Efficiently Monitor Crops With Drones & AI
A Semi-supervised Deep Neural Network for Fruit Yield Estimation for Intelligent Agriculture
Tags: City University of Hong Kong, Hong Kong, Transportation & Automotive, Agriculture & Food
The invention is a semi-supervised deep neural network using limited labeled and many unlabeled images for efficient fruit quantity estimation from drone-captured images. It reduces the time and effort in training image annotation and easily adapts to different crop categories. The network's architecture enhances learning with RGB and grayscale images, improving fruit recognition. Farmers can use this system for precision agriculture, while other stakeholders can benefit from improved data for crop management and economic planning. It also presents an opportunity for drone makers to enhance their product offerings.
IP Type or Form Factor: Patent Pending; Software & Algorithm
TRL: 5 - prototype ready for testing in intended environment
Industry or Tech Area: Drones & UAV; Agriculture
