RfVideoTracker lib v3.1 is a software library for automatic tracking of objects on video and calculating their parameters. This software library makes it possible to implement the STOP-FRAME function when capturing objects for tracking, and to compensate time delays in communication channels when transmitting commands for capturing objects. The source code of the software library is written in C++. The software library comes as source code files containing the description of the RfVideoTracker software class of the automatic tracking algorithm, or as compiled files of the software library for inclusion in C++ projects. One instance of the software class makes it possible to implement one tracking channel. To build multichannel tracking systems, several instances of the class have to be created. The software library contains the following files (for the compiled version): RfVideoTrackerDataStructures.h (description of data structures and constants), RfVideoTracker.h (description of the software class of the tracking algorithm) and compiled RfVideoTracker.o (or *.lib) library file.
- MAIN PARAMETERS
- High-speed operation. The processing time for one frame for the tracking rectangle of 128x128 pixels for various platforms: intel i7-4720 (4 core) – 1 ms; intel atom e3950 (4 core) – 4.1 ms; intel i5-7200U (2 core) – 1.9 ms;
- Tracking of objects having the size from 8x8 to 128x128 pixels and more. Tracking of the part of object.
- Tracking of low-contrast objects (from 10%) against a complex background and under interference conditions.
- Tracking of all types of objects of any shape. Adaptability to changing shapes and sizes of objects.
- Tracking when the interference overlaps up to 50% of the area. Automatic detection of tracking failure.
- Dynamic objects tracking. The allowable offset per frame is up to 52 pixels in any direction.
- Calculation of the position and size of the object in the tracking rectangle. Ability to change parameters without resetting.
- Ability to build both single-channel and multi-channel systems (several instances of the class).
- Stable tracking under a continuous changing of the object brightness. Adaptation to a changing background.
- TARGET PLATFORMS AND COMPATIBILITY
- The RfVideoTracker lib v3.1 software library is written in C++ without using third-party software modules and functions dependent on the operating system or processor. The library is compatible with any operating systems that support C++ compilers (C++11 standard) and OpenMP parallelization technologies (standard 1.0). The library is compiled from source codes and can be used with any types of processors (x86, ARM, RISC-V, etc.).
- source code, technical documents