The Data and Decision Sciences Lab is constantly working with local community partners on several data related projects that have a local and global impact.
Automatic Detection of Maize Tassels from Unmanned Aerial Vehicle Images with Convolution Neural Networks
Community Partner: University of Nebraska–Lincoln – Department of Agronomy and Horticulture
Team Members Involved
Maize tassels play a critical role to indicate the reproductive growth and health of the maize crop. Observation of maize tassels and plant growth is a tedious and challenging task which is mainly done by manual human efforts. In recent years, image-based approaches have received much attention in plant-related studies. The aim of this project was to detect the tassels of the maize crop obtained from an unmanned aerial vehicle (UAV) using convolutional neural network (CNN). The project tested the performance of CNN models with the custom trained convolutional network, and pre-trained CNN models. With the finalized model the project obtained an accuracy of 96.2% on the training dataset and 95.4% on the validation dataset. For the given test images from the UAV, the project generated density plots to show the high probability location of tassels.