Defense Approaches against Adversarial Attacks on Remote Sensing Applications
Primary Investigator (PI) Name
Kazi Aminul Islam
Department
CCSE – Computer Science
Abstract
Remote sensing systems utilize deep learning models to interpret satellite imagery for tasks such as land classification and surveillance. However, these models are vulnerable to adversarial attacks—small input perturbations that can cause critical misclassifications. This work explores the impact of adversarial attacks on satellite image classification. We evaluate several defense strategies against adversarial attacks. Our findings highlight the strengths and insights for improving the robustness of remote sensing systems against adversarial threats.
Defense Approaches against Adversarial Attacks on Remote Sensing Applications
Remote sensing systems utilize deep learning models to interpret satellite imagery for tasks such as land classification and surveillance. However, these models are vulnerable to adversarial attacks—small input perturbations that can cause critical misclassifications. This work explores the impact of adversarial attacks on satellite image classification. We evaluate several defense strategies against adversarial attacks. Our findings highlight the strengths and insights for improving the robustness of remote sensing systems against adversarial threats.