Abstract

The complexity inherent in the study of oceanic particulates is compounded by particle collection devices which fail to preserve the integrity of individual particles (e.g., disrupt or aggregate particles). This results in apparent particle sizes and shapes which are inconsistent with the dynamics and optical properties of the original suite of particles. Hence, it is desirable to develop techniques to identify particles and quantify their optical characteristics in their natural environment with minimal disruption. Any methodology for the identification of marine particles necessarily involves the capture of particles (or their images) and then the successful extraction of information which is not ubiquitous and therefore useful for discrete particle identification within some defined classification system. In our effort toward in situ automatic (machine vision) marine particle classification, we are examining the relationship between apparent optical density of the target particle and its observed/calculated dynamic density, forward scattering and backscattering as functions of particle size and relative index of refraction, laser stimulated fluorescence (LSF), and 2-D spatial recognition.

© 1989 Optical Society of America