Skip to content

analyze

The main command for processing hyperspectral seed imagery.

Synopsis

hyperseed analyze INPUT_PATH [OPTIONS]

Description

The analyze command performs complete hyperspectral seed analysis:

  1. Loads ENVI format hyperspectral data
  2. Applies reflectance calibration with white/dark references
  3. Performs spectral preprocessing
  4. Segments individual seeds
  5. Extracts spectral signatures
  6. Exports results to CSV or HDF5

Arguments

INPUT_PATH (required)
Path to the hyperspectral dataset directory

Options

Output Options

-o, --output PATH (required)
Output file path (.csv or .h5)
--export-plots
Generate visualization plots (distribution, segmentation, spectra)
--export-mask
Export segmentation mask as .npy file

Processing Options

-c, --config PATH
Path to YAML configuration file
--preprocess CHOICE
Preprocessing preset: minimal, standard, advanced, none
Default: standard
--segmentation CHOICE
Segmentation algorithm: threshold, watershed, connected, combined
Default: watershed
--min-pixels INT
Minimum seed size in pixels
Default: 200
--no-outlier-removal
Disable automatic outlier detection and removal

Logging Options

-v, --verbose
Enable verbose logging
-d, --debug
Enable debug mode with detailed logging

Examples

Basic Analysis

hyperseed analyze dataset/sample_001 --output results.csv

With Visualizations

hyperseed analyze dataset/sample_001 \
    --output results.csv \
    --export-plots

hyperseed analyze dataset/sample_data \
    --output seed_analysis.csv \
    --min-pixels 50 \
    --preprocess minimal \
    --export-plots

Using Custom Configuration

hyperseed analyze dataset/sample \
    --output results.csv \
    --config custom_settings.yaml

Disable Outlier Removal

hyperseed analyze dataset/sample_data \
    --output results.csv \
    --min-pixels 50 \
    --no-outlier-removal

Output Files

For input dataset/sample with output results.csv and --export-plots:

  1. results.csv - Extracted seed spectra with metadata
  2. results_distribution.png - Spatial and size distribution
  3. results_segmentation.png - Numbered seed visualization
  4. results_spectra.png - Individual and mean spectra
  5. results_spectra_statistics.png - Statistical analysis
  6. results_mask.npy - Segmentation mask (if --export-mask)

CSV Output Format

seed_id,index,centroid_y,centroid_x,area,eccentricity,solidity,band_1000nm,band_1005nm,...
1,0,234.5,156.2,435,0.34,0.92,0.234,0.237,...
2,1,345.6,234.1,421,0.28,0.94,0.229,0.232,...

Columns: - seed_id: Sequential seed identifier - index: Original label from segmentation - centroid_y, centroid_x: Seed center coordinates - area: Seed area in pixels - eccentricity: Shape eccentricity (0=circle, 1=line) - solidity: Shape solidity (convex hull ratio) - band_*: Reflectance values at each wavelength

See Also