Close
0%
0%

Low-Cost Robot Vision Module

Add artificial vision to any robot in a simple and economical way, with this module. Without limits, you can add it to any platform

Similar projects worth following
We love modifying robot toys, adding new functionalities: that is how this project was born. Thus, we integrate computer vision in a worthy way in all our toys without unnecessary budgets and with a wireless philosophy ‘NO WIRES, NO LIMITS, JUST DISCOVER’.

And that is how we created Low-Cost Robot Vision Module, a module which summarizes the system that processes the images taken by the robot and, at the same time, allows the communication with him.

Nowadays, the modules take the images in a digital way thanks to small cameras or webcams, which means that every robot must have a tiny computer in order to process the images; the costs are incremented, or the robot must be connected with a wire to an external computer. In conclusion, every robot needs both a computer and a digital camera. 

On the other hand, we work with a different technology: the image taken by the robot is sent to our module with no wires (wireless!!!) and after that, the module quickly digitalizes with nothing more than a few tenth of latency. Furthermore, this system allows the user to control 16 robots at the same time because of the fact that the computer can switch automatically the channel of processing, selecting in every moment the signal we want to work with. 

Give All the Robots Eyes to See!

The costs of the implementation are less than 20$, which can be translated into an efficient and accessible system. Adding computer vision is now flexible and simple.

It is the best way to add a camera to any robot

It is possible to configure the power of the emitter: if it is correctly made, we would be able to control our robots more than 100 m away in open environments.

This is how the system works: throughout the microcontroller, the module is synchronized with the PC in the desired channel. Once we collect the image, the Analog-to-Digital block digitalizes the image, like a webcam. At the same time, the USB-UART module allows the bidirectional communication between robots. 

We can process the image from the robot with any program because, in the end, the device works like a webcam. In this case, we use Processing and a library from OCV. The idea is working with the examples of the library, recognizing faces quickly. Nowadays this system has been built in our robot Wall-e, tracking faces with no difficulties.

Without change there is no improvement, this is the evolution of our idea


We are working on the latest model with the following features:

    -AIO system module

    -Compatible with any UART communication module

    -On-screen control

    -Control of channels by UART communication

    -Capacity of control up to 16 robots

    -USB DC supply

    -Compatible with any operating system

To conclude, the way these device works allows the user to add computered vision in a simple way with no needs of expensive or powerful components, making possible the idea of artificial vision in every robot!.  

tracking_windows_v0.zip

Processing sketch for windows

x-zip-compressed - 1.19 MB - 05/30/2018 at 13:01

Download

tracking_linux_v0.zip

Processing sketch for Linux

x-zip-compressed - 2.16 kB - 05/30/2018 at 13:01

Download

  • 2 × HC-05 Bluetooth module to communicate between PC and robot
  • 1 × Analog Camera Capture the images on the robot
  • 1 × Video transmitter Tx video 5,8Ghz
  • 1 × Receiver and video capture Like rotg-01, or it can also be worth a separate receiver and video recorder.
  • 1 × TTL/USB You need it to connect the HC-05 to the computer

  • Design for last prototype

    Innovart05/31/2018 at 13:12 0 comments

    In the last prototype of this module, we include all the functionalities in a single device. We need a USB HUB that allows us to connect all the functions at the same time. We have researched about it and created some pcb to perform tests

  • Not only robots..

    Innovart05/31/2018 at 12:10 0 comments

    With this system you can process the image of any analog video transmitter. This means that you can also collect information from drones that incorporate small cameras. Imagine how awesome this can be. This is a small example of a drone flying with our partner Sara, the image sends it to the computer and recognizes its face. This happens very fast, but you have to improve the processing sketch.

  • Prototype 2.0

    Innovart05/30/2018 at 13:11 0 comments

    We are working to create the latest AIO model "All In One" artificial vision module, which will also allow the selection of video channel from the PC. In this way we can control several robots at the same time by switching the video signal from the PC.

    The prototype of the new module, still in tests incorporates all the components of the scheme.

  • New Version

    Innovart05/30/2018 at 12:56 0 comments

    With the demand of the drones, there are new video capturers that include the capturing and at the same time the receiver, this allows us to receive the images of the robot in a simpler way. This system has almost no latency and works without problems

    We modified our main code to be able to reconcile the camera in windows, before only the camera worked in linux. We are very happy with the tests since it is a very cheap way to add a camera to any robot.

    As you can see in the last video, we use a bluetooth bridge formed by one (PL2303 and HC-05).

  • Example Processing

    Innovart05/30/2018 at 12:38 0 comments

    You can process the image with the program you prefer, in our case we prepare a simple example that allows us to move our little robot to follow the faces. Processing recognizes faces with the OCV library. When it detects a face, it draws a blue square on it. This program is very simple, the only thing it does is compare the position of the square with respect to the window. In this way the robot can know if it has to move to the left or to the right. With the size of the square the program is also able to know if the face is far or near, in this way is able to follow the faces.

    On the other hand, the robot must have a microcontroller as Arduino, which allows it to receive orders from our computer (Brain) and be able to execute them in our robot (Slave).

  • Vision on Walle

    Innovart05/30/2018 at 12:02 0 comments

    We hacked our new toy the "walle-e" robot, thanks to the small size of the analog cameras we could place one in his right eye without much complication. The video signal travels through the black wires that connect the eyes with the body. Inside the body we have a small 200mw video transmitter that sends us the image in real time. Inside the robot we place a bluetooth module that allows us to control the robot remotely. Any analog camera can be used. We use a very small camera to hide it in the eye.

    In the following link you can find the camera that we use in this project:

    https://www.banggood.com/600TVL-8_0MP-14-2_8mm-CMOS-FPV-170-Degree-Wide-Anlge-Lens-Camera-PALNTSC-p-984345.html?rmmds=myorder&cur_warehouse=CN



  • First Prototype

    Innovart05/30/2018 at 11:54 0 comments

    In the beginning we set up the system in a fast way, we look for a box where we can introduce video capture, video receiver, usb hub and a communication module. This system worked well for a long time and we were able to test the idea. We needed a 12v transformer to power the fpv receiver, we also need to select the channel each time we turn on our box. 

                                                        SIMPLE, UGLY BUT FUNCTIONAL


View all 7 project logs

Enjoy this project?

Share

Discussions

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates