Face Recognition Terminal VF105

VF105 integrates face recognition, temperature, card and QR reading plus ID card reading, with RS485, Wiegand, Ethernet and WiFi (4G optional). Face verification is on-device; the face database is stored locally and outputs the person ID on success; card/QR can be configured for transparent output. It features an 8-inch touch screen, real voice prompt and reserved microphone, and supports protocol-mode development.

Key features: ≥97% accuracy at full database, < 0.3s/person; recognition distance 0.3–1.5m, up to 50000 face templates, 50000 cards/QR and 50000 records; mask detection and silent liveness anti-spoofing against photo/video/mask attacks; TCP/IP, MQTT, HTTP; IP66 (non-4G non-temperature version).

Front elements: IR camera, RGB camera, white fill light, IR fill light, 8-inch HD touch screen, QR module and card module. Mounting: turnstile, 86-box, flat-wall and pole.

Avoid backlight, direct sunlight or close strong light on the reading area, which degrades recognition and interferes with the temperature module.

Linux OS, 8GB storage (16G/32G expandable), ARM Cortex A7 MP2 1GHz; 8-inch touch screen 800×1280. Comm: 10/100M port, 2.4G WiFi, RS485, Wiegand 26/34, 4G optional; relay 30V/1A, 2 alarm inputs. Power DC 9–24V (12V recommended), max 16W. See the Specs card on the right.

Transparent mode: face matches local DB only, card/QR output directly with no auth flow. Protocol mode: requires a host receiving data via 485/TCP/HTTP and returning results (protocol doc on official site). Dev mode: currently card read/write only. Whitelist mode: card/QR/face pass after local whitelist match.

After connecting, the screen shows network icon (top-right), device IP (bottom-right), time (top-left) and SN (bottom-left). Put PC and device on the same network, browse http://deviceIP:8060/ to enter the backend; default user admin, password password. After login set the device config password for mode switching, upgrade and reset.

Scenario 1 (online, face-only pass): backend → add person → upload photo or capture → set transparent mode. Scenario 2 (offline): factory default transparent mode, just capture to enroll. Scenario 3 (whitelist): add person ID and validity in whitelist settings, then QR/card opens.

Basic settings: brightness, auto screen-off, screensaver, logo/screensaver image, volume, auto time/zone. Voice & output: per-event voice, relay, 485 and Wiegand for face/card/QR. Face config: enroll/recognize threshold, liveness, mask detection, distance. Auth mode: recognition type (face/card/code), card/face work mode (transparent/protocol/dev/whitelist) and online verification channel. Whitelist, firmware, personnel and records are all managed in the backend.

Enrollment photo: 80–512px, 50–300K, 512–800 wide / 640–1024 high, JPG, natural photo without filters/occlusion. Unused pins must be insulated; short-circuit damage from missing insulation is user responsibility. Avoid backlight and strong direct light; strong light interferes with the temperature module.