Scanning optical organic matter sensor

Organic content can be measured in a variety of ways. The most accurate tests involve measuring the amount of combustable carbon-based matter in a sample by putting it in a chemical oven. In a test like the one pictured below, soil is allowed to sit undisturbed in a falcon tube for a period of time in which the available organic matter floats to the top.

I would like to use a combination of a laser and a photometer to make a scanning optical device that moves upward from the bottom of the tube. An arduino would log the values of intensity of the light that permeates the tube. The intensity would be reduced for the section of the range that is obstructed by organic matter (or mineral matter at the bottom). I will write software to measure the length of this band and then calculate the volume from this value (thickness*pi*r^2). This method would be faster and more consistent than measuring each band individually by eye.

Electrical Conductivity & PH Sensing Platform

This platform will also be designed for bench top use because soil samples need significant time to stabilize before accurate readings can be collected. I am focusing on sensors that connect via BNC because they seem to be the cheapest option for arduino interfacing.

The PH interface will be based on this implementation from Sparky’s Widgets (a great all around resource for electrode sensing applications).

I will try to build this interface for the pH probe first and then see how easy it is to modify to read EC. Sparky’s Widgets offers an integrated solution for purchase, but I would like to reengineer it myself if possible.

Temperature and Moisture Field Sensor

This is the only sensor platform I am designing for field use. Right now I am thinking I will design the temperature portion of the sensor to be used right after the soil core sample is taken. I am planning to use infrared sensing in the resulting ~10cm hole to measure soil temp. There are a lot of options out there for infrared sensing – here is a sparkfun sensor I am looking into. I need to get an idea of what precision is necessary for this application.

For moisture I am planning to use a very robust sensor from Adafruit that has been used by students in the past. Building the circuit for this sensor might be a good place to start since I have models to work off of.

NPK and Munsell Color Optical Sensor

A lot of soil nutrient tests involve chemical reactions leading to color changes. I am interesting in using RGB sensors to find a precise value for the color of a solution so I can determine it’s nutrient density. Munsell color test is the official color metric for soil research. I am now researching ways to use RGB LEDs and photosensor to test color of an object. A sensor capable of detecting color could be used for both bench-top NPK tests and Munsell color testing.

I have been weighing the options of testing soil in the field vs. on the bench for parameters of interest. The parameters I am focusing on are soil ph, temperature, moisture, fertility, conductivity and organic content. Of these seven, only temperature and moisture must be collected in the field. I believe that the others can be more accurately determined in a bench-top environment. I am developing my hardware designs accordingly. The next steps for me are to track down materials to build this ph probe and figure out how to interface it with an arduino board and relay information back to the field science android app through bluetooth. I am also planning to have a conversation with Mike Deibel this week about using optical measurements for organic content and fertility tests (and possibly munsell color tests). I also have a back-burner interest in geophysical surveying techniques for archeological dig sights that is still very much in the research stage.

My current project is to develop a functional and consistent set of sensors and protocols to collect information about important soil parameters such as pH, fertility, conductivity, temperature, organic content, and moisture. In Iceland and Nicaragua, the approach was to construct a unified soil sensor platform using Arduino and Yoctopuce technology that takes readings in the field.

The potential benefits of this approach are that every parameter of interest is collected in the field and immediately saved in .csv format on a nexus device. The downside is that it is not particularly easy (or in some cases, possible) to test every parameter in harsh environments accurately or without causing damage to the sensors.

With this in mind, I am planning to take a sightly different approach: a sensor for use in the field that measures parameters that must be collected in situ (such as temperature and moisture), supplemented by benchwork preformed in the evening to collect other relevant soil parameters. The evening benchwork would also have hardware and software components. For example LED/photometer sensors could be used to measure organic content and other optical sensors could be used to make observations about the color of the soil and its composition.

The next steps for me are:

1. Learn more about the chemistry of soil fertility and ph and find a way to test them cheaply without using standard kits.
2. Consider the hardware element. Look for cheap sensors that are compatible with Arduino boards and begin preliminary designs. I will probably build the temp/moisture sensor first since it most closely follows the paradigm of sensors that have already been developed as a part of this research.