DeepPhenology: Estimation of apple flower phenology distributions based on deep learning

Estimation of phenology distribution in horticultural crops is very important as it governs the timing of chemical thinning in order to produce good quality fruit. This paper presents a novel phenology distribution estimation method named DeepPhenology for apple flowers based on CNNs using RGB images, which is able to efficiently map the flower distribution on an image-level, row-level, and block-level. The image classification model VGG16 was directly trained with relative phenology distributions calculated from manual counts of flowers in the field and acquired imagery. The proposed method removes the need to label images, which overcomes difficulties in distinguishing overlapping flower clusters or identifying hidden flower clusters when using 2D imagery. DeepPhenology was tested on both daytime and night-time images captured using an RGB camera mounted on a ground vehicle in both Gala and Pink Lady varieties in an Australian orchard. An average Kullback-Leibler (KL) divergence value of 0.23 over all validation sets and an average KL value of 0.27 over all test sets was achieved. Further evaluation has been done by comparing the proposed model with YOLOv5 and shown to outperform this state-of-the-art object detection model for this task. By combining relative phenology distributions from a single image to a row-level or block-level distribution, we are able to give farmers a precise and high-level overview of block performance to form the basis for decisions on chemical thinning applications.

Automated summary

The paper introduces a novel phenology distribution estimation method named DeepPhenology for apple flowers, which efficiently maps flower distribution on image-level, row-level, and block-level using RGB images without the need for image labeling. The method was tested on apple varieties in an Australian orchard, achieving average Kullback-Leibler divergence values of 0.23 on validation sets and 0.27 on test sets. DeepPhenology outperformed the YOLOv5 object detection model and provided farmers with a high-level overview of block performance for decision-making on chemical thinning applications.