Deep Learning for Recognizing Mobile Targets in Satellite Imagery

TL;DR

This paper extends CNN classification methods to a sliding window detection algorithm for mobile targets in satellite imagery, achieving over 95% accuracy on the xView dataset, aiding various practical applications.

Contribution

It introduces a sliding window detection approach based on CNN classification for mobile targets in satellite images, improving detection accuracy.

Findings

Detection and classification accuracy over 95% on xView dataset

Effective extension of CNN classification to detection tasks

Applicable to multiple mobile targets in satellite imagery

Abstract

There is an increasing demand for software that automatically detects and classifies mobile targets such as airplanes, cars, and ships in satellite imagery. Applications of such automated target recognition (ATR) software include economic forecasting, traffic planning, maritime law enforcement, and disaster response. This paper describes the extension of a convolutional neural network (CNN) for classification to a sliding window algorithm for detection. It is evaluated on mobile targets of the xView dataset, on which it achieves detection and classification accuracies higher than 95%.