Good and bad characteristics of molina (Triticum durum)

Data Description for Good and Bad Classification of Triticum durum

The dataset used in this project focuses on the classification of Triticum durum (commonly known as durum wheat) into two categories: good and bad quality samples. The goal of the project is to build a reliable model that can accurately differentiate between high-quality and low-quality grains based on selected features.

The dataset comprises a total of 2000 samples, with an equal distribution of 1000 good and 1000 bad samples. Each sample represents a batch or unit of Triticum durum and is labeled according to quality assessment standards used in agricultural or industrial evaluation.

1. Data Collection

The samples were gathered from different harvests and suppliers to ensure variability in grain quality. Good quality samples represent grains suitable for milling and pasta production, showing desirable traits such as uniform size, color, low impurity content, and good hardness. Bad quality samples may include grains that are damaged, discolored, infested, or otherwise not meeting quality standards.

2. Features

Each sample in the dataset is characterized by several features (depending on available sensors or measurement tools), which may include:

Physical Attributes: kernel size, shape, and weight

Color Features: RGB or HSV color values from grain images

Texture Features: surface roughness or uniformity derived from image processing

Chemical/Quality Parameters (if available): protein content, moisture, and hardness index

Image-based Features: edge sharpness, contrast, or other pattern-based metrics (if machine vision is used)

These features may be numerical, categorical, or derived from digital image processing techniques.

3. Data Labeling

Each sample is labeled either as:

Good (Label = 1)

Bad (Label = 0)

Labeling was performed based on expert evaluation or using existing standards in grain quality assessment (e.g., ISO, USDA grading systems).

4. Data Format

The dataset is stored in a structured format such as:

CSV file: with rows representing individual samples and columns for each feature plus the label

Image folder (optional): if visual classification is performed, images are stored in labeled folders (e.g., /good/, /bad/) and may be linked to metadata in a CSV file

5. Purpose and Application

This dataset is designed to train and test machine learning models for the purpose of automating quality control in grain processing and agriculture. Accurate classification of Triticum durum quality can enhance supply chain efficiency, reduce waste, and improve food product quality.