perception is important to be fulfilled, so the results
of agriculture and plantations must be adjusted to the
standard (Kurnianto M Wibowo M Hariono B Wijaya
R and Brilliantina A, 2020).
2 RELATED WORK
An image is a two-dimensional Value function f (x,y),
where x and y are spatial coordinates and f at a point
(x,y) is the brightness level of an image at a point. An
image is obtained from capturing the strength of the
light reflected by the object. The image as the output
of the recording device can be an analog or a digital
image. Digital image processing is a process that aims
to manipulate and analyze images with the help of
computers. The analysis process in image processing
involves visual perception with input data and output
data obtained in the form of images of the observed
object. Image processing techniques can include
several aspects such as image sharpening,
highlighting specific features of an image, image
compression, and correction of out-of-focus or blurry
images (Chen H, 2021).
The step in digital image processing begins with
capturing or capturing images (image acquisition)
using sensors in the form of cameras, scanners, etc.
Then proceed with the preparation process
(preprocessing), such as the process of changing the
size (image resizing) or quality improvement (image
enhancement) before finally being used for a specific
purpose. The next step is to divide the image into its
constituent parts (segmentation). This process is done
to separate the desired object apart from other objects.
Because the result of the segmentation process is the
boundaries between the object to be observed further
with other objects, it is necessary to make further
observations (representation and description) to show
that the area within the boundary is the actual object
being observed. The last stage of image processing is
recognition and interpretation and interpretation).
Recognition is the process of assigning a label to an
object based on the information provided by its
descriptor. In contrast, interpretation includes giving
meaning to an object, a set of recognizable objects.
Image classification can use various algorithms,
such as KNN (Z E Fitri A Baskara A Madjid A M N
Imron, 2022), or Neural Network (R Y Adhitya A
Khumaidi S T Sarena S Kautsar B Widiawan and F L
Afriansyah, 2020). An artificial neural network
(ANN) is a network of a group of small processing
units modeled based on a human neural network.
ANN is an adaptive system that can change its
structure to solve problems based on the information
flowing in the network. One part that needs to be
considered in using ANN is the selection of training
methods and ANN architecture. An ANN architecture
that is too small will result in poor problem modeling,
while an ANN that is too large will cause over-fitting
and long computation time (Wijaya R Hariono B
Saputra T W and Rukmi D L, 2020).
3 SYSTEM DESIGN
The Sorting Equipment was measured w x l x h =
100cm x 60cm x 100 cm and is made of an aluminum
frame and covered by a red polycarbonate board.
Inside the sorting box, two cameras are pointing at a
melon cup. The melon cup consists of a circular board
connected to a servo motor so that it can rotate the
melon position. It aims to obtain several shooting
angles for the test data. 3 LED lights lead to the melon
container at an angle of 45 degrees with a distance of
+-60 cm. The lamp power supply uses a variable
power supply to regulate the light intensity in the
sorting box. This aims to get the right light for
shooting in an enclosed space. Figure 1 is a sorting
box design.
Figure 1: Sortir system design.
Taking pictures using 2 Logitech C922 cameras
with HD 1080P video specifications at 30Fps/720P at
60FPS. 1 camera to take side view pictures, and
another to take top view pictures. An ATMega328
microcontroller is also connected serially to the
computer to adjust the servo working angle on the
melon dish. The computer for image processing has
1GB of Intel Core i5 graphics specifications and 8GB
of RAM. Figure 2 is a system block diagram of the
sorting system used on this paper.