Automatic Road Feature Recognition and Extraction from Remote

  • Slides: 26
Download presentation
Automatic Road Feature Recognition and Extraction from Remote Sensing Imagery E. F. Granzow Iguana

Automatic Road Feature Recognition and Extraction from Remote Sensing Imagery E. F. Granzow Iguana Incorporated & David Fletcher Geographic Paradigm Computing ----------------

Presentation Overview l Research Context l Basic Approach l The IPaver Toolkit l An

Presentation Overview l Research Context l Basic Approach l The IPaver Toolkit l An Example l Findings and Reflections

Research Context l White paper prepared as resource document for NCRST - Safety, Hazards

Research Context l White paper prepared as resource document for NCRST - Safety, Hazards and Disaster Assessment l Research and software developed as part of NASA supported ARC project “Development and Automation of High Resolution Image Extraction Methodologies for Transportation Features”

Research Context Problem Statement Feasibility of automating extraction of transportation features and potential degrees

Research Context Problem Statement Feasibility of automating extraction of transportation features and potential degrees of automation l Development and use of microcomputer based and specially developed software in component environment l Economic rationale for commercial applications in this area l

Basic Approach Key Concepts l Roadway network considered as single object in feature identification

Basic Approach Key Concepts l Roadway network considered as single object in feature identification process l Image elements regarded in terms of estimated probability of inclusion in solution set l A priori assumption elements are not part of road network object l Approach has roots in pattern recognition and computer vision

Basic Approach Key Concepts (Continued) l Based on user directed iterative application of tools

Basic Approach Key Concepts (Continued) l Based on user directed iterative application of tools l Provides immediate feedback on progress l Scaled for interactive usage

Basic Approach Roadway Network as Object - Benefits Flexible Problem/Image Segmentation l Processing Efficiencies

Basic Approach Roadway Network as Object - Benefits Flexible Problem/Image Segmentation l Processing Efficiencies l Global/Reusable Classification/Processing Model l

The IPaver Functions Toolkit Recomputes DNs based on mean and offsets into n equal

The IPaver Functions Toolkit Recomputes DNs based on mean and offsets into n equal l Merges two images with user specified weightings l Calculates given statistic for specified kernel and replace l Deletes image features based on a combination of size an l

The IPaver Toolkit Functions (continued) Identifies and eliminates tenuous connections between fe l Uses

The IPaver Toolkit Functions (continued) Identifies and eliminates tenuous connections between fe l Uses pixel distance map to develop single string represen l Uses width/member seeds to trace and draw road elemen l

The IPaver Toolkit Support Software Iguana. Space - Implements custom IPaver interface and l

The IPaver Toolkit Support Software Iguana. Space - Implements custom IPaver interface and l Scion. Image - Implements macro procedures to view im l IParse - Tiles images to specified size and overla l Evidence - For known solution reports false and true l

Supports both menu and (Iguana. Space) flowchart access to the IPaver to IPaver Interface

Supports both menu and (Iguana. Space) flowchart access to the IPaver to IPaver Interface l Allows direct editing of each function’s input options and l Automatically logs program states and sequences for revie l Will easily accepts changes/additions to IPaver l

IPaver Interface (Iguana. Space)

IPaver Interface (Iguana. Space)

IPaver … An Example 1 Meter resolution USGS D Residential Area Central Albuquerque Panchromatic

IPaver … An Example 1 Meter resolution USGS D Residential Area Central Albuquerque Panchromatic (0 -255) 1/2 square km

IPaver … An Example Classification by Road Material Type Parameters DN Mean - 135

IPaver … An Example Classification by Road Material Type Parameters DN Mean - 135 Group Interval - 15 Number of Groups - 4

IPaver … An Example Statistical Projection in 3 x 3 Kernel Neighborhood Parameters Statistic

IPaver … An Example Statistical Projection in 3 x 3 Kernel Neighborhood Parameters Statistic - Std Deviation Kernel Size - 3

IPaver … An Example Merging two images Parameters Weight I 1 - 1. 0

IPaver … An Example Merging two images Parameters Weight I 1 - 1. 0 Weight I 2 - 1. 0

IPaver … An Example Deletion with Morphological Constraints Parameters Max Object Size - 300

IPaver … An Example Deletion with Morphological Constraints Parameters Max Object Size - 300 Max H/W Object Size - 1500 Min H/W Ratio -. 8

IPaver … An Example String Filtering Parameters Type - cul-de-sac DN Threshold - 255

IPaver … An Example String Filtering Parameters Type - cul-de-sac DN Threshold - 255 Cul-de-sac depth - 2 pixels

IPaver … An Example Centerline Development from Distance Map Parameters None

IPaver … An Example Centerline Development from Distance Map Parameters None

IPaver … An Example Road Edge Tracing In development … Uses seed to identify

IPaver … An Example Road Edge Tracing In development … Uses seed to identify both “essential line” and road edge Constrains trace based on degree of curvature and aberrant section length Controls degree of deviation between EL and road edge path Uses DM based and “source” image together

IPaver … An Example Superimposition of Solution on Base Image Parameters None

IPaver … An Example Superimposition of Solution on Base Image Parameters None

Findings & Reflections The Evidence Model was developed to measure success in delineating image

Findings & Reflections The Evidence Model was developed to measure success in delineating image elements both within and outside the travelway Solution template was developed by hand for 256 x 256 image thumbnail and compared on pixel by pixel basis to IPaver derived solution Evaluation phrased as true and false postives and negatives

Findings & Reflections The Evidence Model (Continued) image size is '65536' total road rasters

Findings & Reflections The Evidence Model (Continued) image size is '65536' total road rasters '9517' percent of image '14. 5' total true positives '7219' ; total true negatives '53621' ; total false positives '2398' ; total false negatives '2298' ; total efficiency '71. 6' pp/tp nn/tn fp/tn fn/tp '75. 9' '95. 7' ' 4. 3' '24. 1'

Findings & Reflections Project Conclusions - Technical • It is possible to automate portions

Findings & Reflections Project Conclusions - Technical • It is possible to automate portions of the transportation feature recognition and extraction process • It’s feasible to do this without use of “legacy” commercial products (i. e. ERDAS Imagine) and large scale hardware • The probable minimum spatial resolution for IPaver is probably about 1 meter

Findings & Reflections Project Conclusions - Economic • Our original conclusion to not pursue

Findings & Reflections Project Conclusions - Economic • Our original conclusion to not pursue commercial options may be obsolete • New interest and funding for transportation feature and centerline extraction may present new commercial potentials • Changes/evolution of image provider licensing policies have enhanced these potentials • Spaceborne imagery’s near total reliance on defense applications and procurements creates continuing commercial uncertainties

Findings & Reflections Some General Observations • New multi- and hyperspectral high resolution imagery

Findings & Reflections Some General Observations • New multi- and hyperspectral high resolution imagery offers avenues to enhance the extraction process • Urban scenes present greatest challenges due to oblique shadow effects of urban canyons and other urban specific issues • Likely applications are for suburban/rural high growth and unmapped areas