This code is a supplementary material for the paper A methodology for detection and localization of fruits in apples orchards from aerial images by Santos & Gebler (2021), presented at SBIAgro 2021 (the XIII Congresso Brasileiro de Agroinformática). This work presents a methodology for automated fruit counting employing aerial-images, including algorithms based on multiple view geometry to perform fruits tracking. DOI 10.5753/sbiagro.2021.18369.
The present code details just the apple's detection part, as presented on Section 2.1.1 in the paper. We have employed a Faster R-CNN network with a ResNet-50 backbone, using tochvision 0.10.0 (model), PyTorch Lightning (training) and Albumentations (augmentations). Check the ADD256 repo for the data.
