Electromagnetic Systems, Inc.
Small R&D and engineering company (founded 1996, El Segundo CA) focused on algorithm development and software for exploitation of airborne and spaceborne radar data. Core work areas include SAR/ISAR image processing, radar automated target recognition (ATR) using deep learning, radar mode design, radar simulation for ML training data, and embedding/optimizing ML-based ATRs for constrained onboard and cloud processing environments. Primary customers include U.S. government and aerospace organizations.
Industries
N/A
Products
ATR software for operational deployment
Software package implementing CNN-based ATR pipelines that accept sensor-independent complex SAR/ISAR data, perform pre-processing and inference, and output detections and geolocated classifications.
Radar simulation software
Software for generating synthetic SAR/ISAR imagery and labeled targets to support machine learning training and validation.
ATR software for operational deployment
Software package implementing CNN-based ATR pipelines that accept sensor-independent complex SAR/ISAR data, perform pre-processing and inference, and output detections and geolocated classifications.
Radar simulation software
Software for generating synthetic SAR/ISAR imagery and labeled targets to support machine learning training and validation.
Services
AI-based vessel classification from SAR imagery
Classification of maritime vessels using deep learning applied to SAR data for maritime monitoring and analytics.
ATR development for third-party processing environments
Development of ATR software tailored for execution within constrained onboard processors and third-party hosted processing environments, including optimization for compute-limited systems.
Radar simulation and synthetic training data generation
Simulation software and workflows to rapidly generate labeled SAR/ISAR imagery for ML training and evaluation, and methods to reduce sim-to-real bias.
AI-based vessel classification from SAR imagery
Classification of maritime vessels using deep learning applied to SAR data for maritime monitoring and analytics.
ATR development for third-party processing environments
Development of ATR software tailored for execution within constrained onboard processors and third-party hosted processing environments, including optimization for compute-limited systems.
Radar simulation and synthetic training data generation
Simulation software and workflows to rapidly generate labeled SAR/ISAR imagery for ML training and evaluation, and methods to reduce sim-to-real bias.
Expertise Areas
- SAR/ISAR image processing
- Automated target recognition (ATR)
- Deep learning for remote sensing
- Radar mode design
Key Technologies
- Synthetic Aperture Radar (SAR)
- Inverse SAR (ISAR)
- Convolutional Neural Networks (CNNs) / Deep learning
- Sensor-Independent Complex Data (SICD) processing