From Drones and Mining to Tractors and Farming, Persistence Data Mining (PDM) Applies Their Hyperspectral Imaging and Data Technology to Solving Soil Health Issues

April 26, 2019 | By Leanna Sweha

Persistence Data Mining (PDM) is a software analytics firm that uses hyperspectral imaging to measure soil nutrients.

PDM’s technology can increase producer profits while reducing fertilizer overuse, which is a major cause of water pollution.

The firm was founded in 2012 by Penelope Nagel and Brian Zamudio. We recently spoke with COO and President Penelope Nagel to learn more.

 

You describe yourself as a 9th generation farmer but your address is La Jolla. What’s the connection?

My family owns farmland in Illinois, where we grow soy and corn. I was born there, but my family moved to California when I was a young child. I inherited the farm from my mother, who managed it from California. We would go back every spring and fall, so it has always been a big part of my life and continues to support my family.

I live on the coast so that I can continue my other passion, which is channel swimming. And now that we are developing a presence in California, it just makes sense.

My love for ocean actually inspired PDM. Within the channel swimming community, I started hearing a lot of blame put on farmers for causing algal blooms in coastal waters, including Southern California. Now, my swims are long – up to 15 hours – and I never experienced algae blooms. I felt that farmers were getting bad rap. On the other hand, I understood that fertilizer runoff is a problem. So, I looked deeper into what was going on.

That’s when I learned about how inefficient and ineffective the current soil nutrient testing is. Current chemistry-based testing uses a single cup of soil for every 4 acres of land. I decided that there had to be a better way to do this, and became motivated to apply new technology to the problem.

 

Is it true that PDM started out as a mining-oriented tech firm? How did it pivot to AgTech?

Yes. I had known my co-founder and our CEO, Brian Zamudio, since college. He was interested in using technology for mineral exploration for oil and gas drilling. He developed a way to do hyperspectral imaging using drones to search for key mineral signatures. When I joined the management team, we were moving forward – we had mineral claims and had gotten an FAA exemption to fly our drones commercially.

When the last administration put the kibosh on a lot of this kind of exploration and drilling, we quickly realized our business model would not be profitable. So our team sat down and asked – how else can we use this technology?

We learned that drones had big potential in agriculture. That’s when I realized that we had a potential solution to the soil testing and fertilizer overuse problem. So, I said, I’m a farmer – let’s look at NPK (key soil nutrients). Even though the rest of the team didn’t know the first thing about NPK, they said, you sound confident, so let’s go ahead.

 

What steps did you take to pivot?

First, we sent our science team to a lab to do spectral analysis on soil and confirmed that we could indeed get a signal for NPK.

Next, we hired our lead agronomist who had contacts with a fertilizer company that allowed him to use its land for further testing. He compared our spectral data with with data from traditional testing to build basic algorithms. His results validated our spectral analysis method for NPK testing. However, we also learned that, although the testing could be done from a tractor or ATV, it could not be done with drones.

Mineralogy of soil differs from area to area, and depends largely on the binding capacity of soil. We have to take all of that into account with our testing method. We were able to build our soil database from USDA data, as well as from our own data. We were also fortunate to be able to access soil data that had been recorded by Climate Corp after it sold its soil labs to Land-o-Lakes.

 

How is data taken and processed?

PDM replaces a more expensive, less efficient, chemistry-based soil test with with an efficient imaging test which gives near-real-time results from our cloud-based analytics.

We image from a tractor or from an ATV either after discing or after spot tilling. Data is sent to the cloud, where our analytics go to work, applying our algorithms to get readings on NPK.

We can take more samples per acre than the current chemistry-based test. We take 4 readings per acre. Why? Because currently, variable-rate sprayers can adjust four times per acre.

Our primary focus is on NPK and micronutrients because they are so important. Take Nitrogen for example. Most farmers don’t test for it because it is expensive. Instead, they follow university recommendations, that can be to general or out of date.

We imagine a future where our technology is taking real-time measurements from a variable rate sprayer that can make an infinite amount of adjustments also in real-time.

 

What strategic partners are you working with?

Our initial deployment strategy involves working with the agronomic service providers who take field samples, as well as the labs who process soil samples. By integrating our test into their services, we can cut their costs.

In 20xx, we had a collaborative agreement with the USDA ARS to validate our technology. We were very excited when our ARS collaborator said our technology could cause a revolution in agriculture.

We have a partnership with w/ Melvern Panalytical, to make our sensors.

On the international front, we have done several pilots with Mahindra in India.

 

You’ve been through several pitch and accelerator programs – Tell us about that experience.

The first competition we participated in was a pitch program at UCI in 2016.

In 2017, we participated in AgLaunch accelerator, which was a 12 week hands-on intensive program. This program provides access to a network of growers who participate in testing and business development. It was a great experience particularly because of the farmer network. We worked with growers in Tennessee, Mississippi and Arkansas. We got verifiable data showing correlation between our sensor-derived data and chemistry-derived lab data. In fact, we have evidence that, in these pilot tests, our data may actually have been more exact than the lab data.

Next, we applied to and were accepted into the 2018 THRIVE accelerator cohort. This greatly helped us build a network of strategic partners, such as the one we forged with Trimble’s agriculture division. Thrive was great – we got so much support from industry and farmers.

 

Where is PDM in terms of going to market?

We are a pre-seed stage company. We are currently raising the last $500 thousand we need to go to market. It’s been a long cycle to get to market – over five years – but we are now ramping up to deploy.

Interestingly, unlike a lot of tech firms, we have not applied for or received grant funding to develop our technology.

We’ve secured a process patent on the method we use for obtaining our algorithms, that is how we interpret the spectral data and all of the variables inside the readings to get to nutrient levels.

We are in it for the long haul – we can see the sun coming up on our company, but it’s taken us a long time. Unfortunately, we have found that there is a disconnect between investment funding in this sector and boots on the ground in the field.

 

Do you have any paying pilot customers?

US Tobacco was our first pilot customer. This involves a project on my land in central Illinois in which we are growing hemp organically. We have been using our technology for validation, as well as for soil mapping.

We also have support from the Ohio Soy Board to test on no-till fields. This involves taking an image from the coring machine, which is used to pull core samples for traditional chemical test.

 

As a farmer, what insights do you have into marketing your technology?

We want to make sure the science behind our technology is good before we go to market. I get farmers – you need to prove out your product before they will buy. Farmers are “one and done.” If a product does not perform the first time, it ends up like the many gadgets from the past 150 years that I have in my farmhouse attic.

 

What is your plan for your California rollout?

First, we want to complete our validation process. We have done some limited testing in California, for example, we’ve done some on irrigated ground in my friend’s citrus grove.

We are in the process of submitting a proposal to work with West Hills College’s Farm of Future program and get some pilots going there.

California farmers are diverse, and that includes whether they do soil testing or not. In the Salinas Valley, growers do a lot of testing. We want to make it easier without additional cost so that it becomes a regular part of the process. Part of our mission is to emphasize the importance of good soil testing – for both NPK and micronutrients.

About the author: Leanna Sweha is an attorney with an agtech background and is Program Director for AgStart. She is passionate about promoting new sustainable agriculture companies.

 

New Innovator Spotlights will appear soon, featuring the innovators and change makers in the sustainable food and ag innovation space. If you know of someone we should be covering, please email us at hello@thevine.io


Stay Updated