SVC (Stereo Vision Core) is a stereo vision engine for real-time tridimensional data extraction. It exploits a technology based on an advanced correlation algorithm in FPGA to recover the tridimensional structure of the scene observed by a couple of cameras. SVC is provided as FPGA IP core.
| Computational stereopsis means processing a pair of images coming from two cameras (watching the same scene from two different points of view) to obtain depth information. Range and accuracy measures depend on both cameras resolution, baseline, and focal length. Thus, the stereo setup must be configured according to your specific application needs. Depth maps are typically obtained using correlation. Getting a dense and accurate depth map is a very high computational cost task. SVC makes the process directly in hardware, parallelizing the computation and reaching a frame rate greater than 150 fps at 256x256 pixels resolution. | ||
SVC is designed for applications requiring robustness and high performance in terms of both quality and speed. Typical scenarios for industrial automation are grasping and positioning of pieces through robot arms, or autonomous guided vehicle (AGV) navigation. Other possible application scenarios are in automotive, domotics, building automation, security, and surveillance. KEY FEATURES
AVAILABLE SOLUTIONS
| ![]() | |
![]() | ![]() | |
For a deeper analysis consult the technical characteristics page:

