The Cranberry Fights Back

Artificial Intelligence – and, in particular, a facet of it called Deep Learning – has been making dramatic progress in tasks that have traditionally required humans to perform. One of the major advances has been in Image Recognition/Classification. Most often these tools have flashy applications, like recognizing cats in YouTube videos. And for most organizations, these tools/techniques seem out of reach, as they require substantial technical abilities to be able to pull off. However, sometimes having the right people and determination to help the user can lead to the development of a solution that helps people with their real problems.
 
In this article, I want to outline how a midwestern marketing agency built a solution for Wisconsin cranberry growers using some of the latest advances in Artificial Intelligence.
 
When it comes to cranberries, it’s easy to close your eyes and think of the fruit and the myriad of products that are made with it. All that intense flavor just waiting for a bite. From sauces, and salads, to desserts and drinks, cranberries truly are an exceptional gift of nature, and we as consumers have easy access to them at grocery stores as fresh, frozen, dried and juiced berries.  The fact that there are real people who grow and care for cranberries is often forgotten, but cranberry farmers work incredibly hard each year to grow a healthy and productive crop.
 
One of the problems facing growers is effectively and sustainably managing pests that could damage the crop. With names like “Black Headed Fireworm” and “Cranberry Blossomworm,”  it’s easy to imagine why these little pests can be a farmer’s nightmare. Recognizing these bugs can be challenging and can mean the difference between a successful crop versus an unsuccessful one. Growers bring in pest consultants and other experts to help identify the insects and appropriate treatments, a costly process with a turnaround time that can sometimes allow insects to mature and spread before treatment.
 
Enter AI and Laughlin Constable. Through LC’s work with the Wisconsin State Cranberry Growers Association, we began to explore whether there would be an easier way for growers to identify insects in the field.
 
The Laughlin Constable software engineering team had been working on a few experiments using AI, and in particular Deep Learning to see how they could solve previously intractable problems. It turned out that they had some of the tools to solve the cranberry grower’s problem. LC determined it could create a pest identification smartphone application. Growers would be able to use the app to take a photo of an insect and receive an immediate identification of the pest.
 
As always, a solution is simple in its conception, but delivering a working product is far more complicated. We had to build out a data gathering application to create an insect image database, which we did using a custom mobile application. We also had to enlist the help of subject matter experts, who could label the images for us, and growers who were willing to help with the data gathering effort. And finally, we had to train and deploy the Deep Learning Model that could help growers identify bugs in the field in near real time.
 
The solution will be rolling out later this year, and we are really excited with the possibilities!

We will be speaking at the Spark + AI Summit 2018 on this topic. We would love to hear from you and share our experiences of using AI to deliver real solutions to real problems for our clients.

For more tips, tricks, or insights on how to take your marketing from now to next, subscribe to our newsletter or contact Nicole Stone – Director, New Business Development at nstone@laughlin.com or 414.270.7235.