Making farming intelligent and data-driven

A unique start up in agricultural domain

When you have a garden or a small orchard behind your house, there is no pain to take care of it. However, what happens if there are hundreds or thousands of trees to look after? Right, it becomes extremely time-consuming to get around the orchard and even more difficult to track the condition of each tree. SeeTree is an innovative product aimed at the agricultural domain. It was founded in 2017 in Tel-Aviv. Its main idea is to help farmers manage their orchards effectively and make data-driven decisions.

It's not magic but cutting-edge technology

The drones fly around the garden and monitor all trees, gather images which are later processed and analyzed with the help of AI algorithms. This way farmers are provided with relevant data about the condition of each tree and its precise location. It helps to make data-driven decisions which result in higher harvest figures and ROI.

A GPS processing pipeline

In order to work with data, this data should be collected. So quadcopters fly around and take pictures of the orchards. Then the images and other GPS data are uploaded with Field upload tool. The tool then verifies directories structure, accounts images, extracts info like mission (farm) name, date, compresses GPS data to .zip, sends GPS data to Google Cloud Bucket. We built this solution using NodeJs / Electron stack that allowed to have stable performance.

Active score learning

SeeTree uses AI and special algorithms to track the conditions of the trees and their fruits. The system is constantly learning and improving to give better results. It can understand how ripe the fruit is and if it is ready to be harvested. It can provide insights on whether the tree is ill or should be watered. Our engineers developed a special Python / Flask wrap that replaced already existing trees health score learning inside Google Kubernetes cluster.

Our DevOps solutions

Because of high traffic volume, our developers had to find a solution for scaling and being stable at the same time. It was where Kubernetes came into help. To schedule and monitor many workflows we used Airflow with Google Cloud. It also had a positive effect on cost, as our client was able to cut down on extra spendings.

Outro

It was a challenging product but it gave our team a lot of opportunities to work with the latest and most advanced technologies. We are proud to make a contribution to something as big as SeeTree. And last but not least, we enjoyed cooperating with the team from all over the world.