Related Experiment Video
Updated: Aug 27, 2025
06:25
Stereoacuity Improvement using Random-Dot Video Games
Published on: January 14, 2020
14.5K
Improvement of AD-Census Algorithm Based on Stereo Vision
Yina Wang1, Mengjiao Gu1, Yufeng Zhu1
1College of Mechanical and Electronic Engineering, Nanjing Forestry University, Nanjing 210037, China.
Sensors (Basel, Switzerland)
|September 23, 2022
Summary
This study improves the AD-Census algorithm for lunar obstacle detection. The enhanced method uses average window pixels and adaptive area growth, improving accuracy and speed for reliable lunar exploration.
Area of Science:
- Computer Vision
- Robotics
- Space Exploration
Background:
- Lunar surface imaging faces challenges like low light, similar colors, and noise.
- Traditional Census and AD-Census algorithms struggle with noise and fixed window sizes, impacting accuracy and speed.
Purpose of the Study:
- To enhance the AD-Census algorithm for more accurate and efficient lunar obstacle detection.
- To address limitations of noise susceptibility and fixed window matching in existing algorithms.
Main Methods:
- Introduced an improved algorithm calculating average window pixels to mitigate noise effects on central pixel values.
- Proposed an area growth adaptive window matching strategy to overcome fixed rectangular window limitations.
- Integrated these improvements into the AD-Census algorithm.
Main Results:
- The improved algorithm demonstrated more apparent object contours and significantly enhanced image edges in disparity maps.
- Achieved an average runtime improvement of 5.3% and superior matching accuracy compared to the traditional AD-Census algorithm.
- Effectively detected lunar surface obstacles in a simulation environment.
Conclusions:
- The improved AD-Census algorithm offers enhanced accuracy and efficiency for lunar obstacle detection.
- This advancement holds significant practical value for improving the feasibility and reliability of lunar exploration missions.
