Fruit grading, value chain, AI and human perception

The correct classification of fruits is essential to add value to the production chain and maximize the economic performance of producers. Grading is a critical task in the fruit supply chain, helping to ensure that produce is sold at fair prices and that customers receive high-quality produce that meets their needs.

The classification of fruits refers to the separation of the same according to their quality, size, weight, maturity, color, defects, organoleptic qualities and other relevant criteria. The classification is usually carried out automatically with the use of grading machines, designed to identify and separate the fruits according to their characteristics.

The correct classification of fruits adds value to the production chain in several ways:

     1. Improves the quality of the final product: Sorting helps separate low-quality fruits. This means that high-quality products can be sold at a higher price, which increases the value of the production chain.

     2. Reduce waste: The classification helps to identify those fruits that do not meet the required quality standards. These fruits can be discarded or processed for other uses.

     3. Facilitates transportation and storage: Classification helps to separate fruits according to their size and weight and other criteria, facilitating transportation, storage, and internal logistics.

     4. Satisfies the needs of customers: The classification allows producers to offer fruits that meet the specific needs of customers, such as fruits of uniform size for the processing industry, packaging machines, juice extractors, adaptation to markets of different economic level. This increases customer satisfaction and brand loyalty.

The human brain is capable of classifying a fruit as «best» based on its visual characteristics, such as color, shape and size. This ability stems from the brain’s ability to process visual information and compare it to previous experiences, allowing the brain to make quick and accurate decisions about the quality of food. It is a strategy similar to that used in autolearning algorithms.

The classification of a fruit as «best» is based on a series of visual characteristics that are common to many types of fruit. For example, high-quality fruits tend to have a uniform, symmetrical shape, vibrant color, and a smooth, unbruised texture. These characteristics are processed by the brain through visual perception and are used to make decisions about the quality of the fruit.

However, the classification of a fruit as «best» based on visual characteristics does not always correspond to reality. Often the quality of a fruit depends on other factors, such as ripeness, freshness and variety, which may not be obvious to the naked eye. For example, a fruit may have an attractive appearance, but may be overripe or taste bland due to a lack of nutrients.

In addition, the perception of the quality of a fruit can be influenced by subjective factors, such as previous consumer experience and cultural expectations. For example, one consumer may prefer a riper or sweeter fruit, while another may prefer a firmer or more acidic fruit.

In general, the classification of a fruit as «best» based on visual characteristics is a complex process that depends on many factors. Although visual perception is an important factor in evaluating fruit quality, it is important to consider other factors, such as ripeness, freshness, and variety, to obtain an accurate evaluation of fruit quality.

The human brain is capable of classifying a fruit as «better» based on its visual characteristics, but this classification does not always correspond to reality. The quality of a fruit depends on many factors, and it is important to take into account factors other than visual perception in order to obtain an accurate evaluation of the quality of the fruits.

Automatic fruit sorting based on artificial intelligence (AI) is a technique increasingly used in the fruit supply chain, which offers a number of advantages and some disadvantages compared to sorting carried out under classic deterministic strategies. Automatic sorting uses computer vision and machine learning technologies to identify and separate products based on characteristics such as size, color, shape, or defects in an attempt to try to mimic the psychological processes a human uses in their strategy. decision. Next, we will analyze the points for and against this technique, as well as its similarities and differences with respect to other more classical approaches.

Advantages of AI-based automatic classification

    1. Greater stability in the selection criteria: Once the system has been trained under the specific needs of the client, the selection criteria remain constant over time, this helps to establish a constant quality and selection criteria.
     
    2. Higher accuracy: Automatic classification can be more accurate than classification done by other methods as self-learning algorithms can identify small differences in product characteristics that may go unnoticed by a human.

    3. Simplicity of use: All the complexity of adjusting the decision pattern is included in the training of the neural network that will establish the decision criteria. The user will not have to periodically readjust the system, the reinforcement of the decision network is constant, progressive and in real time.

Disadvantages of AI-based automatic classification

     1. Cost: Automatic sorting systems can be expensive to implement, which can limit their accessibility for smaller producers and traders. We work in this line to reduce costs and have a universally applicable tool.

    2. Need for training and maintenance: Self-learning classification systems require regular training and maintenance. We have easy-to-use tools to implement this work.
     
    3. Limitations in the capacity of machines: Machines may have difficulty sorting products that have unusual or changing characteristics, which may limit their ability to adapt to new and unfamiliar situations.

In the end, the automatic classification methods will try to replicate the internal decision processes that a human would take, there are similarities and differences that can be important but that over time, will become narrower:
     
    1. Both use classification criteria based on visual characteristics.
     
    2. Automatic classification is faster and more accurate than human classification, but may be limited by the ability of machines to adapt to new and unfamiliar situations.
     
    3. Human sorting can be more flexible and better adapt to new and unfamiliar situations, but it can be slower and more error-prone.

AI-based automatic fruit sorting offers a number of advantages and disadvantages compared to sorting done by a human or by a system based on a classic decision strategy. While AI-based classification may be closer to a human’s decision, it may be limited by the ability of machines to adapt to new and unfamiliar situations. On the other hand, the classification carried out by a human can be more flexible and better adapt to new and unknown situations, but it is infinitely slower, prone to errors and lacks stability in the selection criteria. Ultimately, the decision to use an AI-based strategy or sorting done under a more classical strategy will depend on the specific needs of the supply chain and the resources available to implement each method.

At Ingivision SLL we are definitely committed to new AI-based technologies for our highest-performance system, IngiGrader. AI-based technologies are rapidly evolving and the demand for AI-based fruit sorting solutions is increasing exponentially.

Our bet years ago was for this new approach to classification systems based on AI and we tried to capitalize on all our experience to position ourselves as a market alternative with the greatest technological development.

Many satisfied customers endorse our work.