Artificial Intelligence and Quantum Computing: Early Innovations. Volume 1. Studies in Computational Intelligence,

Abstract

Currently, the agricultural industry exports a large number of fruits and vegetables, which require a good classification and selection for quality control. This process is commonly carried out manually, where workers visually verify if the product has defects and the appropriate size, shape, and color for its final packaging. Although it is carried out semi-automatically in some industries, this is the most traditional way of exporting in the country. The aforementioned entails a considerable duration in production and the risk of not meeting quality specifications. Therefore, this work proposes the design of an efficient system that is capable of automatically selecting and validating agricultural produce in a more controlled manner. In this research, we focus on the inspection of cucumbers using computer vision techniques for the classification of shape, size, and color in an automated manner. Experimental tests are carried out with artificial intelligence tools. These experiments are carried out on-site using a single high-range industrial camera. The proposed system is evaluated using objective metrics yielding high efficiency for practical applications in agricultural produce inspection.