| Object recognition supported | Person, helmet, no helmet and safety vest |
| Camera supported | LILIN cameras (recommended) and ONVIF cameras |
| Supported AI engines | Safety management |
| AIDA computing power capacity | AIDA power 1 / ch |
| Rrecognition speed (per server) | 6 recogs / sec / ch |
| Maximum objects recognized in a frame | 24 objects / frame |
| Zone configuration | 4 zones |
| Color recognition | Supported |
| AI recognition resolution requirement | Object 120x120 pixels |
| Object size filter | Yes |
| SDK | Yes |
| Video resolution | 1920 * 1080 or lower |
| Interface Language | English / Traditional Chinese / Japanese / Italian |
| User management with access rights | Three levels |
| Application | Public safty |
| Recording (optional) | Navigator Corporate |
| Behavior counter | 64 |
| Behaviors Detection | Helmet
No helmet
Safety vest detection |
| NAV statistics report (optional) | Vehicle flow, vehicle types, people counting, behavior counting, and color counting |
| NAV statistics report type (optional) | Web based pie chart, bar chart, and line chart JPEG snapshot supported |
| NAV statistics report time range (optional) | 7 day, 30 days and 60 days |
| NAV statistics summary (optional) | Camera, object classification, or behavior summary report |
| NAV counter (optional) | Support tripwire mode and object tracking |
| HTTP Post notification | LILIN smart IO box, LILIN LED display, third-party network device, or cloud |
| HTTP/HTTPs push notfication with JPEG snapshot | 16 |
| Support Line notfication | Via IFTTT |
| Digital output | Controllable via IP cameras or LILIN Network I/O Box |
| Car park counting for multiple cameras | LILIN LED display |
| Building capacity counting display | LILIN LED display |
| Suggested CPU | Intel® i3 or above |
| GPU/VPU supported | Intel® HD Graphics 630 or above
PMH-XM2280
Intel® Movidius™ NCS2
NVidia 1050 4G or above |
| System memory | RAM 8G (recommended) |
| OS requirement | Windows 10 with DirectX 12 installed |
| Warranty | 1 Year |
