I know this isnt strictly openhab related but it is something I want to implement in openhab.
I grow fruits: tomatoes, apples, pumpkins and rootcrops: potatoes, carrots etc.
What I like to do is make a sensor that keeps track of the growth.
For a tomato for instance I was thinking about a small device with two light arms on the bottom and top of the tomato that get pushed apart more when the tomato grows.
That keeps me informed about the growth of the tomato but also its health: if the size would suddenly decrease that is a sign of not enough water or rot.
Anyway what I like some input on is: What type of sensor could I use. I was thinking about maybe a potentiometer connected to the two āarmsā or maybe a flex sensor, but if anybody has a better idea I would be very happy to hear that.
Just to get the juices flowing I enclose a picture of a commercial sensor. Sure I could get that one but they come as total (expensive) subscription system for large farms and not justifiable for my small needs.
The mechanics is not the problem, but Iād like some thoughts on the sensory element.
Thanks a bundle
You could take a look at a linear displacement sensor such as this one . Iāve not looked too hard at the specs - you want something with a low force or your plants will grow around the sensor ;).
@chris that is definitely a good idea. The stroke in that one is a bit small, but maybe I can construct something with a bigger stroke, could meke that of ferrite in a coil and then measure impedance. Hmm, something to think about
@christoph_wempe
Thank u, sending the info isnt really the problem, I use ESP8266 for that and I am up to date with MQTT. What i mean with āsensory elementā is the active āthingā that transforms the fruit expansion into a measurable unit of some sort, be it voltage, resistance, capacity or induction.
With regard to the pumpkins, indeed I have thought about a pressure sensor or maybe some extendable flex sensor around it.
I was thinking about the stem measurement for the 12mm version, but RS also have a 38mm version for the same price (Ā£1 more) and there are much larger versions, but they have much larger pricetags .
Interesting project though - Iād be interested to see what you doā¦
Actually I hadnt thought of stem measurement yet. Good idea. I have to give that some thought.
yes, seems that where sensors like this are concerned size does matter
It is a bit past growing season, so I have some time to work on this but if I come up with something, I will definitely give some feedback, I am sure i am not the only one into gardening here.
Yes you are right, I presume it is, but it just āhappened to be in the pictureā, I just had not given that one much thought as I was focussing on the fruit.
But it is definitely also a growth indicator that is on my radar now. Seems much easier to construct. I presume there is some tension/spring mechanism in it to hold it on the stem. I presume they turned it upside down for the picture as it probably would be more stable just hanging on the stem.
Obviously I would not put this on every plant, I would just pick a āspecimenā plant in the bed to collect the data on.
My tomatoes are/will be under a āroofā so rain doesnt need to be a problem for the sensorā¦ or for the tomatoes (blight nad stuff)
I think I will order that sensor (Oops, out of order)
Well, thinking outside the box in terms of the solution already presented, I would probably use a number of cameras arranged around the plant and use something like OpenCV to identify and measurer the size of the fruit.
If you have overlapping views, you could even model the fruit in 3D and get a pretty accurate volume. You could even use color to determine ripeness.
Very true Rich, I had been thinking along that path as well, especially what the color/ripeness concerns. Maybe a good moment to start digging into using a camera on the RPi. Have not tried that before.
Opens a lot of new possibilities, also general health monitoring from foliage.
I had not thought of openCV, but that seems pretty powerful, would need to find a way to let the camera stay on level with the fruit. Hmmā¦ getting bigger than I thoughtā¦ but interesting
Looking at my sitemap/items looks like the garden is getting a bigger and bigger part of that
There is no question that modern ML could be helpful GIVEN ENOUGH DATA,
BUT if your real objective is to grow good fruits at a high yield and you do not have thousands of plants (some of which you are willing to sacrifice āfor the dataā) you COULD do single camera views of plants-of-interest (pretty expensive on a per-plant basis),
OR even betterāwalk thru your gardens every day or so with an open eye. Good for the fruit, good for the soul.
Thanks Bob,
I do walk through my garden every day, check the plants, feed the chickens and quail, check out the fish, but what I am trying to do is conceptual to see if keeping track of certain parameters can increase yield and then expand it to a larger scale
That makes sense. I grew up on a farm and my father was pretty obsessive about measures of things. Nothing like modern GPS driven micro-agricultural stuff, but I get that. I like @rlkoshak 's idea with the camera and openCV. Calculate the volume of each tomato in your sample every day and get a chromatic measure of the color. Plot the volume and chromatic value over time. See how the metrics compare with your human observation and the ultimate taste test.
Indeed. No worries, I really get my pleasure out of the hands on work in the garden (and eating fresh produce), but I didnt grow up on a farm (but fairly rural) and I am too old to build up a lot of practical experience so, I just wanna know if what I do translates into ābetterā
Experienced farmers maybe would put their finger in the ground to know it was time to plant the carrotsā¦ I stick in a DS18B20
I would experiment with the IR time-of-flight distance sensors (āLASERā distance sensors). Theyāre cheap enough that you could have one per plant, and their sense distance and resolution seem perfect for the task.
or maybe some extendable flex sensor around it. 