跳到内容

语言

Currency

Codex Helps Automatically Configure the WSL Version of SpireCV-Pro: An Edge Real-Time Perception Platform for Intelligent Unmanned Systems

by 舒大军 16 Jun 2026 0 条评论

SpireCV-Pro is a real-time perception development platform designed specifically for intelligent unmanned systems, mobile robots, and edge AI devices. It does not simply package algorithm models into an SDK. Instead, it breaks capabilities such as cameras, video streams, LiDAR, object detection, segmentation, tracking, gimbal control, result visualization, video recording, streaming, ROS2 forwarding, evaluation, and data iteration into nodes that can be freely combined across multiple platforms, including workstations, NVIDIA Jetson NX, Rockchip, and Ascend. This allows developers to build their own perception workflows as if assembling building blocks.

In one sentence: SpireCV-Pro helps visual perception algorithms move from demos that merely run to engineering systems that can operate reliably in real unmanned systems, support cross-platform deployment, and keep evolving over time.

01

Let Codex Automatically Configure SpireCV-Pro

In Codex on Windows, enter:

Help me install the WSL2 version of the Ubuntu 24.04 subsystem on this machine, then enter the WSL subsystem and configure CUDA and PyTorch.

Next, tell Codex:

In this environment, help me install SpireCV-Pro. Refer to the Wiki: https://spireai.top/wiki. Use SpireMS version 0.8.7 and SpireCV-Pro version 0.2.5, which is the latest version at the time of the article.

02

Demonstration Results

(1) Integrated QR Code Detection

Run the command:

  1. smsrun viddet_landing_pad_fei

 

The result is shown below:

(2) Click-to-Track Object Tracking

Run the command:

  1. smsrun vid_object_click_tracking_cuda
  2. # Left-click the target to start tracking; press ESC to return to detection mode.

 

The result is shown below:

(3) Underwater Image Enhancement

Run the command:

  1. smsrun vid_underwater_enhancement_cuda

 

The result is shown below:

(4) 3D Object Detection from Point Clouds

Download the dataset velodyne_0001.zip, extract it, and then run the commands:

  1. # Open command-line window 1 and enter:
  2. # Modify /home/amov/Downloads/velodyne_0001 to the actual path on your machine.
  3. # repeated controls whether playback loops continuously.
  4. smsrun pbinreader bin_dir=/home/amov/Downloads/velodyne_0001 rate=10 auto_next=1 repeated=1
  5. # Open command-line window 2 and enter:
  6. # Start the detection algorithm.
  7. smsrun pptpillarscuda

 

The result is shown below:

Note: The visualization result needs to be viewed with Foxglove, and smsfox must be enabled.

03

References

· Official SpireCV-Pro Wiki: https://spireai.top/wiki

· Tutorial for using SpireCV-Pro for free under WSL

· Using smsrun to manage node execution: https://spireai.top/wiki#/如何使用smsrun管理节点运行.md

· SpireCV-Pro source repository: https://gitee.com/spirecv/spirecv-pro.git

发表评论

All blog comments are checked prior to publishing

感谢您订阅

This email has been registered!

Shop the look

Choose Options

Recently Viewed

Edit Option
Back In Stock Notification
Terms & Conditions
Please review AMOVLAB's published policies before submitting product inquiries or placing an order. Policy details are available from the footer links and checkout flow.
this is just a warning
登录
Shopping Cart
0 items