Lonza’s Axcela brand is an established producer of agricultural pest control methods focussed on European markets. Their brief was to create an in-the-field tool to help farmers use the correct quantity of their product to help avoid any environmental impacts.
Our solution was to leverage the power of AI to train our app how to spot the minute particles of the product using the device camera, and return the results as a quantity per definitive area. Using specific training data, we were able to push the confidence level close to 99%, thus allowing extremely accurate usage in real life, no matter the light or ground conditions.
Collating high quality training data allowed us to effectively train a highly accurate neural network capable of identifying the minute product pellets ~3mm in diameter from a distance of 1.5m. Field testing proved the model accurate up to 99%.
A strong technical and development team, Sandbar excelled in delivering an app that hits the mark in terms of accuracy as well as producing an unparalleled User Experience.
Adrian Bell
Agricultural Communications, Axcela
