For more information, see this article.

Open-source visual recognition module project (all projects on this site provide complete source code for local download)

FPGA-based visual recognition module

This powerful visual recognition module is developed by the Seattle robotics club in the United States. It can accurately identify and easily locate the target object. The author discloses all source files of hardware and software to share with you.

 

Cmucam3-visual recognition module developed by Carnegie Mellon University

This should be the most famous Machine Vision Recognition Project in the robot community, especially in the Robot Football community. This project has gone through the first and second generations, and is currently the third generation, that is, cmucam3. The major feature of the Third Generation cmucam is that it is open source code and greatly improves the computing capability and new functions.

 

Avrcam open source AVR Microcontroller visual recognition module

John R. Orlando, an American robot enthusiast, developed the visual recognition module based on the AVR Microcontroller. This module is fully open-source and provides comprehensive information for download. This avrcam uses a color-based object recognition algorithm with high efficiency.

 

Falcon open-source object recognition system (module)

The Falcon Object Recognition System can track up to 12 different objects at the same time, and has a definition that exceeds six times that of cmucam. Falcon is one of the most powerful visual systems.