Enhancing Lettuce Classification: Optimizing Spectral Wavelength Selection via CCARS and PLS-DA

This dataset has been utilized in a manuscript titled Evaluating and Optimizing Wavelength Selection Using the CCARS Algorithm and Spectroscopy Signals for Lettuce Classification Leveraging PLS-DA Model. The study leverages this detailed spectral data to evaluate and optimize wavelength selection and to classify lettuce health using advanced analytical models, such as PLS-DA, supported by the CCARS algorithm for enhanced signal processing and feature selection.

It comprises spectral measurements collected from lettuce plant leaves, with each row representing a distinct reflectance reading taken during a specific acquisition session. The structure is organized with the following key attributes:

Date & Day. The Date column indicates the exact day of measurement, while the Day column may represent a sequential experimental day or a relative timeline indicator. These two columns are highly correlated.

 Acquisition. This column records the acquisition number, signifying the instance of measurement sessions performed on a given day. It helps differentiate multiple sessions that might occur within the same day.

Position. Each plant is assigned a unique identifier in the Position column, ensuring that measurements can be traced back to individual subjects. This is essential for analyzing variability across different plants.

N. The N column indicates the number of samples recorded during a single acquisition on the same plant, providing insight into the sample size and measurement frequency.

Wavelength. Spectral data is recorded at various wavelengths, and this column specifies the wavelength at which each reflectance measurement was captured, expressed in nanometers (nm).

Reflectance. This column holds the actual measured reflectance values, which are the core data points used to analyze the spectral characteristics of the lettuce leaves, expressed as absolute values.

Class. The binary Class column categorizes the plant's health status, as Healthy or Unhealthy, enabling further analysis or predictive modeling based on spectral data.