Add to Chrome
Add to Firefox
Add to EdgeDeep Transfer Learning For Plant Center Localization
Apr 29, 2020
Plant phenotyping focuses on the measurement of plant characteristics throughout the growing season, typically with the goal of evaluating genotypes for plant breeding. Estimating plant location is important for identifying genotypes which have low emergence, which is also related to the environment and management practices such as fertilizer applications. The goal of this paper is to investigate methods that estimate plant locations for a field-based crop using RGB aerial images captured using Unmanned Aerial Vehicles (UAVs). Deep learning approaches provide promising capability for locating plants observed in RGB images, but they require large quantities of labeled data (ground truth) for training. Using a deep learning architecture fine-tuned on a single field or a single type of crop on fields in other geographic areas or with other crops may not have good results. The problem of generating ground truth for each new field is labor-intensive and tedious. In this paper, we propose a method for estimating plant centers by transferring an existing model to a new scenario using limited ground truth data. We describe the use of transfer learning using a model fine-tuned for a single field or a single type of plant on a varied set of similar crops and fields. We show that transfer learning provides promising results for detecting plant locations.
