Knowledge Base & FAQ

We're working on simplifying our calculations, but they're often off because the operator forgets to pause, resume, or stop a session, (a session = 1 field), once they start.

In 2023, we saw many use cases where an operator started a session, finished a field, went to another, or went back to the shop to park the combine for the day, and forgot to stop the Farmwave session. This means the system keeps taking pictures every 3 seconds, and if it's counting a lot or no grain at all, this will definitely impact your loss numbers and calculations.

**NOTE**

We often get asked why we can't connect, or tie this into, the machine harvest mode. We're working on it, but this takes a lot more engineering work with the OEM's than we, or they, can handle right now, but we are looking into it.

In 2025, we'll be experimenting with a new GPS sensing feature. Here's how it will work:

• Farmwave GPS senses you have not moved in x seconds

  • pause Farmwave

  • alert operator session has paused

• Farmwave GPS is paused and senses you've begun moving

  • in X seconds resume Farmwave

  • alert operator sessions has resumed

We consider a session a field.

Start a session when you start a new field, and stop a session when you finish that field. Should something happen in between, (e.g. machinery breakdown, stop to wait on the grain cart, etc), you can pause the session and then resume when you're going again.

Reports in farmwave.app portal are saved as sessions.

Currently, Farmwave works with:

• corn

• soybeans

• edible beans

• wheat

• peanuts

• cotton

Crops in beta:

• canola

Crops in training:

• barley (you can try using the wheat model)

• oats

• lentils

• peas (you can try using the soybean AI model)

If your idea of validating Farmwave involves getting a drop pan, throwing it out at about the exact same time as one of the cameras takes a picture, and then counting the kernels on the ground and comparing it to what is seen in that Farmwave image ... you're overthinking it.

Farmwave has never claimed to, and never will be, perfect.

As mentioned in another article in this knowledge base, we do not see underneath material other than grain (MOG).

Given it takes a picture every three seconds, there is the time interval in between where we may not see something.

There is the off chance that lighting is just a certain way, a piece of grain is just a certain way in flight or on the ground, that we miss one.

It's possible that a peanut shell, (as an example), is actually opened so it's only half the shell but in flight or on the ground it looks like a full peanut so we counted it as such therefore it's a false positive.

Farmwave, nor any other kind of AI, is perfect.

It's the trends on the graph that make the difference.

The graph is where it's at. Conditions change throughout the day, from field to field, and from crop to crop. Setting your thresholds to where you are willing to allow for loss, (because nothing is perfect), is one of the more important aspects of setting up Farmwave.

As the graph moves, it will go up if more loss is seen, and if it leaves your acceptable loss line, (the green zone), that's when you want to look at the images and see what changes can be made and where it's coming from; header or combine. When you make those changes, the graph will trend back down, letting you know that the changes you made worked!

At the very core of Farmwave, that's it!

The advantages of Farmwave are the amount of counts we do in a single acre, (300+), or even in a single day, (>20,000), without stopping the combine and having to get out of the cab. 

This is not something a human could do, as well as concurrent loss checks, when changes are made to see if they worked or not.

Farmwave is an AI-assisted tool to allow for ongoing loss checks without having to stop the combine and get out of the cab. It also adds more value compared to other loss checking devices by accounting for header loss sensing which is nonexistent today. In certain crops, such as peanuts and cotton, it is the only loss sensing technology available on the market today.

Our goal at Farmwave is to make your life and your job easier by integrating value-added technology. We want to keep it simple.

Don't overthink it.

Yes. Right now you can download in JSON format which is very commonly accepted in most other software programs. You can also email reports and download to PDF.

You can also share, via email, reports from the farmwave.app portal to others.

In addition, you can download data in CSV format for uploading into things like Power BI, or another data analytics tool.

We anonymize and retain copies of each image for training purposes of the AI only. Otherwise we do nothing with your data. You own it. We do not sell it, use it for marketing, or otherwise.

No system is stabndard. We custom quote everything due to the numerous variations in not only make and models harvesters, but headers, custom setups, and more.

The base system includes:

• The Farmbook, our AI-centric all-in-one touchscreen tablet built for AI at the edge. 

• AI magnet mount cameras

• Power Assembly Disconnect (PAD)

Absolutely. Pricing is per combine. So each combine will require it's own hardware and annual license fee.

The Farmbook is a custom dedicated AI edge machine. It's a machine dedicated to running AI tasks, and even learning, at the edge.

Your entire farm operation is a business, and it's getting more and more autonomous over time. The Farmbook is being built for the future to integrate with the machines themselves, as well as other platforms, to bring the true power of machine learning and artificial intelligence right into the cab, and in your hands ... without the need for cloud connectivity.

The possibilities are almost endless.

What is currently working in the cab to monitor harvest loss is just the beginning.

Make sure that your header cameras are pointed close to straight down to be able to see the ground.  If the camera is angled too far back, the field of view will be blocked by the stalks.

Ideally the field of view should capture part of the tail end of the rear of the combine with part of the sky as the backdrop.  

This is a personal preference, but the concept of Farmwave is to capture consistent data and show averages, so placing header cameras midway on each side (18”-24” off the ground)  of the header is optimal for giving the best estimates of harvest loss.

Each combine cab will have different options for mounting.  Additionally, mount placement will depend on other screens/tech in the cab.  Ideally the Farmbook should be placed where the screen can be viewed consistently while running to monitor loss.

Click on the graph for the camera you’d like to view. This will open a detailed view of the graph with thumbnails. Click the thumbnail you’d like to view.

Log in to the tablet

• The only available account right now is the vision user.

• Password: password
  
  

Start the Farmbook app

• You’ll find it at the top of the tray on the left side of the screen.
  
  

If the app hangs on first startup

• Swipe down from the top of the screen to minimize it.

• Reboot the tablet:

• Tap the top-right corner.

• Select the power button icon.

• Choose Restart/Reboot → Yes.

Log back in

• Sign in again as the vision user.
  
  

Create your new account

• You’ll be prompted to sign up.

• This account is new for everyone, so all users need to register—even if you had an account before.
  
  

⚠️ Note:

• The first-time startup issue should be resolved in the next update.

• Eventually, logging in as the vision user will no longer be required, but this will come in an upcoming application update.

There's a configuration change required for your systems. Sometimes, users encounter issues with the application's keyboard due to the system's on-screen keyboard blocking it but not being visible. This change is quick. Please put your system online, and we'll make the change. We apologize for any inconvenience caused.

Factors to consider

• Assumed yield (low, medium or high)

• Ground cover (excessive, average or minimal)

Thresholds can be adjusted at anytime before starting a session.  Just select your configuration (corn, soybeans) and adjust your min/max

To best determine your minimum or GREEN threshold settings for corn and soybeans

• Use assumed yield by associating a 1(low), 2(medium) or 3(high)

• Use ground cover by associating 1(excessive), 2(average), 3(minimal)

• Add together your values and that will give you the minimum setting

  • e.g. regular yield soybeans with minimal ground cover (2+3 = 5)

To best determine your maximum or RED threshold settings for corn and soybeans

• Use assumed yield by associating a 2(low), 4(medium) or 6(high)

• Use ground cover by associating 2(excessive), 4(average), 6(minimal)

• Add together your values and that will give you the max setting

  • e.g. regular yield soybeans with minimal ground cover (4+6 = 10)

In the examples above, the settings for the min/max would be 5 & 10

• </= 5 = GREEN

• >5 <10 = YELLOW

• >/=10 = RED