OSINT and Image-Based Intelligence
Open-source intelligence is the collection and analysis of data gathered from open sources to produce actionable intelligence and gain situational awareness. Different types of data originate from sensors, devices, video, audio, networks, log files, transactional applications, web and social media — much of it generated in real time and at a very large scale.
However, image and video analysis is very time-consuming and expensive. Humans cannot monitor and process the existing data-saturated OSINT streams, which remain active around-the-clock, without the aid of artificial intelligence.
Computer vision is a field of artificial intelligence that trains computers to interpret and understand the visual world. Using digital images from cameras and videos and deep learning models, machines can accurately identify and classify objects in images, allowing for an effective and efficient way of categorising and filtering information
Introducing ODCT by HENSOLDT Analytics
Object Detection, Classification and Tracking – ODCT for short – is the new HENSOLDT Analytics product, which combines three important computer vision techniques to heighten the open-source intelligence gathering capabilities of the Media Mining System. They are cornerstones in the development of complex image and video analysis solutions that mimic or even surpass human visual system.
At HENSOLDT Analytics we have a dedicated team working on deep learning computer vision systems and constantly developing models for ODCT. This technology allows our customers to extend their open-source intelligence gathering capabilities to real-time videos, empowering them to make truly informed decisions.
ODCT Use Cases
Object detection, tracking and classification methods are in high demand for vehicle traffic control systems, surveillance systems for detecting unauthorized movements, mobile robot applications, or animal tracking. In the security and defence area, these methods can be used for:
- Area surveillance
- Automatic alarms
- Target detection
- Multiple object detection – the tracking algorithm tracks every single object of interest in the video
- Autonomous Flight control of unmanned air vehicles (UAVs) – surveillance, detection, and capture of drones
ODCT - Definitions
Object detection – the detection of an object of interest in a video is the most important step in a system for tracking and classifying objects. An object can be detected either when it first appears in the video or in every frame, depending on the requirements. Since it is more meaningful to detect the moving objects than to detect all the static objects in a video sequence, most methods focus on detecting such objects.
Object classification – A single frame in a video can have multiple objects that need to be identified. The algorithm learns the features of different objects during a training process and can generalize the feature for the object and detect the object. The classification method is assigning a unique identification for each object (ID).
Object tracking – Object tracking is a deep learning process where an algorithm tracks the movement of an object. In other words, it is the task of estimating or predicting the positions and other relevant information of moving objects in a video.