Skip to content

quickstart.md

Latest commit

 

History

History
210 lines (146 loc) · 7.61 KB

quickstart.md

File metadata and controls

210 lines (146 loc) · 7.61 KB
Raw
title Quick Start: Deployment, Sign-In, and First Detection
description Deploy or connect to CosmoEdge, add a video, configure a scenario task, and verify the first AI detection.
prev
text link
Using CosmoEdge
/en/tutorials/
next
text link
Scenario Task Configuration
/en/tutorials/02-scenario-config/scenario-config

Quick Start: Deployment, Sign-In, and First Detection

Item Details
Who this is for First-time CosmoEdge users, deployment engineers, and developers
What you will accomplish Sign in, add one offline video, assign an algorithm, and observe a detection result
Prerequisites An x86 host with Docker, or an edge device with CosmoEdge preinstalled
Estimated time About 15–30 minutes for a first x86 build; 10–15 minutes for a preinstalled device
Device required Choose an x86 Docker host or a preinstalled edge device; a camera is not required
Final acceptance result The channel is running, the live view shows algorithm output, and matching events can be queried

The goal is not merely to open the console. You will complete one verifiable first detection:

  1. Deploy or connect to CosmoEdge.
  2. Sign in to the console.
  3. Add a test video.
  4. Assign and save an algorithm.
  5. Verify the live result and event record.

Changing the default password, connecting a real camera, and tuning business parameters are important follow-up tasks, but they are not required to prove the first detection.

1. Deploy or Connect to CosmoEdge

Path A: Docker on an x86 Host

Use this path on a Linux x86_64 host. Windows users should use docker-compose.x86.windows.yml; the remaining steps are the same.

git clone https://github.com/cosmo-wander-ai/cosmo-edge.git
cd cosmo-edge
docker compose -f docker-compose.x86.yml up -d --build
docker compose -f docker-compose.x86.yml ps

Windows PowerShell:

git clone https://github.com/cosmo-wander-ai/cosmo-edge.git
cd cosmo-edge
docker compose -f docker-compose.x86.windows.yml up -d --build
docker compose -f docker-compose.x86.windows.yml ps

The initial build downloads dependencies and compiles the application, so its duration depends on the network and host.

Docker building the CosmoEdge service images

Expected state:

  • docker compose ... ps reports the services as Up or running;
  • http://127.0.0.1:8080 opens in a browser;
  • for remote access, replace 127.0.0.1 with the host IP and allow TCP port 8080 through the host firewall.

CosmoEdge containers in the running state

Path B: A Preinstalled Edge Device

  1. Connect the device to power and Ethernet.
  2. Connect the operator computer to the same LAN.
  3. Open the management address supplied with the device.

Expected state: the sign-in page opens. If the device address is unknown, check the delivery label or the router's DHCP leases before considering any reset.

2. Sign In

The default x86 address is http://<host-ip>:8080. Initial credentials are:

  • Username: admin
  • Password: admin

CosmoEdge sign-in page with username and password fields

After signing in, Operation Overview should open. Change the default password after the first sign-in. This is a security requirement, although it does not block this tutorial's first-detection acceptance.

3. Add the Test Video

The repository contains a reproducible test asset:

data/test-video/Safety Helmet.mp4
  1. Open Video Access.
  2. Select Add.
  3. Choose Offline Video as the source type.
  4. Enter a channel name such as “First helmet detection”.
  5. Upload Safety Helmet.mp4, then save.

Adding an offline test video on the Video Access page

Expected state: the new channel appears in the list without a persistent connection or decoding error.

::: tip A camera is optional After the first detection works, replace the offline video with an RTSP/RTSPS network source. Starting with the local asset isolates CosmoEdge from network, credentials, and camera codec differences. :::

4. Create the First Detection Task

  1. Select Allocate Task for the new channel.
  2. Select No Safety Helmet from the available services.
  3. Under Detection Area, cover the relevant part of the frame. For this first check, use the main work area.
  4. Keep the built-in parameter defaults.
  5. Select Save.

Selecting a scenario task for the test video

Saving associates the channel with the algorithm and starts analysis. Return to Video Access and verify that the running switch is enabled.

The channel in the running state after the task is saved

Expected state:

  • the channel is assigned to No Safety Helmet;
  • its running switch is enabled;
  • no persistent start failure is displayed.

5. Verify the Result

5.1 Live Display

Open Live Display and select the channel. Allow time for video playback and model initialization.

Safety-helmet inference overlays in the live view

Pass criteria:

  • the video plays continuously;
  • bounding boxes, labels, or algorithm status overlays appear;
  • the channel can be reopened after refreshing the page.

5.2 Event Center

Open Event Center and query by channel and time. An event is created only when the image matches the detection condition and also satisfies the alarm interval, count, duration, and deduplication rules. “Live overlays are present” and “an event was generated” are therefore separate checks.

Detection records listed in Event Center

6. First-Detection Acceptance Checklist

  • The sign-in page and Operation Overview are reachable.
  • Safety Helmet.mp4 exists as an offline video channel.
  • No Safety Helmet is assigned and the channel is running.
  • Live Display plays continuously and displays algorithm output.
  • A matching segment produces a queryable Event Center record.
  • The deployment method, access address, and verification time are recorded.

The first detection is complete only when every item passes.

7. Troubleshooting

Containers Do Not Start

Check in this order:

docker compose -f docker-compose.x86.yml ps
docker compose -f docker-compose.x86.yml logs --tail=200
docker system df

Verify that Docker is running, disk space is available, and port 8080 is free. See Troubleshooting for the broader diagnostic path.

The Sign-In Page Does Not Open or Authentication Fails

  1. Use host port 8080, not a container-internal port.
  2. For remote access, verify the host IP, firewall, and network segment.
  3. Use admin / admin only for the initial sign-in. If the password has been changed, use the new value instead of rebuilding containers.

The Saved Video Has No Picture

Check file readability → channel state → decoder logs → algorithm state:

  1. Start with the repository MP4 so that RTSP networking is not another variable.
  2. Confirm that upload completed and the channel is enabled.
  3. Inspect logs for file-access, codec, or decoding errors.
  4. If video plays but no overlay appears, confirm that the scenario task was saved and is running.

Live Detections Appear but No Event Is Created

This is usually not a connection failure. Confirm that the clip matches the rule, then inspect the alarm interval, alarm count, detection duration, and stationary-target deduplication settings. The next guide explains these controls.

Next Step

Continue to Scenario Task Configuration to set detection areas, parameters, and alarm-related rules intentionally.