Commercial
AI and ML-based solutions
R&D. Development. Consulting.
MENU
FDSF
Feed Detection System for Farms
HOME / OUR WORK / FDSF
/ task
Develop a system for detecting feed quantities on dairy farms. Utilize the customer’s existing video surveillance infrastructure to determine the presence and volume of feed.
/ Solution
A neural network classifier is trained not only to detect the presence of feed in a designated area but also to accurately determine its quantity across 4 classes for timely replenishment.
/ technologies
Technological complexity: The neural network classifier is trained not only to detect the presence or absence of feed in a designated area but also to accurately classify its quantity into 4 categories for timely replenishment. Achieving a quality level above 80% required significant initial work with large datasets. As a result, the final outcome exceeded the planned performance.