Hanjun Song
cross_1.png

Pedestrian Crossing Aid Device

Undergraduate Research Program (URP)

Development of Pedestrian Crossing Aid Device for The Visually Impaired

Summer 2013 - Fall 2013

 

Publication

[C-6] H. Song, P. Wachirawit, “Pedestrian crossing aid device for the visually impaired,” in Young Researchers Consortium, 2014 International Conference on Computers Helping People with Special Needs, Paris, France, Jul. 2014. [pdf] [video]

Demo Video

Motivation

Sound generating systems are installed only 7.2% of crosswalks in South Korea and 46.9% of those lack of regular maintenance, possibility of safe crossing is remarkably low for the visually impaired.

 
 

Purpose

This research aims to establish a real-time robust visual object detection algorithm and develop a pedestrian crossing aid device based on that algorithm.

 
 

Concept Design

 
concept.png
 
 

Development Process

 
 

Pedestrian signal detection and tracking

 
 
Flowchart of detection and tracking system

Flowchart of detection and tracking system

Visualization of detection process

Visualization of detection process

  • Detection module is composed of Color-based segmentation and machine learning recognition.

  • Continuously Adaptive Mean Shift is used for tracking module.

 
 

Color Interpretation

Flow diagram of color interpretation

Flow diagram of color interpretation

 
 
Visualization of color interpretation

Visualization of color interpretation

  • Cascade color segmentation is used for speed and reliability.

  • Window setting method is used to shorten the processing time.

 
 

User Interface

 
 
Finite state machine

Finite state machine

 
 

  • System input: presence of the pedestrian signal and its color

  • System output: guiding voice

 
 

Hardware Implementation

r2.png
 
 

Result

  • Total Ground Truth: 1850 images

  • Detection Rate = 0.7735 (FAR = 0.0124, FPR = 0.2189, Specificity = 0)

  • Color Interpretation Accuracy = 0.9622

  • Overall system accuracy = 0.7443