Overview

Equipment Description and Features

The Csmart-Digit is a digital classifier for raw and/or roasted coffee, designed to establish quality parameters in samples through computerized systems. It extracts relevant information by performing individual analysis of seeds using computer vision algorithms and artificial intelligence models.

The primary goal of the equipment is to verify and analyze the quality of the seeds through statistical sampling of the lot.

Main Features:

• Counting seeds in the sample.

• Classification and distribution by sieves, according to categories and criteria established by relevant standards.

• Identification and percentage calculation of intrinsic and extrinsic defects, based on classification tables for processed coffee (raw seeds), considering defect equivalence and categories.

• Classification of processed coffee (raw seeds) based on defect types or standards defined by regulatory norms.

• Classification of roasted coffee based on defect types or standards defined by regulatory norms.

• Acquisition and storage of individualized seed images for subsequent creation of labeled or unlabeled datasets, used to train and enhance neural networks.

• Extraction of image descriptors (shape, color, and texture) enabling comparisons between analyses, using dimensionality reduction techniques applied to individualized descriptors.

• Generation of PDF reports containing all structured data, as well as data export in XLS, CSV, or JSON formats.

Note: As this is a sample-reading device, the raw material to be analyzed must be properly homogenized and free from stratification to meet the characteristics required for accurate extrapolation of results. The representativeness of the sample must align with the statistical criteria outlined in relevant literature.