Automated Grading of Tomatoes Using Artificial Intelligence: The Case of Zimbabwe

Introduction

The rate of production of horticultural produce has been increasing over the past century owing to the increase in the world population. Present world statistics show that tomatoes are the second largest produced fruits by volume and represent up to 100 million tonnes produced on 3.8 million hectares (Montanati et al., 2017; BRESOV, 2018; FAO, 2018; BTGA, 2018). Tomatoes are currently cultivated in 144 countries with China being the leading country in their production. The vegetable has the highest global consumption rate due the fact that most meals include the fruit. Figure 1 shows a tomato and its constituent nutrients (Haifa, 2018)

Figure 1.

Nutritional summary for a tomato

Due to the high rate of production of tomatoes and their perishability it is preferable that immediately after being harvested they are be washed, graded and packed. It is a cumbersome exercise for large-scale stakeholders such as farms and wholesalers to manually sort and grade them. Hence, there is a vital need for machines such as automatic tomato graders to perform this tiresome and repetitive task. Automated tomato graders would group the fruits by size and color; one of the critical stages in the collection and marketing of the fruit (Hernandez-Hernandez et al., 2016). Depending upon the organic attributes of the assortment, the natural product is partitioned into four, five, or more evaluations. Reviewing is manually complete - contrasting with standard examples - or on evaluating machines - as indicated by weight or size. Hand grading is usually combined with sorting or packing; machine grading is done after sorting (Londhe et al., 2013). The machine performs the task at a faster rate with minimum human supervision compared to the manual methods. This helps large farm owners to quickly harvest and dispatch the product to their consumers without delay.

The rate of manually grading tomatoes by small to medium enterprises (SMEs) is very low and inefficient compared to the demand of the fruits both nationally and globally hence there is a need to develop a system of grading tomatoes using artificial intelligence.

Based on this need the principal objectives of this chapter were as follows;

• 1.

  To grade tomatoes using artificial intelligence systematically.

• 2.

  Use machine vision technology to grade tomatoes.

• 3.

  Achieve process efficiency greater than 80%.

According to the authors the benefits of this methodology are manifold;

• 1.

  The machine is environmentally friendly hence it can be greatly appreciated globally,

• 2.

  Reduces the damage of the fruits during the sorting process,

• 3.

  An economy in electricity,

• 4.

  Rapid and efficient,

• 5.

  Affordable to both big and small organizations,

• 6.

  Tomatoes are going to be sorted and packed rapidly before they rot.

Background

Tomatoes are believed to have been originally discovered in South America where they grew wild (Jones, 2017). They were domestically planted by the early explorers with some ornamental curiosities but not eaten. In some parts of Europe, they were grown as decorative plants in the gardens and the plant was regarded to being poisonous. According to Isaac (2015) tomatoes were accepted as a crop in 1840 even though in some countries they were still debating whether the plant was fit for human consumption. Global tomato production drastically increased in the past century after it was proven to be an edible fruit, and could be marketed to generate income.