Anzeige
Erschienen in:
2017 | OriginalPaper | Buchkapitel
Object Detection with Proposals in High-Resolution Optical Remote Sensing Images
verfasst von : Huoping Ding, Qinhan Luo, Zhengxia Zou, Cuicui Guo, Zhenwei Shi
Erschienen in: Intelligent Data Engineering and Automated Learning – IDEAL 2017
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Abstract
Detecting object in remote sensing images remains a challenge due to multi-scale objects, complex ground environment and large image size despite of the fast development of machine learning and computer vision technology in recent years. The primary difficulty lies in the fast and accurate location of candidate bounding boxes from a large-size remote sensing image. In this letter, we propose a novel remote sensing object detection method inspired by the recent-popular technique, Object Proposals, to quickly generate high-quality object bounding box locations in remote sensing images. A simple but effective objectness measurement, based on the image gradients and its variants, is proposed. Moreover, to evaluate the effectiveness of our method, we complete the subsequent detection flow based on the convolution neural networks as a standard detection baseline. Experiments show that our method is able to produce high-quality proposals with a desirable computational speed.