Camera Management

Camera List

The camera list is the first main page shown after login.
It shows camera status, recent detection activity, and monitoring state per camera.
Click a row to open that camera’s detail page.

Screen Structure

The camera list includes the following fields.

Item	Description
Favorites	Mark frequently used cameras
Name	Registered camera name
Assigned Detection Scenario	Scenario currently applied to the camera
Category	Category of the most recent detection event
Recent Detection Alert	Latest detection alert content
Alert Time	Time of latest detection event
External Alert Channel	Teams/Slack/n8n integration status
Monitoring Status	Detection ON/OFF status

• Favorites

  You can bookmark frequently checked cameras.
  Click the flag icon to toggle favorite status.
  Favorite cameras can be filtered via Favorites in the top filter.
  Favorite status is per user and does not affect others.

• Name

  Shows the camera name entered during registration.
  Use names that clearly represent location or purpose.

• Assigned Detection Scenario

  Shows scenarios applied to the camera.
  Up to 4 scenarios can be registered per camera.

• Category

  Shows the category of the most recent event.
  Useful for quick event-type awareness.

• Recent Detection Alert

  Shows latest alert content.
  

  Example:

  • Worker not wearing a helmet.
  • Person entered restricted area.
  • Smoke detected.

• Alert Time

  Shows when the latest event occurred.
  Helps quickly identify recency.

• External Alert Settings

  Shows linked external channels.
  Supported channels: Microsoft Teams, Slack, n8n.
  When configured, alerts are sent via webhook.

• Monitoring Status

  Per-camera monitoring ON/OFF can be toggled.
  Controlled by toggle button.
  
  

  • ON: Real-time analysis is active
  • OFF: AI analysis is stopped (camera info remains, can resume by turning ON)

Camera List Functions

• Camera Filters

  Use top filters to view only matching cameras.
  Useful in large deployments.
  Supported filters:
  
  

  • All
  • Disconnected Cameras
  • Monitoring ON
  • Monitoring OFF
  • Favorite Cameras

• List Sorting

  Camera list is automatically sorted by most recent detection event.
  When a new event occurs, that camera moves to the top.
  

  Example:

  1. Event occurs on Camera A
  2. Camera A moves to top
  3. Event occurs on Camera B
  4. Camera B moves to top

• Camera Search

  You can also search by camera name, scenario, category, or alert keywords.

---

Camera Registration

Camera registration is the first step to start AI detection in EVA.
After registration, EVA receives video, runs detection scenarios, and sends alerts.
Registration has 4 steps.

1. Camera Info
2. Detection Scenario
3. AI Model
4. Detailed Detection Settings

Registration Flow

1. Go to camera list after login.
2. Click + Add New Camera at top right.
3. Open the registration wizard.
4. Complete each step and save.

Step 1. Camera Info

Enter basic connection information.

Fields

• RTSP URL
• Camera Name

RTSP URL

RTSP URL is the streaming address EVA uses.

Example:

rtsp://username:password@192.168.1.100:554/stream

Usually includes:

• Camera account ID
• Camera password
• Camera IP
• Streaming port
• Stream path

RTSP format varies by vendor/model. Check vendor docs or device settings page.

Camera Name

Internal identifier in EVA. Use location/purpose-friendly names.

Examples:

• Main Entrance
• Warehouse A
• Production Line 1
• Parking Gate

EVA does not enforce duplicate checks for camera name/RTSP URL.

Connection Test

Click Apply after RTSP input to run connection test. If successful, you can check:

• Still image
• Resolution
• FPS
• Bitrate
• Other camera info

FPS/bitrate values are measured at connection time and may vary by network/camera/streaming conditions.

Step 2. Detection Scenario

Configure scenarios for situation/object alerts. You can generate by prompt or load a saved template. Up to 4 scenarios can be registered.

• Load Scenario Template: Reuse templates from template settings
• Prompt + Generate: Auto-generate from short requirement text
• Scenario Name: Unique name for identification
• Scenario Category: Applied as alert label/color
• Detection Targets: Add/edit/delete from auto-extracted baseline
• Advanced Settings: Review/edit mode and frame settings

Detection mode examples:

• simple: object-presence centered
• default: step + exception based
• ppe: PPE compliance check
• thinking: risk reasoning with false-positive checks

Frame mode examples:

• single: single image
• sequence: consecutive frame flow
• timeline: time-window frame bundle

Step 3. AI Model

Select an AI model that fits the scenario. You can use AI Model Recommendation for scenario-based suggestions.

Step 4. Detailed Detection Settings

Set AI inference interval.

• Longer interval (for example 120s): saves resources, good for simple object detection
• Shorter interval (for example 10s): better for fast-changing scenes

If interval is too short, resource usage and duplicate alerts may increase.

---

Bulk Camera Registration

In large environments, standardize operational rules before bulk registration. Download the provided Excel template, fill camera name, RTSP URL, and scenario template, then upload. Up to 50 cameras can be registered at once.

Recommended standards:

• Naming conventions
• Default scenario template
• Default model/inference interval
• Default alert channels

After bulk registration, start monitoring gradually from representative groups.

---

Camera Detail Page

Camera detail page consists of streaming view, current detection info, and alert/chat area.

Streaming View

Shows RTSP live video and real-time target detection state based on configured scenarios.

Detection Info Panel

For selected scenario, check:

• Detection situation
• Detection targets
• Category
• Applied model
• AI inference interval

EVA Chat Panel

Receive live alerts and interact with EVA for guidance and limited setting control via chat. You can review alert message/time/image and submit true/false feedback to improve quality.

Example command-based controls

• Adjust target sensitivity: Change sensitivity of target OOO to 0.5
• Change inference interval: Set AI inference interval to 50 seconds
• Start/stop monitoring: Start monitoring / Stop monitoring

You can also view alert history and enable "Show Detection Alerts Only" mode.

Settings Menu

From detail page, manage:

• Camera info
• Detection scenario
• Detection targets
• Detection ROI
• Alert channel settings
• Screen brightness
• Camera deletion