The Effects of GLCM parameters on LAI estimation using texture values from Quickbird Satellite Imagery

Area of Science:

• Forestry
• Remote Sensing
• Image Analysis

Background:

• Spectrum saturation limits Leaf Area Index (LAI) estimation accuracy in forests with LAI ≥ 3.
• Grey Level Co-occurrence Matrix (GLCM) texture analysis offers potential for improving forest structure estimation from satellite imagery.

Purpose of the Study:

• To conduct a sensitivity analysis of GLCM parameters for optimizing forest LAI estimation.
• To identify optimal texture features and parameters for estimating broad-leaved forest structure using high-resolution imagery.

Main Methods:

• Calculated GLCM texture parameters (ASM, ENT, COR, CON, DIS, HOM) from Quickbird panchromatic imagery.
• Evaluated texture features across four orientations, seven displacements, and seven window sizes.
• Performed sensitivity analysis to determine the influence of each parameter on LAI estimation accuracy.

Main Results:

• An orientation of 90° and a displacement of 3 pixels were optimal for black locust forest LAI estimation.
• A 3x3 moving window size yielded the highest adjusted r² for ASM and ENT.
• Orientation significantly influenced the window size dependency of CON, COR, DIS, and HOM.

Conclusions:

• Optimal GLCM parameter selection is crucial for accurate forest structure estimation, especially when dealing with spectrum saturation.
• This research provides valuable guidance for parameter selection in texture-based forest remote sensing applications.