About the project
BBLeap is an organization that focuses on technological development within the agricultural sector...
At BBLeap they believe in “Farming On Plant Level”. In this they work towards a new leap in agriculture. Every plant has a different need and with their tools they want to anticipate these individual needs and give what it needs at plant level.
In this case, the focus is on using new camera systems, based on AI,
to spray crops in 'Real-Time'.
Challenge
The hardest part inside this project was implementing multiple aspects in one plaform which are in line with eachother and still needs to be clear to use for farmers. Guiding through the owned land, marking photo’s to send to developers and giving information through the photos about the land. Creating the structure and simplicity was challenging and changed multiple times due to the input of the clients.
Using AI isn’t a one direction flow. Farmers have to inform algorythm-developers their needs when spraying the crops, so the development will be as accurate as possible. Ofcourse it’s mostly impossible to create a algorythm what the farmer will find usable in the field. Therefore, a feedback-loop between farmer and developer is created.
The farmer can simulate the algorythm versions through a simulation. This was one of the criteria the farmers had when using AI. The big question was “How can we stay in control and use our expertise when creating algorythms?”
Algorythms can be simulated on existing photos and matched with the opinion of the farmers. This creates a performance rate inwhich the user can see how well his spraying will function.
One of the most important aspects inside the platform is navigating through a map. While interviewing clients, I’ve been told that long lists and a lot of text insn’t beneficial when searching through land and tasks.
The goal after a long day working, is to turn the data of the land into information for algorythm developers very quick as possible. The farmers have long and busy days and having a quick workflow online will be a low thresshold for using this platform and optimizing AI in their business.
Methods
-
There are more companies trying to use AI as a spraying solution, using HDD to transfer data via letterpost. The disadvantage is letterpost is time consuming and a higher risk of losing data. However, the outcome is up to 98% of using less chemicals while spraying the crops.
-
The most important element in this journey, understanding the target audience. The farmers taught me multiple things about their businesses and lives. First, it is important to know they work 10 - 14 hours if not more in peakseason. Time management and efficiency is highly important when implementing other activities in terms of optimizing their business. Second, the user want to stay in control over the outcome. If algorythm ain’t working as they want, it isn’t useful.
-
Best: Using AI can be helpful when targeting/dosing pesticide more accurate and is easy to combine with other types of technology. The healthy crops will be protected from chemicals, while weeds can get a higher and effective dosage.
Good: AI has a lot of skills, but it is still important to train properly when using effectively.Bad: Blind trust in AI can lead to wrong decisions. it is important to use AI as a validation of human expertise
-
Trends are a major aspect when creating a product. Digitalization, healthy products and sustainability are growing in the agriculture. these trends not only provide insight, they also validate previous used methods.