F' Andrew Beal

1
AUV 2008 - Vision Subsystem
F. Andrew Beal
fab052000_at_utdallas.edu
Department of Electrical Engineering Erik Jonsson
School of Engineering Computer
Science University of Texas at Dallas Richardson,
Texas 75083-0688, U.S.A.
Project Results
Project Goals

• Basic Algorithms
• Pipe Detection
• Light Detection
• Bin Detection
• Stereoscopic Range Estimation
• C Code
• RGB to HSV Conversion
• Edge Cleanup
• Median Filtering
• Region of Interest Filtering
• Area, Center of Area and Orientation Calculation

• Matlab Algorithms Completed
• C Code for Most Components Completed

Project Overview
int median_filter(int buffer, int length, int
n, int height, int width, int FLG1) if(height
0 width 0 buffer null length
0 n lt2) return null int result calloc
(length, sizeof(int)) int range
ceil(n/2) int w null int h null int x
null int y null int sum null for(x
range x lt (width-range) x) for(y range y
lt (height-range) y) sum 0 for(w x
- range w gt x range W ) for(h y -
range h gt y range h) sum sum
BW(dxdywidth) if(sum gt n n - range
range) resultxywidthtrue else result
xywidthfalse return result

• The purpose of this project was to facilitate an
  autonomous
• underwater vehicle navigate a course using
  vision.
• The course required multiple object detection,
  recognition and
• orientation determination.
• Pipe Detection
• Light Detection

Project Goals

• Complete C Code for All Components
• Implement C Code on Davinci Board using TI
  Libraries
• Layout PC Board for Davinci Chip
• Program Davinci Chip with Final Code
• Test Code in Real Life Situation
• Integrate PC Board with submersible