Skip to content

Tackling the challenges of off-road environment navigation through Attention Guided Off-Road Semantic Segmentation.

Notifications You must be signed in to change notification settings

OSSome01/semseg

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

20 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Attention Guided Off-Road Semantic Segmentation

Problem Statement

Navigating in off-road environments presents a significant challenge for autonomous vehicles due to their unpredictable and noisy nature. Conventional methods often fall short in addressing the complexity of these environments, leading to suboptimal performance and safety concerns. Robust and lightweight methods are needed to accurately identify terrains, obstacles, and objects for safe and efficient navigation. This problem is particularly crucial for applications like search and rescue missions, precision agriculture, and autonomous driving.

Project Overview

This project focuses on tackling the challenges of off-road environment navigation through Attention Guided Off-Road Semantic Segmentation. The objective is to develop methods capable of accurately identifying and classifying various terrains, obstacles, and objects for improved autonomous vehicle navigation.

Models Explored

Three different models were compared for semantic segmentation in challenging off-road environments:

  1. U-Net: Baseline model.
  2. Attention U-Net: Model enhanced with a visual attention module.
  3. Segformer: Model chosen for its performance potential.

Inference Outputs

Inference Output

Key Findings

  • The hypothesis that adding a visual attention module enhances the performance of baseline models was validated.
  • Segformer demonstrated the highest Mean Intersection over Union (mIoU), outperforming both Attention U-Net and U-Net.
  • While Attention U-Net performed better than U-Net, the performance gap was smaller than anticipated, prompting further investigation.

Conclusion

In conclusion, this project explored attention-guided semantic segmentation models for off-road environments. The findings supported the effectiveness of incorporating visual attention, with Segformer emerging as the top performer. This research contributes to the development of robust and efficient methods for enhancing autonomous vehicle navigation in challenging environments.

About

Tackling the challenges of off-road environment navigation through Attention Guided Off-Road Semantic Segmentation.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages