autonomous golf ball collector

hi im currently invloved with my final design for my Bsc Elec eng degree, in South Africa. my design is a autonomous golf ball collecting robt, could any one advise me as to how i could get this accomplished, bearing in mind that, I want to do all the detection and processing of info on the microcontroller on board the robot, thanks, much appreciated

What are you asking? Do you want us to do your entire design for you? Why don't you just get a HUGE vacuum cleaner and have the robot run around randomly sucking up golf balls.

Really, people here are more than happy to help, but when you come out with a question showing that you've exerted just about zero effort, we are not inclined to offer anything.

The vacuum cleaner with golf balls was always a test my buddies would suggest, whenever I was swinging. But in Africa they may need a snake discriminator.

http://www.gizmag.com/golf-ball-wrangler-harvest-balls-from-water-traps/18841/

**broken link removed**

Robot can haul this like roller.

so barry, I was not very specific, I have idea's and have done research to some aspects of my design, but the actual ball detection is a problem, I want to use image processing on the robot MCU itself.

- - - Updated - - -

can you guide me as to how i can achieve this.

How will be image processing hidden balls which cant be seen or seen partialy ?

What will be time needed to collect lots of balls from big field, one by one ?

SunnySkyguy nicely mention snake separator I dont want even to imagine little creatures such as Gabonica Viper inside ball boxes and someone how to trying to reach some ball from boxes.


See this project :

http://coecsl.ece.illinois.edu/ge423/spring09/Group4_Sandra_Eeshan_Chris_Mike/index.html

I just came back from golfing, somtimes you have to be stepping on them in tall grass to feel them. Seeing them alone may be an issue. The robotic field scanner with pull cart disc rake seems to make more sense. Like pattern plowing of farmers fields using GPS.

I would imagine you want to take a 2D image of the area ahead. Look for light-colored pixels. For such a method to work, the robot's eye must be several inches above the ground, and the grass must be short.

The robot moves toward the light-colored pixels.

It might help to use two eyes, for stereoptic distance judging.

I can't advise how to connect a microcontroller to a photodetector array, nor to read it.

A simpler hardware hookup would involve a single vertical line of photodetectors, enough to take in a 10 degree vertical field of view... the robot scans around, looking for a light-colored pixel or two, between green ones? It would complicate the search routine, but it would simplify the optical hardware hookup.

To figure out how much resolution you need in a photodetector array, take some test pictures of a lawn with a digital camera held several inches above ground. Examine a frame extending 10 degrees above horizon to 10 degrees below. See how many pixels need to occupy the frame.