Mastery of the latest technologies in computer vision, neural networks, and graphics accelerators has enabled SITEP to develop an in-house solution that allows it to finely control performance and adapt it to different operational scenarios.
Customization allows the versatility of the system to be exploited for different uses, such as: threat identification and classification, tracking of targets and/or objects
(automatic target tracking), collision avoidance, augmented reality
The integration of mixed GPGPU-FPGA architectures enables the real-time application of DSP algorithms suitable for enhancing the image from video sensors before it is processed by AI or viewed by the end user.
These particular technologies also allow the integration of different types of sensors, for the implementation of multispectral video analysis (in the various IR bands and the visible).
- Low-level video acquisition via frame-grabbers
- Accelerated proprietary modules for video sensor management via FPGAs
- Pre-processing of video for first conditioning
- Accelerated video encoding in popular formats
- Implementation of state-of-the-art algorithms for target/object tracking and identification
- Integrated neural network with possibilities for further customization and optimization
- Customized user interface for system visualization and control
- Integrated diagnostics
- Extensibility and integrability with other systems (e.g., CMS, C&C and the like)
