Question

Which of the following statistical measures CANNOT be computed from the multispectral image histograms?

• A. \( \text{Mean, skewness, kurtosis} \)
• B. \( \text{Covariance matrix} \)
• C. \( \text{Co-occurrence matrix} \)
• D. \( \text{Correlation matrix} \)

Hint:

Remember that histograms show the distribution of values within a single band, while covariance, correlation, and co-occurrence matrices describe relationships between bands or spatial arrangements of pixels.

Answer 1

The Correct Option is C

A histogram represents the frequency distribution of pixel values for a single band of a multispectral image. From individual band histograms, we can compute statistical measures like mean, skewness, and kurtosis, which describe the central tendency, asymmetry, and peakedness of the distribution, respectively. The covariance matrix and correlation matrix describe the statistical relationships between different spectral bands, requiring the joint distribution of pixel values across multiple bands, not just the individual band histograms. The co-occurrence matrix (also known as the Gray-Level Co-occurrence Matrix or GLCM) analyzes the spatial relationship between pixel values within an image, considering the frequency of co-occurring pixel values at a specific offset and angle, which goes beyond the information present in individual band histograms.