Removing glint with video processing to enhance underwater target detection

摘要

Remotely sensed imagery of large bodies of water is often dappled with bright patches known as glint. Solar glint is light originating from the sun that reflects off the water surface directly into a camera's sensor. Glint reduces the ability to observe the water at depth, making complicated problems such as in-water parameter retrieval, benthic mapping, and submerged target detection especially difficult. The purpose of this research is two-fold. First, it is hypothesized that the latency between spectral bands on typical pushbroom imaging systems can be utilized to remove glint. The experimental concept of using video and basic image processing techniques is explored using a monochrome camera. Secondly, ongoing efforts are focused on characterizing the key features of glint (size, shape, intensity, and duration) to provide insight for improved glint removal algorithms.