Submarine Imaging Real-time Enhancement (SIREN)

The US Navy is in the process of phasing out conventional optical submarine periscopes in favor of digital sensor masts that transmit video (and other sensor data) to the bridge electronically. Sensor masts can host multiple sensor modalities including both electro-optic (EO) and infrared (IR) for day/night operations. Digital video can be fed directly into a pipeline of automated image processing algorithms including target detection, tracking, and classification for enhanced situational awareness. Unfortunately, the imagery from submarine sensor masts often suffers from a variety of artifacts, including etched reticle lines, interlacing artifacts, and smudging from water droplets. These artifacts affect image quality and negatively impact the performance of downstream processing algorithms. The current method of manual video correction is highly labor intensive and cannot be performed in real time. We propose a Submarine Imaging Real-time Enhancement (SIREN) system to detect and correct these artifacts in real time. In addition to a set of gold standard video enhancement algorithms and novel artifact removal techniques, SIREN features an image analysis module that detects which artifacts are present and automatically applies the correct enhancement algorithms. In this Phase I effort, we will develop a software prototype and evaluate on available periscope video data.