Investigating the Challenges of Class Imbalance and Scale Variation in Object Detection in Aerial Images

TL;DR

This paper addresses the challenges of object detection in aerial images by modifying the Faster R-CNN architecture, experimenting with backbones, data augmentation, anchor sizes, and loss functions to improve detection accuracy.

Contribution

The study introduces specific modifications to Faster R-CNN, including backbone selection, data augmentation, anchor tuning, and loss function adjustments, to enhance detection of small and densely packed objects in aerial images.

Findings

Achieved 4.7 mAP improvement over baseline

Enhanced detection of small objects

Improved robustness to scale variation

Abstract

While object detection is a common problem in computer vision, it is even more challenging when dealing with aerial satellite images. The variety in object scales and orientations can make them difficult to identify. In addition, there can be large amounts of densely packed small objects such as cars. In this project, we propose a few changes to the Faster-RCNN architecture. First, we experiment with different backbones to extract better features. We also modify the data augmentations and generated anchor sizes for region proposals in order to better handle small objects. Finally, we investigate the effects of different loss functions. Our proposed design achieves an improvement of 4.7 mAP over the baseline which used a vanilla Faster R-CNN with a ResNet-101 FPN backbone.