Improving Aerial Instance Segmentation in the Dark with Self-Supervised Low Light Enhancement

TL;DR

This paper introduces a self-supervised low light enhancement method for aerial images that improves detection and segmentation performance in dark conditions, while also providing a new GAN-generated dataset for evaluation.

Contribution

It presents a novel self-supervised enhancement technique integrated with detection and segmentation, and creates a new low light aerial dataset using GANs.

Findings

Enhanced detection and segmentation accuracy in low light aerial images.

Low overhead in computational and memory resources.

Effective evaluation dataset for low light aerial vision tasks.

Abstract

Low light conditions in aerial images adversely affect the performance of several vision based applications. There is a need for methods that can efficiently remove the low light attributes and assist in the performance of key vision tasks. In this work, we propose a new method that is capable of enhancing the low light image in a self-supervised fashion, and sequentially apply detection and segmentation tasks in an end-to-end manner. The proposed method occupies a very small overhead in terms of memory and computational power over the original algorithm and delivers superior results. Additionally, we propose the generation of a new low light aerial dataset using GANs, which can be used to evaluate vision based networks for similar adverse conditions.