Client Profile

Our Client is a major player in the Agtech industry, headquartered in Switzerland. They are a well-diversified agriculture enterprise having interests in seeds, genetic research, agrochemicals, and other crop protection products. They are helping farmers around the globe to meet the challenges of the future: to grow more from less with affordable, simple, and turnkey solutions.

Client Requirement

Our Client wanted a fast and reliable application to count corn seeds and to eliminate the manual and conventional way which is tedious and time-consuming.

Feathersoft Solution

Feathersoft assisted Our Client to overcome critical concerns, reducing time and effort, and maximizing output. A machine learning model was introduced using cutting-edge technologies to accurately detect individual corn seeds.

  • A reliable and quick iOS application was developed to detect seed trait data from the image of the corn.
  • ¬†With the help of deep learning techniques, we detected every single corn seed on one side of the corn and then calculated the whole seed count in that entire corn.
  • A decisive platform was developed that could collect seed trait data at any location without any prior experience.

Key Challenges

  • Building the perfect dataset for training the models
  • Training and checking different machine learning models for better, decisive results.
  • Training the model by changing image resolutions to achieve accuracy.

Impact on business

  • Superb convenience: An iOS device with a stable internet will do the challenging job of counting seed trait data which was otherwise a very cumbersome and manual activity.
  • Massive productivity gain: what was a very time-consuming and tedious job, can now be done within seconds and without the help of any professionals.

Technology Used

Deep Learning
Computer Vision
python
Tensorflow
ios