Seahaven Thermal Synthetic Data Generation Module and Thermal Object Detection Model for U.S. Army Thermal Sensors

To address the U.S. Army's demonstrated need for synthetic data generation capabilities, Lexset proposes an SBIR Phase II effort to develop an Army-specific “Seahaven Thermal Synthetic Data Generation Module.” This module will produce synthetic thermal sensor data that is optimized for integration with the Army’s Machine Learning Operations (MLOps) and Computer Vision (CV) training pipelines, based on critical desired thermal sensor integrations identified by the Army. The Seahaven module will enable the generation of high-fidelity synthetic data tailored to Army-specific operational scenarios, enhancing the training and performance of CV models without reliance on extensive real-world data collection. As a proof-of-concept to demonstrate the utility and benefits of this thermal synthetic data generation pipeline, Lexset will also deliver a bespoke "Thermal Object Detection Model," leveraging state-of-the-art detection algorithms, tailored for thermal sensor collections. This model will autonomously identify objects within raw, collected thermal sensor data, validating the effectiveness of the synthetic data produced by the Seahaven module. By integrating the Army-specific Seahaven Thermal Synthetic Data Generation Module with the Army’s MLOps and CV training pipelines, and proving the superior efficacy of Seahaven’s synthetic data in a sensor-integrated Thermal Object Detection Model, Lexset will provide a scalable, adaptable solution that meets critical operational needs, supporting real-time object detection and tracking capabilities on U.S. Army thermal sensors in varied operational environments.