Database
Purpose
The database serves as the core for:
- Storing pre-processed FT-IR spectra of known pesticides.
- Enabling quick lookups and ranked results based on spectral similarity metrics.
Structure
The database is implemented using SQLite for local usage and simplicity. Its main table includes:
| Column | Description |
|---|---|
id |
Unique identifier for each compound |
compound_name |
Common name of the pesticide |
wavenumbers |
Array or string representing spectral x-axis |
absorbance |
Corresponding absorbance or transmittance values |
💡 The structure may expand to include more fields as needed for future features.
Population
All spectral data was obtained from a free and open-access public database. Each spectrum goes through the following preprocessing pipeline before insertion:
- Baseline correction using airPLS.
- Normalization of transmittance values.
- Band filtering to retain only relevant peaks.
These steps ensure consistency across the dataset and improve classification performance.