Vision-based Low Cost Field Demonstrable Paint Restriping ...

[Pages:62]California AHMCT Program University of California, Davis California Department of Transportation

Vision-based Low Cost Field Demonstrable Paint Restriping Guidance System

David C. Slaughter, Nelson Smith, Chris Gliever, Garrett Jones, and Justin Schlottman Biological and Ag. Engineering University of California, Davis

AHMCT Research Report UCD-ARR-01-09-14-02

Final Report for Contract RTA 65A0041

February 2002

This work was supported by the New Technology and Research Program of the California Department of Transportation (Caltrans) and the Advanced Highway Maintenance and Construction Technology (AHMCT) Center at the University of California, Davis.

Copyright 2011, AHMCT Research Center, UC Davis

1. Report No.

2. Government Accession No.

Technical Report Documentation Page

3. Recipient's Catalog No.

4. Title and Subtitle

VISION-BASED LOW COST FIELD DEMONSTRABLE PAINT RESTRIPING GUIDANCE SYSTEM

5. Report Date

February 2002

6. Performing Organization Code

7. Author(s)

David C. Slaughter, Nelson Smith, Chris Gliever, Garrett Jones, and Justin Schlottman

9. Performing Organization Name and Address

Biological and Ag. Engineering University of California, Davis Davis, CA 95616

12. Sponsoring Agency Name and Address

California Department of Transportation New Technology and Research, MS-83 Sacramento, CA 94372-0001

15. Supplementary Notes

8. Performing Organization Report No.

UCD-ARR-01-09-14-02

10. Work Unit No. (TRAIS)

11. Contract or Grant No.

RTA 65A0041

13. Type of Report and Period Covered

Final Report 4/99 to 9/01

14. Sponsoring Agency Code

16. Abstract

This report describes a developmental feasibility study for the automatic guidance of a lane restriping system to automatically apply paint on top of worn traffic lane boundary striping with a lateral tolerance of +/-13mm and a longitudinal tolerance of +/-102mm for dashed lane striping. This system would assist the California Department of Transportation in their effort to enhance safety, reduce worker stress, improve restriping efficiency, and reduce traffic flow impacts of striping maintenance. Machine vision recognition systems employing both hardware and software based neural network lane stripe recognition were evaluated. Preliminary results show potential for automatic machine vision location of worn traffic lane boundary striping, however additional study is needed to fully evaluate the accuracy of this system. A non-contact radar displacement sensor was found to be an acceptable alternative to a traditional ground-driven encoded wheel sensor for longitudinal control of dash length.

17. Key Words

18. Distribution Statement

Lane Striping, Machine Vision, Automatic Guidance, Displacement Sensing, Radar.

No restrictions. This document is available to the public through the National Technical Information Service, Springfield, VA 22161

20. Security Classif. (of this report)

20. Security Classif. (of this page)

21. No. of Pages 22. Price

Unclassified

Unclassified

55

Form DOT F 1700.7 (8-72)

Reproduction of completed page authorized

i

Copyright 2011, AHMCT Research Center, UC Davis

TABLE OF CONTENTS

PAGE

TECHNICAL REPORT DOCUMENTATION (DOT-F-1700.7)

i

LIST OF FIGURES

iii

LIST OF TABLES

iv

DISCLOSURE STATEMENT

v

DISCLAIMER STATEMENT

v

ACKNOWLEDGEMENT

vi

INTRODUCTION

1

OBJECTIVE

3

OUTRIGGER LATERAL POSITION CONTROL SYSTEM

3

PRINCIPAL OF ZICAM OPERATION FOR LATERAL OUTRIGGER GUIDANCE

8

SYSTEM RESPONSE REQUIREMENTS FOR LATERAL OUTRIGGER GUIDANCE 12

ZICAM DEVELOPMENT CHRONOLOGY AND PERFORMANCE EVALUATION 15

DASH LENGTH CONTROL SYSTEM

18

OPERATOR INTERFACE

21

CONCLUSIONS AND RECOMMENDATIONS

22

APPENDIX A ? CAD DRAWINGS FOR ZICAM ENCLOSURE

23

APPENDIX B ? ZICAM MANUAL

33

APPENDIX C ? RADAR SENSOR SPECIFIICATIONS

44

APPENDIX D ? COST BENEFIT ANALYSIS

52

APPENDIX E ? CD-ROM MOVIE OF ZICAM OPERATION

55

ii

Copyright 2011, AHMCT Research Center, UC Davis

LIST OF FIGURES

Figure 1. Caltrans District 10 lane restriping vehicle. Figure 2. Caltrans lane restriping operator adjusting restriping controls while leaning

out vehicle window to observe striping performance. Figure 3. Diagram of lateral position control system for the lane restriping outrigger. Figure 4. Machine vision system mounted inside the white enclosure attached to the

edgeline outrigger on the starboard side of the striping vehicle with the outrigger in the normal operating position Figure 5. Machine vision system mounted inside the white enclosure attached to the centerline outrigger on the port side of the striping vehicle with the outrigger in the retracted (non-operating) position for storage or high-speed transport. Figure 6. Enclosure interior showing the configuration of the two Zicam machine vision systems. Figure 7. View from the rear operator's window showing the port outrigger in operation and shadows caused by trees adjacent to the roadway. Figure 8. An example of a Radial basis function mapping, where (P1, P2) is stored knowledge from prior training and (V1, V2) is a new visual input to be classified by the network. Figure 9. A raw image of a worn edgeline collected from the Zicam is shown in (a), the image in (b) is the result of averaging columns of pixels from (a) to reduce the intensity variation in (a) due to wear and cracks, the graph in (d) is the vertical profile of the intensity in a column from (c), the red line in (b) is the estimated center of the lane stripe using the vertical profile from (d). Figure 10. Spectral reflectance curves for fresh white and yellow paint and worn asphalt. Figure 11. Typical spectral sensitivity of a monochrome CCD video camera. Figure 12. Visual Basic program interface used to determine the relative position error of the outrigger when operated without the automatic guidance system. Figure 13. Lateral position error of the outrigger when the striping vehicle was operated by an inexperienced driver without the aid of the guidance mirror seen in figure 1. An error of zero indicates that the paint nozzle is located directly above the center of the lane stripe. Figure 14. Power spectral density of the outrigger position when the striping vehicle was operated by an inexperienced driver without the aid of the guidance mirror. Continued on page iv

PAGE 1 2 4 5

6

7 7 9

10

11 11 13 14

15

iii

Copyright 2011, AHMCT Research Center, UC Davis

LIST OF FIGURES CONTINUED

Figure 15. Step test response of starboard outrigger, valve offset = 2083, proportional gain =8.

Figure 16. Step test of starboard outrigger showing the effect of changes in proportional and derivative gain levels.

Figure 17. Software interface for automatic recognition of worn lane stripes developed by UC Davis for a standard industrial machine vision computer without the use of the ZISC hardware.

Figure 18. Striping vehicle conducting a test of the longitudinal accuracy of dash line length.

Figure 19. Redesigned operator interface, showing traditional paint valve switches and a new touch screen display.

PAGE 16 17 18

19 21

LIST OF TABLES

Table 1. Precision of longitudinal control of dash and space lengths.

21

iv

Copyright 2011, AHMCT Research Center, UC Davis

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download