Mapping plant growth forms using structure-from-motion data combined with spectral image derivatives.

Area of Science:

• Ecology
• Remote Sensing
• Geospatial Analysis

Background:

• Chaparral shrublands in southern California require effective monitoring for conservation.
• Accurate classification of plant growth forms is essential for understanding these ecosystems.
• Traditional methods can be labor-intensive and less scalable.

Purpose of the Study:

• To assess the effectiveness of plant height data from aerial imagery for distinguishing plant growth forms.
• To compare classification accuracies using spectral data, canopy height models (CHMs), and hybrid approaches.
• To enhance the monitoring of Mediterranean-type plant communities.

Main Methods:

• Derivation of Canopy Height Models (CHMs) using structure-from-motion photogrammetry from high-resolution aerial images.
• Application of a multi-criterion, knowledge-based thresholding approach for classification.
• Utilizing spectral data (NDVI, hue, intensity, focal texture) and CHM data individually and in combination.

Main Results:

• Classification accuracy using spectral data alone ranged from 66.0% to 69.0%.
• Classification accuracy using CHM data alone ranged from 72.0% to 75.5%.
• Hybrid approaches combining spectral and CHM data achieved the highest accuracies, ranging from 80.5% to 82.5%.

Conclusions:

• Combining multi-spectral and canopy height data significantly improves the classification of plant growth forms in chaparral ecosystems.
• Canopy height models derived from aerial imagery are a valuable tool for ecological monitoring.
• This hybrid approach offers a more robust method for characterizing Mediterranean-type plant communities and detecting changes.