Click on “Download PDF” for the PDF version or on the title for the HTML version.


If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Mask R-CNN Based King Flowers Identification for Precise Apple Pollination

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan www.asabe.org

Citation:  2021 ASABE Annual International Virtual Meeting  2100670.(doi:10.13031/aim.202100670)
Authors:   Xinyang Mu, Long He
Keywords:   Apple flowers, Deep learning, King flowers, Object detection, Precise pollination

Abstract. To help achieve optimal precise pollination for apple flowers, identification of individual king flowers is one of the most critical steps in developing a robotic pollination system. Despite plenty of studies have provided direct methods and approaches to detect flowers in the clusters on the apple canopies, none of them managed to identify the king flowers. Typically, each cluster has 5-6 individual flowers, and king flowers can be occluded by the lateral flowers because of its central position in a flower cluster. Meanwhile, the king flowers share the identical features (i.e. color, shape, size, etc.). While in most cases flower clusters open sequentially from the king flower to the lateral flowers. This spread in time of anthesis presents an opportunity for selective pollination. To accurately determine the pollination targets and timing, it would be important to monitor the flower blooming stage. In this study, firstly we established a dataset for apple flowers throughout the blooming stage from first king bloom to full bloom. Then a deep learning algorithm was developed to identify and locate the king flower thus to calculate the percentage of the king blooming and its distribution in the tree canopy. Along with the horticultural knowledge and concurrent climate condition, the outcome from the study is expected to provide decision-making information for robotic pollination.

(Download PDF)    (Export to EndNotes)