1. Introduction
The Milesight VMS Enterprise already has AI capabilities enabled.The intelligent attributes of humans and vehicles can still be recognized and alarmed by the server-based AI video analytics, regardless of whether the front-end IPC is AI or non-AI. The significance is that it is compatible with a large range of third-party cameras and that it can be easily and quickly integrated into your current project. 
2. The Components of VMS-end Smart Analytics
Version | Function | Module |
V1.5.0.0 | Motion Detection, Region Entrance, Region Exiting, Line Crossing, Loitering | VCA |
V1.5.0.1 | Cross Line People Counting, Regional People Counting, Group People Counting | People Counting |
V1.7.0.0 | ANPR | ANPR |
V1.8.0.0 | Video Synopsis | Video Synopsis |
3. Composition of AI Analytics Server in VMS Installation Package
Version | Installer | Description |
V1.10.0.0 | VMS Enterprise 32-bit or 64-bit | The VMS main program installation package includes the following components: VMS Client (including Video Synopsis), VMS Server, Video Wall, and Analytics Server. Note: Only the 64-bit version of Analytics Server and Video Synopsis is available. |
Operating Requirements Description: ① Video Synopsis includes only the installation files related to the upper-layer application. This functionality can be accessed within the playback interface. When activating this feature, the system will check if the runtime environment meets the specified requirements, which include the presence of an NVIDIA graphics card. It is recommended that the graphics card has a compute capability of 7.5 or above and a minimum of 6GB GPU memory. ② Analytics Server includes only the installation files associated with the upper-layer application. After installation, a desktop shortcut will be created, and running the application will display a tray icon. When launching Analytics Server from the tray icon, the system will verify whether the runtime environment meets the defined requirements. It necessitates an NVIDIA graphics card, preferably with a compute capability of 7.5 or above, and a minimum of 6GB GPU memory. | ||
Run Time(Win64) | The runtime environment for VMS Analytics Server includes algorithm models and libraries related to NVIDIA graphics card's CUDA runtime environment. |
Note:
lCurrently, VMS Analytics Server under Version 1.11.0.2 can only supports NVIDIA graphics cards with compute capability of 6.1, 7.5, or 8.6.
lVersion 1.11.0.2 and above of the VMS Analytics Sever can automatically adapt to GPU computing power, generating corresponding AI models. It supports NVDIA graphics cards with compute capabilities of 6.1, 7.5, 8.6, and additionally supports compute capability 8.9.
lThe number of analytics channels supported by graphics cards with the same computational power is consistent.
lFor more information, please refer to https://developer.NVIDIA.com/cuda-gpus#compute
4. Performance parameters of VMS analytics.
4.1 VCA & People Counting analytics performance
Graphics Card | Analytics Performance | |||
H.264, 512kbps 640*360@25FPS | H.264, 4096kbps 1080P@25FPS | H.265, 512kbps 640*360@25FPS | H.265, 4096kbps 1080P@25FPS | |
GTX1660, 6GB | 16 | 10 | 16 | 10 |
RTX2060, 6GB | 22 | 13 | 22 | 13 |
RTX3060,12GB | 31 | 13 | 31 | 13 |
4.2 ANPR analytics performance
Graphics Card | Analytics Performance | |
H.264, 4096kbps, 1080P@25FPS | H.265, 4096kbps, 1080P@25FPS | |
GTX1660, 6GB | 22 | 20 |
RTX2060, 6GB | 27 | 24 |
RTX3060, 12GB | 30 | 29 |
4.3 Video Synopsis analytics performance
Currently, VMS only supports Video Synopsis for a single channel at a time, utilizing the graphics card's performance to its maximum capacity.
—— END ——