Good day to all, I am currently looking for information regarding a sensor capable of collecting Nitrogen, Potassium and Phosphorus. I am currently looking into creating a portable system that would collect this specific data and use the data to give recommendations to improve the soil nutrient content for better crops.
Thank you! 
Have you read this ?
Good Day! Thank you for replying to my inquiry.
Sadly, I have read those threads before starting this thread. The threads did not have the information
I was looking for. 
Nitrogen, phosphorus and potassium, sounds like a job for a chemical assay. Soil nutrient testing is this?
Sensor I don't you'll find, but some automated lab equipment might exist to do this routinely - won't be cheap, probably not portable.
Nitrogen: you're probably looking to detect nitrate. Or also nitrogen in protein or other organic molecules?
Phosphorus has the same problem.
Potassium is relatively easy as it will normally come as K+ ion, so all you need is a K+ selective membrane and something that can detect the presence of these ions, probably by potential. This would be quite similar to a pH probe.
The most accurate way of detecting NPK in a sample is by taking a sample and analysing it in an element analyser. Even a simpler mass spectrometer won't work unless you know exactly what molecules are present in your sample.
So yes, you'll need a quite well equipped lab to accurately detect the different nutrient concentrations.
When it comes to soil and what nutrients it needs, it's the local farmers that know this exactly. Unfortunately nothing of this is written down, it's all in their heads, so if you lose even one generation it's all gone...
Sweet stuff, but note that this is for watery solutions as used in hydroponics, not for soil. And very expensive.
This is something I'm also really interested in. Take a look at this paper:
https://ieeexplore-ieee-org.ezproxy1.library.usyd.edu.au/document/8312001/?part=1
They used an arduino for sampling three optical transducers (leds with specific spetral output)
"An optical transducer is developed to measure and to detect the presence of Nitrogen (N), Phosphorus (P) and Potassium (K) of soil."
...
"The optical transducer is implemented as a detection sensor which consists of three LEDs as light source and a photodiode as a light detector. The wavelength of LEDs is chosen to fit the absorption band of each nutrient."
They seem to have moderate success
"Testing on various samples of soils, showed that the optical transducer can evaluate the amounts of NPK soil content as High, Medium and Low."
I'd like to give this a go some day, it doesn't seem too hard. This is not something you could buy as a breakout board yet, but seeing as this paper is 2018, I'm hopeful it might become a product in the near future!
This is about nitrates only. Seems you'll need some expensive optical bandpass filters in the UV:
There is a sensor called TCS230 Color recognition which absorbs the reflected light and converts into the current.
Now N,P and K absorbs certain amount of light and reflects the left out of it. TCS230 absorbs that reflected light using photo-diode and converts into current.
So, you can use it.
Also for more clarity, look at this article:
The link above, posted properly: Detection of nitrogen, phosphorus, and potassium (NPK) nutrients of soil using optical transducer | IEEE Conference Publication | IEEE Xplore. The paper is available only to paying subscribers, but I looked though it.
It is utter nonsense, of no scientific or practical value and containing no quantitative information. Below is the complete description of the material used for "testing" the "sensor".
Three types of the soils with different nutrient were obtained from a nursery shop while the other three were taken from residential areas. The sample specification is listed in TABLE II.
Finally, the article states that a photodiode was used as a detector, but the schematic diagram shows an LDR. Complete waste of time!
jremington:
Complete waste of time!
Of course.
First poster resurrecting a necrothread... that's usually something nonsensical.
Who knows? The poster doing the resurrecting could be one of the paper's authors. It is hard to imagine someone else thinking that the publication provides "clarity".
Any found more information on this? I have several studies downloaded but none of them show the exact circuit and how was the solution with the soil measured via RGB leds..
bump. Anyone has any update on this? I am looking for Soil NPK Sensor as well that can be interfaced with arduino. Thanks in advance.
Still doesn't exist.
what about method described in this video??
Try it and report the results.
definitely will, already expecting delivery for color sensor
I don't know how well and precise is the color sensor method?
I do not think it "quantifies" the values of N, P and K and only shows High/Medium/Low.
jremington:
The link above, posted properly: Detection of nitrogen, phosphorus, and potassium (NPK) nutrients of soil using optical transducer | IEEE Conference Publication | IEEE Xplore. The paper is available only to paying subscribers, but I looked though it.It is utter nonsense, of no scientific or practical value and containing no quantitative information. Below is the complete description of the material used for "testing" the "sensor".
Finally, the article states that a photodiode was used as a detector, but the schematic diagram shows an LDR. Complete waste of time!
As per this earlier post, it does not seem very promising.
I don't know how well and precise is the color sensor method?
It doesn't work at all, as the nutrients of interest are colorless, but we should certainly encourage experimenters discover that for themselves.