I have been working on a project that detects the presence of plants using an Arduino and other stuff. I’m not all that far along. My current dilemma Is finding the right sensor to detect whether a plant or dirt is under the sensor. It would be extremely helpful for this sensor to have a fairly large detection area. I was thinking about connecting a camera to the Arduino to detect the green of chlorophyll (which would work for my purpose since I will not be using this in the fall or when a plant is not green, such as fall)
My questions are: What is a good camera with a decent range and resolution that could be used with an Arduino? How would I get this camera working? Or is a camera the best option for this plant detection purpose? If not, what is?
Lots of green things are not plants and many plants are not green. Dead ones often have the same colors as dirt, and dirt often has lots of plant material in it.
You need to define more clearly what you mean by “detects the presence of plants”.
“detect whether or not a seedling (that is guaranteed green) is sprouting”
These plants are guaranteed to be in a field or other environment where the only background colors are brown dirt and potentially some yellow corn or bean trash from last year in the instance of the field.
These images will be taken some distance from the plants in order to get a larger swath of detection.
The whole point of this project is to see whether or not a certain field or patch of seeded earth has sprouted yet, and give a general idea as to how many, although exact counts are unnecessary. relative amounts will do.
If you are committed to analysis of images taken from some distance away, that will require a high quality, high resolution color camera, and probably also the use of carefully selected bandpass filters in front of the lens. The Arduino is completely out of the question for such a project, due to limited memory and low speed.
The reliable detection of just about anything in the real world is often a very difficult task.
You keep talking about seedlings and guaranteeing that they are always green against a non-green background.
Rephrase your question to “detect a green object” and see if all your assumptions still hold. Also see if with that restatement, any other options appear viable such as, I dunno… shining a white light downward and measuring how much green light is reflected?
As far as a camera goes, you can make an extremely high resolution camera with a single pixel sensor: you just have to focus it on the minimum resolvable area and move it over the scan area in very small increments. The problem now shifts from one of memory and processing power to motion control. It’s up to you to decide which is an easier problem to solve.