Advanced Computational Methods

Enhancing FragTrack and ParticleTrack with Machine Learning and Computer Vision for High-Speed Video and Test Data Analysis

 
 
 

High-Speed Video Analysis

Deriving meaning from high-speed video (HSV) test observations is often a labor-intensive, tedious task. Furthermore, high-speed video observations are often used in environments with high-levels of obscuration (testing energetic materials, for example), which make object classification and tracking difficult.

New machine learning techniques in computer vision have been recently been developed, and proven successful in reducing noise, obscuration, and artifacts when enhancing images. The Synthetik team are leveraging these newly developed methods to enhance Protection Engineering Consultants' (PEC) FragTrack and ParticleTrack codes

Computer Vision and Machine Learning

Synthetik Applied Technologies is working with PEC to develop and implement new methods of image and video de-noising, classifying, tracking and analyzing fast-moving objects in high-speed video using advanced computer vision (CV) and machine-learning (ML) techniques.

These new developments are showing improvements in object detection rates of up to 60%, and are being integrated into PEC's advanced FragTrack and ParticleTrack high-speed video analysis codes.

Machine-learning and computer vision-enhanced high-speed video analysis, object detection, classification and tracking methods implemented in PEC's   FragTrack   software. TOP: Original detection and classification algorithm; BOTTOM: enhanced detection algorithm.

Machine-learning and computer vision-enhanced high-speed video analysis, object detection, classification and tracking methods implemented in PEC's FragTrack software. TOP: Original detection and classification algorithm; BOTTOM: enhanced detection algorithm.