The SAR Image Storage and Recognition System Based on Cloud Platform

摘要

As the modern battlefield environment has become increasingly complex, the traditional SAR recognition methods are too dependent on the training data source to be robust and universal, which makes it can not meet the demand of modern warfare. So how to automatically interpret the fast increased SAR images becomes an urgent problem. The emergence of big data, cloud computing and deep learning technology makes the automatic and intelligent interpretation of large volume of SAR images become possible. This paper proposes a SAR image storage and recognition system based on cloud platform to automatically obtain and identify all kinds of military targets from a complex scene. The system combines the cloud-based platform and deep learning method to achieve real-time recognizing analyses and batch processing of data. The seamless integration of distributed storage and cloud services meets the needs of large-scale data recognition and management requirements. The assessment shows that the method is more efficient in terms of performance, storage, and fault tolerance.