ewenwan/mvision

★ 8,630C++Audience · researcherComplexity · 1/5Setup · easy

Mindmap

mindmap
  root((repo))
    Computer vision
      Object detection
      Scene segmentation
      Object tracking
      Depth estimation
    SLAM
      ORB-SLAM2
      LSD-SLAM
      Map building
    Autonomous driving
      Environment perception
      Path planning
      Collision avoidance
    Resources
      Research papers
      Course slides
      Code links
      Chinese tutorials

mindmap root((repo)) Computer vision Object detection Scene segmentation Object tracking Depth estimation SLAM ORB-SLAM2 LSD-SLAM Map building Autonomous driving Environment perception Path planning Collision avoidance Resources Research papers Course slides Code links Chinese tutorials

Click or tap to explore — scroll the page freely

Things people build with this

USE CASE 1

Use it as a reading map when learning SLAM systems like ORB-SLAM2, with curated links to papers and code in one place

USE CASE 2

Find curated links to autonomous driving datasets like KITTI and open-source platforms like Baidu Apollo

USE CASE 3

Follow structured notes on object detection and semantic segmentation as a Chinese-language computer vision curriculum

Tech stack

C++Python

Getting it running

Difficulty · easy Time to first run · 5min

In plain English

MVision is a Chinese-language collection of resources, notes, code links, and tutorials covering machine vision and robotics. The title translates to Machine Vision. It is not a single software library with a defined API, but rather a curated study guide assembled by the repository owner, covering a wide range of topics that come up when building robots that can see and understand their surroundings. The content is organized around two main domains. The first is computer vision applied to mobile robots, including techniques for detecting objects in camera images, segmenting scenes into labeled regions (for example, distinguishing road from pedestrians from buildings), tracking moving objects over time, and estimating depth from stereo cameras. The README discusses both traditional approaches and deep learning methods for each of these tasks. The second major domain is SLAM, which stands for Simultaneous Localization and Mapping. This is a class of techniques that lets a robot figure out where it is in the world while also building a map of that world at the same time, using only sensor data. The README links to several well-known SLAM systems such as ORB-SLAM2, LSD-SLAM, and SVO, and includes explanations and code analysis links in Chinese. A significant portion of the content is focused on autonomous driving. The README covers the major technical challenges involved: how a self-driving car perceives its environment, how it plans a path through traffic, how it avoids collisions, and how multiple autonomous vehicles might coordinate at intersections. It references datasets like KITTI (a widely used autonomous-driving research dataset) and open-source platforms like Baidu Apollo. Throughout, the repository links to conference proceedings, university course pages, lecture slides, research papers, and open-source code repositories, almost entirely in Chinese. It functions as a reading list and knowledge map for a Chinese-speaking developer or student learning robotics and computer vision.

Copy-paste prompts

Prompt 1

Using the MVision resource list as context, find the best starting tutorial for ORB-SLAM2 and summarize the key steps to run it on a stereo camera dataset

Prompt 2

Based on the MVision autonomous driving section, list the main perception challenges a self-driving car must solve and link each to a relevant open-source tool

Prompt 3

Using the MVision notes on depth estimation, explain how stereo cameras calculate distance and write a Python OpenCV example to compute a disparity map

Open on GitHub → Explain another repo

← ewenwan on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.