Shop OBEX P1 Docs P2 Docs Learn Events
Sensor following servo — Parallax Forums

Sensor following servo

saxosaxo Posts: 2
edited 2014-05-07 10:58 in General Discussion
Hi there, I am new to this forum and I hope I am posting in the correct section. I have been searching the net for something similar to the Pixy CMU, I want to attach a camera to a tilt and pan mechanism but it should follow a wireless sensor. Like the Pixy tracking an object this item should track and follow a sensor that I will attach on a moving object. Can any of you please help me with this item, to build one or where I can buy one.. Thanks in advance!!!

Comments

  • Martin_HMartin_H Posts: 4,051
    edited 2014-05-07 05:37
    If the object is close to the tilt/pan mechanism you can do it with IR sensors. Here's a project thread where I built one using directions from "Robot Builder's Bonanza" http://forums.parallax.com/showthread.php/133845-IR-Compound-Eye-from-RBB-4-and-Let-s-Make-Robots-(Video)

    If the object is far from the camera then machine vision would probably be the way to go.

    With regard to the wireless feature. What sort of range do you need? Bluetooth is cheap but short range, so are Pololu wixels.

    With regards to the tilt/pan mechanism. How heavy a camera will you be using? If it's a lightweight web cam, or something like the Pixy, then a servo based tilt/pan should be fine.
  • saxosaxo Posts: 2
    edited 2014-05-07 08:27
    I want to connect a small video camera like the Go Pro. The purpose of this item it to make a race video while drag racing, so the camera will automatically track the other car, max distance should be about 200m apart.
  • Martin_HMartin_H Posts: 4,051
    edited 2014-05-07 10:58
    saxo wrote: »
    I want to connect a small video camera like the Go Pro. The purpose of this item it to make a race video while drag racing, so the camera will automatically track the other car, max distance should be about 200m apart.

    Those are pretty light, so a servo tilt/pan would probably be fine. But the task you want to apply machine vision to is probably beyond the Pixy. I have the previous CMU cam generation and they work best when they are tracking an object that is different from its surroundings. Think a red ball against a black background. It's easy for the machine to locate it because the signal to noise ratio is so high. A car in a race environment probably wouldn't exhibit such good properties.

    You would probably need something better, like a RaspberryPi, a webcam, and much better machine vision software. I haven't used anything like that so it is beyond my ability to suggest something.
Sign In or Register to comment.