Revista Científica Interdisciplinaria Investigación y Saberes

2023, Vol. 13, No. 3 e-ISSN: 1390-8146

Published by: Universidad Técnica Luis Vargas Torres

How to cite this article (APA):

Perez, J., Robalino, F., Torres, P., López, X. (2023) Free

RGB colorimeter for the sensing and classification of pigmented spheres, Revista

Científica Interdisciplinaria Investigación y Saberes, 13(3) 44-71

RGB free colorimeter for sensing and sorting of pigmented spheres

Colorímetro libre RGB para la sensorización y clasificación de esferas pigmentadas

Julio Perez

Master's Degree, Universidad Técnica de Ambato, Ambato, Ecuador, jperez6656@uta.edu.ec,

https://orcid.org/0000-0002-0119-5570

Freddy Robalino

Master's Degree, Universidad Técnica de Ambato, Ambato, Ecuador, frobalino@uta.edu.ec,

https://orcid.org/0000-0001-8774-4560

Paulo Torres

Master's Degree, Universidad Técnica de Ambato, Ambato, Ecuador, pc.torres@uta.edu.ec,

https://orcid.org/0000-0001-9804-5693

Xavier Lopez

Master's Degree, Universidad Técnica de Ambato, Ambato, Ecuador, mx.lopez@uta.edu.ec,

https://orcid.org/0000-0001-8159-8603

The present work shows the design, development and testing of a

first prototype system for classifying color pigmented spherical

objects. The objective is the analysis of efficiency and error that digital

photoelectric devices can have. Based on the detection of color

patterns called RGB model (Red, Green & Blue) the system is

composed of a digital color detector sensor TCS3472 controlled by

an open source electronic platform Arduino. The ATMega328P

microcontroller processes the data obtained by the sensor, which

then executes the actions responsible for moving the servomotor that

controls the selector mechanism which, in turn, consists of an access

and output channel to three containers. An external control panel with

four push buttons: start, pause, reset and emergency; the latter

configured for the total disconnection of the system in case of

possible jamming failures. Finally, the system delivers the results of

the sphere separation by color on the control panel and physically in

the repositories of the structure.

Abstract

Received 2023-04-25

Revised 2023-06-13

Published 2023-09-07

Corresponding Author

Julio Perez

jperez6656@uta.edu.ec

Pages: 44-71

https://creativecommons.org/lice

nses/by-nc-sa/4.0/

Distributed under

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

Keywords:

colorimeter, color sorter, TCS3472 sensor.

Resumen

El presente trabajo muestra el diseño, elaboración y pruebas de un

primer sistema prototipo clasificador de objetos esféricos

pigmentados de colores. El objetivo versa sobre el análisis de

eficiencia y error que pueden tener los dispositivos digitales

fotoeléctricos. Basado en la detección de patrones de color

denominado modelo RGB (Red, Green & Blue) el sistema se compone

de un sensor detector de color digital TCS3472 controlado por una

plataforma electrónica de código abierto Arduino. El

microcontrolador ATMega328P procesa los datos obtenidos por el

sensor, que seguido ejecuta las acciones encargadas de mover el

servomotor que controla el mecanismo selector que, a su vez, consta

de un canal de acceso y salida hacia tres contenedores. Un panel de

mando externo con cuatro pulsadores: inicio, pausa, reseteo y

emergencia; este último configurado para la desconexión total del

sistema por posibles fallos de atascamiento. Finalmente, el sistema

entrega los resultados de la separación de esferas por color en el

panel de mando y físicamente en los repositorios de la estructura.

Palabras clave:

colorímetro, clasificador de colores, sensor

TCS3472.

Introduction

With the vertiginous advance of technology in recent years, various

electronic systems have been designed to help automate industrial

processes in different fields, such as production, food, manufacturing,

maintenance, etc. The classification of objects is part of a production

process, in this case, the color of the object is the main parameter to

take into account in this process. (Filote Razo, 2016).

A more advanced application of the system is the detection of objects

by shape and its application in supermarkets. Here, the cashier has to

distinguish several dozens of products without barcodes, such as

vegetables, sausages, breads, etc. Remembering all these goods is a

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

heavy task and even more so if it changes every day. Therefore,

automated recognition requires a system that can reduce the cashier's

workload. (Morimoto, 2018) The development of the present project

is accompanied by a previous study of articles concerning the use of

color sensors for object classification that have considered a TCS3472

sensor, and that provides perceived wavelength values of light

according to the RGB model.

A spherical object has been defined as the subject of study for the

classification. This is replicated in 4 bodies with a different color (blue,

red, green, yellow). Thus, after determining the model and color of

the object, we proceed to establish the mathematical model to be

used for color detection based on the saturation of the color and the

luminosity reflected on the body.

By being able to classify objects based on their coloration, future

applications for this classification system are thought of, being the

manufacture of fruit sorting and packaging.

In 2020, a group of researchers from the National University of

Colombia developed and evaluated a prototype for measuring the

color of fresh vegetables, using DHT11 sensors for temperature and

TCS230-3200 for color. Through tests carried out it was deduced that

the color measurement with the prototype is dynamic, because it

avoids the use of a lighting system, since the color sensor has

integrated LED devices. After the analysis of the results it was

concluded that the prototype developed by integrating components

such as sensors and data acquisition card and free software, allows to

evaluate the color in a different way; 2) the average error recorded by

the prototype was 15.57 %, lower value than that recorded by the

commercial colorimeter was 27.45 %. (Sarria-Dussán, Garzón-García,

& Melo-Sevilla, 2020).

Similarly Hermoza Llanos (2018) in his research conducted in Sacha

Inchi producing companies in Lima, determines as main objective the

design of a peeled fruit sorting system by color. All materials in

contact with the seeds were considered as non-contaminating, the

final output flow is 200 Kg/h distinguishing between dark brown and

almond colored seeds. A TCS3472 RGB color sensor, a vibrating table

and an LCD screen were used as materials.

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

A research conducted in 2018 in Italy to determine the impact of light

pollution on a vehicular driver used the RGB sensor type TCS34725

as a luminance meter as it allows the possibility to adjust the data

acquisition rate, between very high or low rates the latter range of

obtaining is necessary for situations where there is little traffic or

partially uninhabited areas in order to reduce the amount of data.

(C.D. Galatanu, 2018)

Tests carried out showed that the structure of the system is suitable

for modifications to an image processing system capable of

controlling the quality of the seeds to be processed. It is concluded

that the cost, dimensions and final processing capacity of the system,

make it competent compared to other options currently existing in

the market, in addition to increasing productivity in the seed selection

process, since it translates into shorter production times and therefore

greater commercial gains. (Hermoza Llanos, 2018). According to

these authors, the electronic color interpretation system that is

intended to be realized is feasible and has tools that are within reach

such as, for example, the RGB sensor TCS3472, microcontrollers of

the ATMEGA series, etc. The communication between these devices

and the visualization through a physical graphic interface can be done

since, in the analyzed researches, the architecture used and the

programming algorithms used can be observed.

Methodology

The classifier system has an RGB sensor, a servomotor, a 7-segment

display to visualize the number of hits of the classifier, programming

algorithms to meet the main objective which is to classify objects

according to color. The RGB sensor is the main element of the system,

since it can be used to determine the color of the object, which is not

possible without the help of the code. By means of these, the

acquired data can be processed and functions can be sent to the

actuators to perform the classification procedure. These actuators are

the servomotors that direct the object to a specific place according to

the detected color. The visual interface is a 7-segment display that

shows the count of each classified object. The following figure shows

the architecture of the exposed process.

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

Figure 1

Block diagram of operation

Test objects

Figure 2

RGB color sensor

The bodies selected for sorting are made of wood, because of their

easy sizing to the main storage funnel, and the shape that future

research will entail. These, at the time of the tests, have a diameter of

2cm and a weight of 70gr covered with blue, green, yellow and red

acrylic paint. The importance of these wooden objects lies in the

similar and irregular shape of their surface, like a vegetable. And their

color, which will inevitably change due to natural aging. These color

effects are visible over time and not necessarily at the time of

experimentation.

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

The color of a surface is described by a color composition: red, green

and blue. Thus, in a colored image each pixel is represented by a

certain value of these components. In the RGB color space, the pixel

p (i) is defined by an ordered triplet of red, green and blue

coordinates (r(i), g(i), b(i)), representing the red, green and blue light

intensities respectively. The intensity value varies from 0 to 255. (A.

Kanade, 2015)

Table 1.

Color analysis

Color



(nm)

Colors

analogs



(nm)

Color type

Blue

436-495

Blue

Turquoise

Blue

Violaceous

489.04

380-

500

Monochromati

Red

620-700

Red

Orange

600-

500

Monochromati

Green

495-570

Green

Blue

Yellow

Greenish

555-

574

Monochromati

Yellow

566-589

Orange

592-

620

Monochromati

RGB sensorization

This device has the ability to provide a digital RGB return from its

incident light in a sharp and accurate manner. Thanks to its integrated

infrared filter in the photodiodes along with its high sensitivity,

dynamic range place it as an ideal color sensor solution for use under

lighting conditions and through attenuating materials. (TAOS, 2012)

Figure 3.

RGB color sensor

The data taken by the RGB sensor are sent to the processing board,

where through the Arduino IDE the data will be processed to calibrate

the sensor, control the speed of the servomotor and count the sorted

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

objects. It is necessary to carefully set the speed at which to work and

install the necessary libraries to avoid problems in the execution of

the program.

These algorithms are created through a flow chart shown in the

methodology section. The importance of the algorithms is due to the

fact that, the classification process will be performed correctly and

through tests it will be possible to reduce the margin of error, if any.

The servomotors will be in charge of routing the route with the specific

angle to where the classified sphere should arrive according to its

color, the programming plays an important role in this stage since it

will be the physical basis of the classification.

Through the use of 7-segment displays, the number of objects that

have been sorted can be visualized. These are placed with labels to

identify which color corresponds to a certain quantity.

Mathematical model

• The parameters taken into account to describe the color are:

• L: brightness

• a: Red/green coordinates where +a indicates red and -a

indicates green.

• b: Yellow/blue coordinates in which +b indicates yellow and -

b indicates blue.

• c: represents the chroma key.

• h: corresponds to the hue defined as an angle (in degrees) on

the color wheel.

Once the color data is obtained, it is compared with other samples to

evaluate their differences (Ruíz Martinez, 2002). (Ruíz Martinez, 2002).

If these differences are called !", !#, y !$for the axes L, a and b

respectively, the total distance between two colors is given by !

whose formula is:

'!#

'!$

(1)

With the data acquired from the RGB sensors, it is divided for 255 in

order to have values in the range of 0 to 1 as shown below.

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

()%

(

*++

(2)

,)%

*++

(3)

-)%

*++

(4)

With this data, the pitch is calculated with the following formula:

44444444444444444444444444444444!%3

,)8-)

9:;5

$%&

%()

-)8()

'*<=>

$%&

%,)

()8,)

$%&

%-)

(5)

The brightness is calculated depending on the luminosity:

@%A

34=44444444444444!%3

*"8B

=!DE3

(6)

Where:

S: saturation.

!total color differential

L: Luminosity.

For the calculation of the percentage error, a standard statistical

formula is used in which the actual value and the approximate value

are used:

(

!"#$

)(

%&'(

(

!"#$

6B33G

(7)

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

System development

Principle of operation

Figure 4

Control Program

In this way, the principle of operation of the system is represented.

Thus, the sphere enters the upper part of the structure through a

circular duct of 2.3cm in diameter until it reaches the TSC34725

sensor. This is in charge of measuring the frequency of the visible

spectrum, determining its color. If there is a problem such as external

light leakage inside the cavity, the system will not work properly and

will show values not established in the programming, i.e. an alert.

Therefore, the user must re-enter the object to be classified and verify

that there is no filtration of ambient light inside the mechanism. If the

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

measurement made by the sensor is within the programmed

parameters, the sphere will move to the classification stage. Here, the

servomotor will move within the four positions previously defined and

depending on the value provided by the sensor. Each one will be

taken to the containers according to its color. For this action a Me-

0634 detector sensor is provided, which allows the system to be

automatic, even if it has a manual mode.

Since the spheres are sized according to the design, possible

blockages are reduced. Even so, it could be the case that the servo

runs through the propeller prematurely, then it will be necessary to

intervene with an anti-jamming function. The second and third points

focus even more on the action of the RGB sensor TSC34725. This is

responsible for measuring the frequency of the visible spectrum by

extracting the color of the sphere. As the same can have external

variations such as light and ambient temperature, the system is

designed to cover and place the sensor on the basis of displacement

of the sphere to reduce to the minimum possible sudden changes.

Therefore, if it were to fail while controlling these parameters, a

complete module replacement would be necessary.

Finally, this point four focuses on the mechanical part of the system,

which includes the servomotors, may have failures either

programming, physical and electrical, these should be corrected first

by programming, if this process does not work, the status and

operation of the servomotors should be verified.

Table 2.

Main components of the system

Hardware

Reason for selection

Sensor

TCS34725

It has an activatable IR signal

filter.

ATMega 328P

High performance, low

power consumption and

optimized for C compilers.

HiTec

Servomotor

Load handling up to 4Kg

7-segment

display

Efficiency and

implementation

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

Potentiometer

50k𝛺

For adjusting the current

flow.

Figure 5

Electronic structure of the system

The electronic boards define the communication structure of the

classification system by stages. In this way the logic stage, contains

the programming code (algorithms) that allow the proper operation

on control parameters that exerts on the physical components, i.e.

servos, mechanical stage that according to the sensor data locates the

objects (spheres) in the place established in the structure, in angle and

time.

Figure 6

Electronic control circuit and serial connection

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

Figure 7

Control and display panel circuit.

After the electrical communication tests on the breadboard, the

printed circuit boards are designed. These provide stability and

reliability to the control and data visualization system, since the

movement to which it is exposed by the servomotors can directly

affect the connection and its collateral effects.

Source code and programming

The systematization of the sensor, the processing of the data obtained

by it and the configuration of the selector mechanism is performed on

a programmable Arduino Nano board, with the use of the

Adafruit_TCS34725 library for the interpretation of the recorded

values.

First, the necessary libraries are included according to the chosen

sensor, this is done from the IDE library. Then, we proceed to declare

the variables for the counting of red, yellow and green spheres,

additionally we create a buffer where the counting of the spheres is

stored to send them by serial communication to the displays.

At this point, the variables in which the hit count of classified spheres

will be stored and a buffer for displaying the hits on the screen are

established.

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

Number of occurrences or events

Then the pins to which the modules will be connected are declared,

which consist of LEDs for the start and end of the process, as well as

the button that allows the system to start.

Table 4.

External Actions

At this stage the code that is executed once (void setup) is set. It is

usually set to mark general operating parameters such as module

startup and preconfigurations. There may be more stages of the same

type called functions.

Counting variables

int countR=0;

int countN=0;

int accountG=0;

int accountA=0;

byte buf[4];

Control Outputs

const int led_start=3;

const int ledSensor=4;

const int led_pare=5;

const int

buttonStart=6;

const int enced;

serovo=13;

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

Now, for the motion control, the pins for the servomotors are

assigned, in this case the pins (7 and 8) for PWM outputs will be used.

Afterwards, the Arduino serial communication is declared at the 9600

bits per second (baud rate) required for the microcontroller.

Table 5

. Position Control

In the "void setup" the programming that does not depend on

repetitions is established, that is to say that they will be executed only

once. A clear example is when the operation range is defined for each

frequency (color), from there it is calculated for all those used (red,

green, blue and yellow) which in turn will be stored in a buffer

parameter (buff) the number that will be recorded. If one of the

conditions is not met, the counter will return its previous value, i.e.

unchanged, this is because it can represent values with variation due

to external changes such as ambient brightness. In case of errors, the

hexadecimal code is printed in the serial monitor, besides serving as

an aid to verify that it is working correctly.

In this part, you set the limits of the data obtained from the sensor

that the algorithm accepts to describe the color. If one of the

conditions is not met, the counter will return the last correct value. If

the received data is outside the parameters, the received value is

printed on the Arduino IDE serial monitor.

Table 6.

Color detection

Servo control

topServo.attach(7);

bottomServo.attach(8);

Serial.begin(9600);

Color ranges

if((r < 2.15) & (r > 1.8) & (g < 0.75)&(g > 0.45)){

color = 1; // Red

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

buf[0]=1+countR++;

cont=0;

}

if((g < 1.4) & (g > 1) & (b < 0.6)&(b > 0.35)){

color = 2; // Yellow

buf[1]=1+countN++;

cont=0;

}

if((r < 1.15) && (g > 1.20) && (b < 0.8)){

color = 3; // Green

buf[2]=1+countG++;

cont=0;

}

if ((r < 1) && (g < 1.2) && (b > 1)){

color = 4; // Blue

buf[3]=1+countB++;

cont=0;

}

else{

cont++;

if(cont==5){

/*buf[0]=accountR++;

buf[1]=countN++;

buf[2]=G++ account;

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

In this stage, functions are created and executed in loops depending

on the data obtained from the sensors. It starts with the analog and

digital readings from the RGB sensor and the power button,

respectively. An anti-bounce function is created so that the state of

the button remains stored applying a delay of 20 milliseconds, this

allows that when the button is pressed, only one state remains stored

and not the other states that are presented at the moment of short-

circuiting positive and negative with the button.

After the previous function we proceed to perform an 'if' conditional

where it is indicated if the status of the button is 1 or if the infrared

LED detects a sphere, the main function called "main_system" will be

executed, otherwise the function "OFF" will be executed. When using

a push-button, an anti-bounce function must be performed in the

code since the residual currents generated when pressing it can send

wrong values to the board, this function provides a 20ms delay that

allows saving only one value at the moment of pressing the push-

button.

buf[3]=countB++;*/

output=0;

cont=0;

}

Serial.write(0xff);

for(int i=0;i<sizeof(buf);i++){

Serial.write(buf[i]);

delay(2);

}

return color;

}

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

Table 7.

Off Control

Shutdown Control

void loop() {

ReadBall=analogRead(DetecBall);

status=digitalRead(startButton);

//ANTI-REBOUND FUNCTION

if(status==HIGH && statusPrevious==LOW){

output=1-output;

delay(20);

}

statePrevious=status;

if(output==1||ReadBall<7){

systemPrincipal();

}else{

shutdown();

}

Table 8.

Color libraries

Acquisition of sensor

readings

Description

ReadBall=analogRead(De

tecBall);

Take readings

from the sensor

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

connected to

analog pin A1.

Servo movement

Description

digitalWrite(encedico_ser

ovo, HIGH);

Controls the

section that

transports the

sphere to one of

the containers.

Sensor data storage

Description

tcs.getRawData(&red,

&green, &blue, &clearcol);

Stores the RGB

spectrum data in

the assigned

variables for

further

processing.

Table 9.

HTML color analysis

HTML

#7828

RGB (r,g,b) B

(120, 40, 140)

CMYK(c,m,y,k)C

(70, 100, 0)

HSV (h, s, v)

(288H, 71%, 55%)

Violet: between 420 and 400 nm wavelength. Attenuated violet:

between 400 and 380 nm, it constitutes a band with a double

connotation, because on the one hand it is considered as part of the

visible violet light and on the other hand it is part of the ultraviolet

radiation (UV). Visible ultraviolet light: between 380 and 310nm.

Although by definition UV radiation is not visible, a part of this

radiation is called UVA or near ultraviolet.

The instrument (colorimeter) is composed of two devices: the control

and sorting device. The first, composed of controls that allow the

system to activate, pause, restart, accelerate or stop its operation. The

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

second has a section for transporting, sensing, transporting and

locating.

Figure. 5

3D CAD control

Table 10.

Elements and symbology

N°

Element and function in the system

7-segment display: counters

Green Led: Indicates that it is working

Red Led: Indicates that it is stopped

Button: Starts the system

Button: Pauses the system

Button: Resets the system

Potentiometer: Accelerates or slows down

speed

Button: Emergency

Red Led: ON = emergency; OFF = normal

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

Figure. 6

Control and visualization implementation

The transport section, designed specifically for the test objects and

their dimensions, carries the spheres to the duct in which the digital

sensor is located. Under these placement conditions, the component

extracts the colors of each of the objects arranged in the chute as they

arrive. When the color is known, the servomotor locates them at the

correct angle in their respective repository, and in parallel sends a

signal to the optoelectronic devices for counting and numerical

display.

Results

The group of non-degraded color tests chosen, allows us to establish

the degree of efficiency in the specific detection of the solid colors of

a wide gamma that it is understood would be detected. In the end,

this allows us to know if the system composed mainly by the sensor

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

based on the criteria of sensitivity and luminosity configuration, as

well as its location, are acceptable for this type of process. For this,

three primary colors and a secondary one are chosen, which will allow

us to know the number of hits in the detection of these essential

colors.

On the number of tests (1089 per sphere) separated in ranges of 99

consecutive repetitions for each one. And, taking into account the

effectiveness of the sensor that is given by the saturation and

luminance of the object, as well as the configuration divided into three

stages, with percentages of tone in the saturation and luminance of

50%, 75%, 100%, obtaining the following:

The unforeseen initial conditions between 0 and 50% allow us to

perform a test with error reduction and adequate illumination and

saturation conditions.

Table 6.

Hit and Miss T1 (50%)

Samples

Colors

Hits

Saturat

ion

Brightness

Error

1089

Blue

101

50%

0,060%

Green

103

0,049%

Yellow

102

0,060%

Red

100

0,073%

Total

408

Total

0,242

In all the events, the saturation and luminosity are maintained. And

from this, a total of 273 failures are obtained divided into 73 for the

blue color with an error percentage of 0.060%, 54 for the green color

with an error of 0.064%, 66 for the yellow color with a percentage

error of 0.066% and 80 failures for the red color with an error

percentage of 0.073%.

Table 7.

Hit and Miss T2 (75%)

Samples

Colors

Hits

Saturation

Brightness

Error

1089

Blue

1031

75%

0,053%

Green

1024

0,059%

Yellow

1036

0,048%

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

Red

1031

0,053%

Total

4119

Total

0,213%

In all the insertions the system maintains the saturation and luminosity

stable with the parameter set to 75% and this results in a total of 4119

hits and 237 misses divided into 58 for the blue color, with a

percentage error of 0.053%, 65 for the green color with an error of

0.059%, 53 for the yellow color with a percentage error of 0.048% and

61 misses for the red color with an error percentage of 0.053%.

Table 8.

Hit and Miss T3 (100%)

Samples

Colors

Hits

Saturation

Brightness

Error

1089

Blue

1071

100%

0,016%

Green

1054

0,032%

Yellow

1066

0,021%

Red

1067

0,020%

Total

4258

Total

0,089%

Stabilizing the brightness and saturation to 100%, the system shows

no variation in this factor, and from this we obtain a total of 98 failures

divided into 18 for the blue color, with an error percentage of 0.016%,

35 for the green color with an error of 0.032%, 23 for the yellow color

with a percentage error of 0.021% and 22 failures for the red color

with an error percentage of 0.020%.

Figure. 7

Efficiency increase

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

Figure. 8

Error Decrement

Figure. 9

3D CAD classifier

Color Yellow

Series 1, limited to 50% sensitivity and luminosity, detects 1023 blue

spheres out of a total of 1089 tests of this color on the sphere. That

is, a difference of 66 misses corresponding to only 6.1% error in the

expected sensing, which, on the contrary, shows 93.9% effectiveness

in its determination (efficiency).

Series 2, defined at 75% sensitivity and brightness, detected 1036

blue spheres out of a total of 1089 tests of this color on the sphere.

On the other hand, there were 53 misses, which corresponds to 4.9%

of failures in detection, with 95.1% of effectiveness in detection.

Series 3, established at 100% sensitivity and luminosity, detected

1066 blue spheres out of a total of 1089 tests of this color on the

sphere. That, in its defect presents 23 errors that correspond to only

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

2.1% of error in its discovery; with a 97.9% of effectiveness in its

detection.

Color Blue

Series 1, limited to 50% sensitivity and brightness, detects 1016 blue

spheres out of a total of 1089 tests of this color on the sphere. That

is, a difference of 73 misses corresponding to only 6.7% error in the

expected sensing, which, on the contrary, shows 93.3% effectiveness

in its determination (efficiency).

Series 2, defined at 75% sensitivity and brightness, detected 1031

blue spheres out of a total of 1089 tests of this color on the sphere.

On the other hand, there were 58 misses, which corresponds to 5.3%

of failures in detection, with 94.7% of effectiveness in detection.

Series 3, established at 100% sensitivity and luminosity, detected

1071 blue spheres out of a total of 1089 tests of this color on the

sphere. That, in its defect presents 18 errors that correspond to only

1.65% of error in its discovery; with a 95.1% of effectiveness in its

detection.

At the end of the test series 1 with an intensified sensitivity and

brightness from 50% to 100% shows an increase of 5.1% in the

detection of the number of blue spheres, i.e. 55 additional spheres

with an equal and inversely proportional reduction of the error.

Color Red

Series 1, limited to 50% sensitivity and brightness, detects 1009 blue

spheres out of a total of 1089 tests of this color on the sphere. That

is, a difference of 80 misses corresponding to only 7.3% error in the

expected sensing, which, on the contrary, shows 93.3% effectiveness

in its determination (efficiency).

Series 2, defined at 75% sensitivity and brightness, detected 1031

blue spheres out of a total of 1089 tests of this color on the sphere.

On the other hand, there were 58 misses, which corresponds to 5.3%

of failures in detection, with 94.7% of effectiveness in detection.

Series 3, established at 100% sensitivity and luminosity, detected

1067 blue spheres out of a total of 1089 tests of this color on the

sphere. That, in its defect presents 22 errors that correspond to only

2% of error in its discovery; with a 98% of effectiveness in its detection.

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

Color Green

Series 1, limited to 50% sensitivity and brightness, detects 1035 green

spheres out of a total of 1089 tests of this color on the sphere. That

is, a difference of 54 misses corresponding to only 4.9% error in the

expected sensing, which, on the contrary, shows 95.1% effectiveness

in its determination (efficiency).

Series 2, defined at 75% sensitivity and brightness, detected 1024

green spheres out of a total of 1089 tests of this color on the sphere.

On the other hand, there were 65 misses, which corresponds to 5.9%

of failures in its detection, with 94.1% of effectiveness in its detection.

Series 3, established at 100% sensitivity and luminosity, detected

1054 green spheres out of a total of 1089 tests of this color on the

sphere. That, in its defect presents 35 misses that correspond to only

3.2% of error in its discovery; with a 96.8% of effectiveness in its

detection.

Testing of the classification system

Figure 10

Complete sorting device

The classifier system contemplates a sequential scheme where there

is a point of entry of the spheres, then it is directed with an actuator

to the TCS3472 sensor, this is responsible for determining based on

algorithms a correct decision (code), determining the position of the

servo in the following degrees (red 58°, yellow 92°, green 134° and

blue 175°), at the end the sphere descends to be placed on the base

of which were presented f, as follows:

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

The spheres, by complying with regular proportions and in

accordance with the design of the structure, did not present

inconveniences as expected in the circuit of travel of this one, giving

an optimal work flow for the operation tests in the first 152 samples.

To guarantee a better result, a stopper was added in the ejection zone

so that it does not hit the surface of the base and consequently there

is no greater wear on the paint.

In test sequence 153, the sorter jammed, requiring the emergency

button to be pressed to bring the prototype to a complete stop. This

behavior was repeated in a subsequent range of 130 to 160 runs on

the mechanism.

When running the 752 test, the classifier suffered the jamming of a

green sphere, being necessary to press the emergency button to stop

the mechanism, however, it got stuck in its cavity, so it was necessary

to restart the system and disconnect the power of the prototype to

unjam the button.

Conclusions

The use of the sensor itself, the white LED, helped to avoid the

principle of color saturation, which supports the principle of

luminosity, which contributed to the fact that by bouncing the light

from it on the sphere, the real wavelength value of the color used on

it can be obtained. At the end of the test series 1 with an intensified

sensitivity and luminosity from 50% to 100% presents an increase of

5.1% in the detection of the number of blue spheres, i.e. 55 additional

ones with an equal and inversely proportional reduction of the error.

At the end of the test series 2 with an intensified sensitivity and

luminosity from 50% to 100%, there was an increase of 1.7% in the

detection of the number of green spheres, i.e. 19 additional spheres,

with an equal and inversely proportional reduction of the error. At the

end of the test series 3 with an intensified sensitivity and luminosity

from 50% to 100%, there was an increase of 3.9% in the detection of

the number of blue spheres, i.e. 19 additional spheres with an equal

and inversely proportional reduction of the error.

At the end of the test series 4 with a sensitivity and luminosity

intensified from 50% to 100% presents an increase of 5.3% in the

detection of the number of blue spheres, i.e. 58 additional spheres

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

with an equal and inversely proportional reduction of the error. The

tests performed were able to determine that the higher the

percentage of hue, saturation and luminosity, the greater the number

of spheres classified. With a percentage of 100 % in each of the above

mentioned parameters, the error could be reduced from 6.26 % to

2.27 % with respect to the average error among all colors.

The saturation of the spherical object was an important factor in the

classification. The blue color is greatly affected by this parameter,

since at first it had an incidence of error equal to 6.70 % with

saturation equal to 50 %, which could be reduced to 1.74 % with a

saturation of 100 %. The mechanical prototype, although it works

properly in 96%, since it has only two displays for each color to count

the number of hits, it does not allow to display more than 99 tests with

positive results continuously for each one before having to restart the

system.

When the emergency button got stuck and the system was

continuously restarted, it was determined that the cavity that houses

it is narrower on the inner edge due to a fault in the construction of

the container, making it necessary to replace it with a switch that

provides more safety in its operation because the mechanism does

not interact with the structure that contains it.

Reference

Ruíz Martinez, J. (2002). Fisica Y Quimica. Profesores de Enseñanza

Secundaria. Murica: MAD-Eduforma.

Filote Razo, J. D. (2016). RGB sensor to detect color changes in fruit

skin. Acta Universitaria, 7.

Ohno, Y. (2000). CIE fundamentals for color measurements. Society

for Imaging Science and Technology, 2000, 540--545.

Serrano Tierz, A., & Biel Ibáñez, M. P. (2013). Aesthetics of the

industrial product and its graphic representation. Zaragoza:

Presses of the University of Zaragoza.

Sarria-Dussán, S., Garzón-García, A. M., & Melo-Sevilla, R. E. (2020).

Development and evaluation of a prototype for color

RGB free colorimeter for sensing and sorting of pigmented spheres

Revista Científica Interdisciplinaria Investigación y Saberes , / 2023/ , Vol. 13, No. 3

measurement in fresh vegetables. Information Technology, 253-

260.

Hermoza Llanos, E. A. (2018). Design of a peeled sacha inchi seed

sorting machine by color with a capacity of 200 kg/h.

Dissertation. Pontificia Universidad Católica del Perú, Lima.

A. Kanade, A. S. (2015). Development of Machine Vision based system

for classification of Guava fruits on the basis of CIE1931

chromaticity coordinates. Symposium on Physics and

Technology of Sensors. Pune, India.

C.D. Galatanu, L. C. (2018). Measuring the driver exposure to the light

pollution. IEEE International Conference on Environment and

Electrical Engineering and 2018 IEEE Industrial and Commercial

Power Systems Europe (EEEIC / I&CPS Europe). Palermo, Italy.

TAOS (2012).

Morimoto (2018). An Object Recognition Method using RGB-D

Sensor. Second IAPR Asian Conference on Pattern Recognition.

Himeji, Japan.