Flemart Systems

Convection Task

 Fire is a big disaster, especially for the ship in the sea. Each ship design should be optimized to prevent fast fire speeding in case of disaster. It’s possible to simulate conflagration at the ship, using computers and software developed by our engineers, and make improvements of fire safety at the project level.

    Software does not have complicated user interface, but encapsulates simulation of complicated physical processes and mathematical calculations. Significant scientific work was done by scientists of Belarusian State University to get main calculation equations especially for this project, and was successfully implemented by our developers. The main equations used in the software are governing energy balance equation and convection equation. It calculates air flows which can assist in fire spreading. Flemart cooperation with the best scientist of our country allowed us to build such pretty complicated scientific solution in a short time. 

    The software uses 3D model of the ship built with AutoCAD for visualization and airflow calculations. The main conception is to use any construction 3D model object. Thus 3D model of any ship or building can be tested for fire spreading. Visualization realized using DirectX instruments allows improving performance operating with the big objects of complicated structure, and gives the possibility to apply additional graphical effects if needed. After definition of fire source or sources, software calculates the overall time for complete object destruction by fire and can simulate conflagration in real-time.

Tags: 3D, ATL, COM, DirectX, WTL

Human motion detection system

  Solution based on Intel Open Computer Vision library which provides image recognition tools. However, we were forced to add our innovations for video filtration. The problem is that image recognition does not provide clear results if image affected with optical effects of the water. At the same time, cameras installed under the water get image with patent optical artifacts. So, after video filtration we succeeded to make recognition system working fine. 

The system features algorithms developed by Flemart engineers for the following tasks: 

• Creation of 3D image using the flat pictures from 4 cameras;
• Reconstruction of target location in the space;
• Movement recognition based on Markov chain;
• Water glares recognition and filtration;
• Underwater survey filtration;
• Neuronet building and binding with 3D model.

Our experience can be useful in every possible recognition system, including human faces recognition, text and signs recognition, different optical measurement systems etc…

Tags: 3D, ATL, COM, Intel, Movement Detection, Movement Recognition, Open Computer Vision Library, OPL, Recognition, WTL