### Author Topic: How efficient is a Color Sensor ?  (Read 1107 times)

0 Members and 1 Guest are viewing this topic.

#### John Lee

• Beginner
• Posts: 3
##### How efficient is a Color Sensor ?
« on: November 01, 2011, 09:28:02 AM »
Hello all, first time here

I am currently constructing a robot that will move around and avoid obstacles as well as identify other objects via an LED. The robot used is iRobot Create
.
Im using an ultra sonic sensor to measure distance, and i want to avoid using a visual camera (to cut on costs) so im planning to use a color sensor to detect what is infront of it. The lightening condition is fair, it wont be completely dark.

The obstacles and LEDs will have specific colors. The issue is I want the robot to detect the objects within a distance of 1m. Obstacles shall be detected at a distance of <0.5m

With further research I managed to find that color sensors respond to 50-150mm ??! Is this true ?

Does that mean I will NOT be able to accomplish my objectives while using a color sensor ?

I would appreciate replies based on expertise in this field please ! Appreciated
« Last Edit: November 01, 2011, 09:41:02 AM by John Lee »

#### Soeren

• Supreme Robot
• Posts: 4,672
##### Re: How efficient is a Color Sensor ?
« Reply #1 on: November 01, 2011, 04:57:51 PM »
The issue is I want the robot to detect the objects within a distance of 1m. Obstacles shall be detected at a distance of <0.5m

With further research I managed to find that color sensors respond to 50-150mm ??! Is this true ?

Does that mean I will NOT be able to accomplish my objectives while using a color sensor ?
What has this got to do with efficiency?

Any light sensor will have a specified sensitivity and you can detect  whatever is above that level - nobody can claim a certain distance, unless both the emitter and the detector is 100% specified.

However, except for LASER, which at least in theory is coherent, any light emitter and sensor will have an emission angle of >0°, which means that the farther the distance, the less a percentage of the radiated light will hit the detector (square law).

The demands on the receiving circuit rises with increased distance and circuit noise needs to be dealt with - do that, and you can receive down to a bit over the noise floor of the sensor. Cooling the sensor will keep the noise down and oversampling can dig out signals that is otherwise drowned in (random) noise.

I've seen an IR link function at more than 1km between emitter and detector, with optics and aligned towards each other.

If you want to "see" eg. a circle of 5cm in diameter, at a distance of 1.5m, without interference from objects farther away, you need the beam of detection to be around 13°.
This will give you the most signal, but will also mean that you need to scan a lot to be sure you covered the entire area in front of the robot.

Distances as short as you mention is mostly a question about how much time and how much cash you're willing to spend, but I don't think you'd save much (if anything at all) by avoiding a camera, as they're extremely cheap nowadays.
Regards,
Søren

A rather fast and fairly heavy robot with quite large wheels needs what? A lot of power?
Engineering is based on numbers - not adjectives

• Beginner
• Posts: 3