Mobile Phone Use Detection

 

打电话检测01

Phone Call Detection System

1. Background

In environments such as oil depots, gas stations, and chemical plants, the use of mobile phones can trigger fires or explosions due to the presence of flammable materials. Similarly, in medical facilities, mobile phones may interfere with sensitive medical devices. During driving, phone use can distract operators and endanger lives. Manual supervision is limited and often ineffective—therefore, intelligent video monitoring is essential.


2. Product Overview & Workflow

Product Overview

Fuwei Vision’s AI Phone Call Detection System is an intelligent image recognition solution based on real-time video streams. Using advanced deep learning and big data analytics, it replaces manual supervision by automatically detecting and alerting unauthorized phone usage.
The system utilizes existing surveillance cameras — no new hardware is required — and provides real-time detection and instant alerts for rapid response.

Workflow

  1. Front-end cameras continuously upload live video streams to the central server.

  2. The AI detection engine analyzes the video feed in real time.

  3. When a phone call is detected, the system captures the event, saves evidence, and triggers alerts at the control room or local monitoring points.


3. System Features and Technical Specifications

(1) Phone Call Detection

  • Automatically detects if a person is using a phone (holding a phone to the ear).

  • Supports both indoor and outdoor environments.

  • Detects multiple individuals within the same camera view.

  • High adaptability to various lighting and postures.

(2) Alarm Function

  • Real-time alerts via on-site IP speakers or central monitoring systems.

  • Millisecond-level response time.

  • Customizable alert methods (audio, visual, or mobile notifications).

(3) Technical Parameters

  • Supports RTSP video streams from Hikvision, Dahua, Uniview, and other major brands.

  • Fully automatic video analysis — no manual intervention required.

  • Recognition accuracy: >95%.

  • Detection frequency adjustable based on server configuration.

  • Supports 4 to 50 channels per workstation, depending on hardware performance.


4. Hardware and Operating Environment

  • Supported Cameras: Hikvision, Dahua, Uniview, and other RTSP-compatible cameras.

  • Installation Angle: Camera tilt angle less than 60° from horizontal.

  • Server Configuration: CPU: Intel i7 (11th Gen), Memory: 32GB, Storage: 1TB SSD.
    One workstation can handle up to 32 camera streams, subject to configuration.

  • Operating System: 64-bit Windows 10, Windows 11, or Windows Server 2016 and above.

  • Deployment: Flexible for on-premise or cloud integration.


5. Typical Project Cases

  • Hengqin Macao New Neighborhood General Contracting Project

  • Shenzhen Nanshan Tsinghua Information Port Project

  • Zhengzhou New Era Plaza Project

  • (More cases available upon request)

打电话检测02