Electronic Thesis and Dissertation Repository

Degree

Master of Science

Program

Geography

Supervisor

Dr. Jinfei Wang

2nd Supervisor

Dr. John M. Kovacs

Joint Supervisor

Abstract

This study integrated multi-temporal, multispectral optical and L-band synthetic aperture radar (SAR) imagery to classify agricultural crops throughout a single growing season in northeastern Ontario, Canada. Various optical and SAR band/date combinations were tested to identify optimal dates and datasets for crop classification at various phenological stages using both object-based decision tree rulesets and traditional per-pixel strategies. Object-based decision tree classification of 2 pairs of SPOT-5 optical and L-Band ALOS SAR imagery yielded crop identification accuracy results comparable with hierarchically masked per-pixel classification, with corn classes regularly achieving high classification accuracies (+90%). Regardless of classification approach, results indicate that at least one complimentary date of optical imagery be used in combination with an mid- season optical/ SAR imagery pair to optimally classify northeastern Ontario agricultural landscapes.

Share

COinS