TECHNICAL FIELD

The invention is related to an artificial intelligence-controlled food drying unit that monitors the color and volume changes that occur during the drying process of foods (fruit, vegetable, meat, cheese, etc.) in order to automotically apply the gradual temperature change. With this drying unit, it is possible to reduce or limit the color change in the product and the formation of substances that may have negative effects on health, and also, dried food at the same standard quality can always be obtained regardless of the raw material properties.

PRIOR ART

Drying, with its general definition, is one of the most used methods for the food to be consumed during longer periods of the year by removing water from the food. With drying, the resistance of the food to microbial spoilage is increased. In addition, since the weight of the product will decrease after losing moisture, the transportation process becomes easier. Convective drying is one of the most common food drying techniques applied. In this technique, high temperature and low humidity conditions are used. While the ambient temperature is increased, the humidity is decreased as much as possible, thus removing water from the food.

Color change in food of plant (fruit, vegetable) and animal origin (cheese, meat, etc.) and quality problems caused by the increase in the concentration of unhealthy substances (hydroxymethylfurfural, etc.) known to occur during drying are the important quality problems for the dried food industry.

The control of drying rate and weight change in the drying of fruits and vegetables with convective drying applications, which is a known technique, cannot be achieved effectively because the moisture content of each fruit/vegetable is not the same. Therefore, there are problems in determining the drying time or ensuring adequate drying of the food (products with different humidity ratios). These problems result in excessive energy consumption or the production of dried foods that are not resistant to microbiological spoilage.

The hot air used in the drying systems is controlled by the operator. Since this means that the operator adjusts the hot air given during drying process depending on his knowledge and experience, excessive use of energy or operator-related problems in the product cause serious financial losses. According to the experience of the operator, in the known technique, the quality of the final productis directly affected, there may even be differences in the quality of the products that come out of the same production line in different shifts.

At the same time, hot air drying applications have a negative effect on product color, although it is widely used by dried fruit and vegetable producers. As the color of the product is important for the preference of the buyer, the variation in the color of the dried product is one of the important problems of the conventional (present) drying process machineries.

LIST OF FIGURES

Figure 1 : Assembly Overview

The corresponding numbers in the figures are as follows;

1. Drying Unit

2. Camera

3. Camera Housing

4. Temperature Controller Printed Circuit Board

5. Lighting System

6. Mini Computer

DETAILED DESCRIPTION OF THE INVENTION

The invention includes drying unit (1 ), camera (2), camera housing (3), temperature controller printed circuit board (4), lighting system (5) and mini computer (6).

The drying unit (1 ) is an oven in which the drying process is carried out, with a temperature setting range of 20-90 °C. The fruits are dried by placing them on the trays in the oven. It ensures that products in standard quality is obtained by performing uniform heating process.

The camera (2) is the device used for monitoring agricultural products and obtaining images of the products. At least one is mounted on the top of the oven and in places suitable for capturing images. The image output is transferred to the mini computer (6). Image transfer is provided by at least one of the wired or wireless communication protocols. During the drying process, the images of the products are taken continuosly by the cameras, and processed within the artificial intelligence based algorithm by the help of mini computer (6).

The camera housing (3) will protect the camera (2) from high temperature. It is transparent and light-permeable. Thus, camera performance is not affected by high temperature and agricultural product color quality evaluation process is continued without any problems.

The temperature controller printed circuit board (4) is a microcontroller-based board where the temperature control of the drying unit (1 ) is implemented. It makes the temperature adjustment instantaneously and automatically according to the outputs coming from the mini computer (6). It is the module where the temperature is precisely controlled with the feedback received. Thus, the color differences in the final product are autonomously prevented.

The lighting system (5) provides the illumination needed for the camera (2). In this way, the images taken with the camera will be evaluated in a standard way. The quality difference among the products depending on the light intensity will be prevented.

The mini-computer (6) is the unit containing the software in which the image is transmitted and artificial intelligence algorithms are run. It instantly processes the data coming from the camera (2) and the temperature controller printed circuit board (4), instantly monitors the size and color change of the product in the drying unit (1 ), and makes the temperature adjustment accordingly in real-time. The mini computer (6) classifies the dried products in real time, using the pre-trained model file with the data set previously created by taking thousands of product images from the same type of dried food. Accordingly, artificial intelligence classifies the stage of drying in real time. It will also provide human-machine communication, minimizing the influence of the operator on the process and preventing human-related errors.

The product images in the drying unit (1 ) tray are collected by the camera (2). With these images, the images of the raw material, the stages during drying and the final product are classified. The data set is created by the artificial intelligence-based software on the mini-computer (6) from the images that have been classified considering the different drying stages of the prosess. This data set is also enriched with existing data sets and made ready for the use. The size and color characteristics of the products to be dried are obtained by the artificial intelligence-based software on the mini computer (6), with image processing techniques. With image enhancement techniques, raw images are converted into a more suitable form for the measurements to be taken. By using artificial neural networks and deep learning techniques, the features of the prepared data set are evaluated and obtained. By using the classification method, the stage classes of the product is determined, and the temperature adjustment is carried out intermittently or near contiuosly during drying using measured time-temperature values. Product color and moisture content are standardized with the decisions taken as a result of this network, and quality and process are managed. When the product reaches the desired drying level, the oven is turned off and the process is terminated automatically.

For example, the drying scenario is as follows;

• Bananas with similar physical properties that are planned to be dried are placed on the tray.

• It is placed in the chamber inside the drying unit (1 ).

• The drying unit (1 ) and the mini computer (6) are powered.

• The user selects the banana from the predefined product types from the screen of the system.

• By pressing the start button of the system, the drying process is started depending on the product type.

• The software on the mini computer (6) activates the artificial intelligence model created previosly.

• In this process, the lighting system (5) is adjusted by the software on the mini computer (6) to provide the appropriate light for the banana.

• The images of the products on the tray are taken with the camera (6) in real time.

• The software on the mini computer (6) classifies the stage of the drying product and creates a temperature-time plan according to its characteristics (volume, color, etc.).

• The temperature of the drying unit (1 ) is controlled with the temperature controller printed circuit board (4) in accordance with the temperature-time schedule. • When the products on the tray reach the final product stage, the oven is automatically stopped by the smart algorithm of the software on the mini computer (6) and the operator is informed with an alarm.

• The drying process is completed.

The problems that exist in the state of the art are solved by providing the products to dry sufficiently in a short time with the artificial intelligence supported invention according to the dimensions and colors of the food products. With the invention, errors caused by the experience of the operator and the resulting financial losses are eliminated and the need for labor is reduced.