Backstory on scouting

Scouting a field has always been a manual, tedious process. It involves going out to a row of crops, evaluating each individual plant’s growth stage, spacing, seed placement, and more. This then must be repeated several times over a range of a farmer’s thousands of acres of land—all to accurately evaluate an equipment’s functionality and potential outcomes for the grower.

POGO

Precision Planting created a tool called the POGO Stick to aid in this process for our internal agronomy teams. This came along with its fair share of hardware issues, almost making the scouting process worse at times. But it did provide one important solution: the ability for a user’s data to be stored in the cloud and for calculations to run automatically.This allowed the POGO Stick to be used beyond our agronomy teams and then became used by our Sales team and Premiere Dealers.

Later, a “light” version of this product came out called POGO+. This eliminated the hardware all together and was only a mobile app. This allowed users to quickly evaluate plants’ growth stages, but eliminated any ability to measure spacing or seed placement.

Original POGO Stick used to calculate emergence, spacing, and population

SmartScout

Meanwhile, our R&D teams were developing sprayer technology which used a camera system mounted to sprayer arms and machine learning to detect plants in order to automatically spray for weeds. We adopted this same technology and applied it to scouting fields. The data model allowed us to build an all new app for iPad which leveraged the built-in cameras and allowed a user to scout two rows at once without any manual inputs or calculations.

Early versions of the app we slow (refreshing detection about 4x per second), but as improvements were made, the detection is quick enough to allow a user to walk at a normal pace while picking up every plant along the way. After years of iterations and updates, SmartScout is now available to Precision Planting’s network of Premiere Dealers. Over the past two years, I have contributed to the design and direction of the latest release of new features, primarily as they relate to the augmented reality scouting experience:

• “Manual” scouting

• Active Run UI

• Mid-run options

• Available on mobile

Manual Mode

The model used to detect plants is currently only trained on corn crops in their early stages of emergence. Manual Mode allows users to individually select any crop at any stage and indicate its level of emergence. Spacing values are automatically calculated—a similar experience to the native Measure app.

Alternative button layout options tested for manual mode

UI Refresh

Early versions of the app contained UI that was highly technical, intending for engineers to see a lot of data in real time. The latest live release includes a simplified, accessible UI which allows users to see exactly what they need without the distraction of what’s happening behind the scenes.

Mid-run Options

This refresh also included some additional smaller features including in-run notes, image capture, and data table. The notes and images are geo-tagged and available for further review after the run is complete.

Mid-run table displaying current progress of scouting

“When can I download it on my iPhone!?”

This was the number one question we received. The iPads were a great start from an engineering perspective, but with the capability improvements of hardware and the improved performance, SmartScout was finally ready for mobile. So the natural next step was to convert the designs to a mobile-friendly view.

Next stop… Android!

Challenges

While not exhaustive, here are a handful of things we encountered and had to consider as we continued making updates to the AR experience:

• Lighting conditions

• Wind

• Ground texture

• Battery life and over-heating

• Angle at which the user holds the device

• Undo mistakes in Manual Mode

• Users who are new to AR experiences

• Users’ trust in the accuracy of the data model

• Calculations for each crop type