What are the problems?
Big data is a big problem. The amount of available data is essential for decision making in a modern world. But it also means more effort and resources are required to extract what is most useful.
This is particularly important for operational decision making. Full Motion Video (FMV) and other assets are essential for providing operational decision-making information. But, the analysts tasked with extracting this information are heavily burdened by multiple persistent video feeds. While highly trained and very good at what they do, the amount of information an analyst uses is growing and they are often tasked with monitoring video feeds for over 10 hours at a time.
What is C-CORE doing?
C-CORE has developed a novel framework for the automatic detection, tracking, and pattern analysis of vehicles in FMV. The resulting information are stored in a spatio-temporal asset catalog (STAC) for visualization in GIS software.
The detection and tracking of vehicles were based on modern machine learning techniques but were identified as components that will continuously update, upgrade, and become better over time. By ensuring that all components of the framework understand the asset catalog, the framework allows for these components to be changed when needed. It also allows for the detection, tracking, and pattern analysis of other objects by adding these components to the framework. The STAC catalog uses best practices and open source structures to handle data, meaning that integrating into modern or legacy systems is possible.