Volume 8, Issue 5, May – 2023 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165

Automatic Garbage Segregation and

Management System
Ahalya R1 Rishab Mahato3
UG Scholar – ETE UG Scholar – ETE
Bangalore Institute of Technology Bangalore Institute of Technology
Bangalore, Karnataka Bangalore, Karnataka

Anigani Nikhila2 B Sudha4

UG Scholar – ETE Assistant Professor - ETE
Bangalore Institute of Technology Bangalore Institute of Technology
Bangalore, Karnataka Bangalore, Karnataka

Abstract:- The increasing population, rapid growth of According to the statistics, the world generates about
industries and urbanization have led to the daily 2.24 billion tons of solid waste per day, which amounts to a
generation of a humongous amount of waste, which may footprint of 0.79 kilograms per person per day. At this rate,
cause fatal diseases in living beings and also cause by 2050, waste generation is expected to increase by 73%,
pollution. Hence the collection and segregation of which levels to 3.35 billion tons of waste per day. In India,
garbage is a major issue faced by the cities. The the waste generation amounts to 0.1 million tons per day.
segregation of waste at junk yards is mundane and Solid waste is disposed of by several methods, the most
laborious which consumes a lot of time and labor. To prominent among them is the incineration of the solid waste.
prevent this, we propose a system to segregate garbage Incineration of waste leads to the generation of a large
at the domestic level based on the nature of the waste. amount of greenhouse gas like carbon- di-oxide. The other
The proposed model is a microprocessor-based IoT method followed widely is dumping solid waste into
device where the bins are embedded with sensors that landfills, which pollutes the land and the underground water
help in identifying the type of waste. The garbage is by percolating through the layers.[1]
segregated using a robotic arm that uses object detection
through Image processing to pick and place the garbage Proper management of waste is necessary to build
into their respective dust bins, then using the ultrasonic sustainable cities. But it remains a challenge for many
sensors the garbage level is monitored and updates the underdeveloped and developing countries as effective waste
data in the cloud using Blynk IoT application. The main management requires an integrated system that is effective,
objective of our model is to provide an efficient and cost- efficient, and sustainable. According to the 'Zero Landfill'
effective solution for the current problem in waste policy, 100% of the recyclable waste must be recycled.
management i.e., proper segregation of the garbage. There is a need to develop a time-targeted action plan for
waste management to prevent the impending environmental
Keywords: Pollution; Segregation; Microprocessor; Iot; crisis.
Sensors; Robotic Arm.
II. LITERATURE SURVEY
I. INTRODUCTION
As shown in Table 1, from paper 1 and 4 image
The metropolitan cities of India are littered with processing technique is referred, from paper 2 MQTT is
garbage, leading to many problems like serving as a breeding referred, from paper 3 CNN and Deep Learning is referred
ground for many infectious diseases, pollution, etc. Often and from paper 2 and 5 IoT techniques are referred. From
the waste is collected only from significant locations of the various papers it is observed that, the limitations of the papers
cities or towns, leaving the other places unattended and are waste must be thrown separately and one at a time and
choking with the waste collected over time. However, waste also limited data sets of components images lead to
management has been ineffective, dumping them into inaccuracy. Future scope is to develop an automated waste
unattended landfills. segregation and it can be done by robotic arm and to train
the image processing model with large data sets.

IJISRT23MAY1474 www.ijisrt.com 2237

Volume 8, Issue 5, May – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
Table 1 Comparative Study of IEEE Papers

 Problems Identified By following the above methods, it can be made sure

From the Literature Survey carried out, several to promote a healthy and clean environment, and it reduces
problems were identified in the existing process of waste the risk of spreading infectious diseases due to human
segregation and management. They are: interaction at the dump yards and landfills.

 Manual segregation is tedious and time-consuming.  Software and Hardware Requirements

 Human interaction with garbage leads to many Hardware and Software requirements are shown in
infectious diseases,deteriorating the health of humans. Component Diagram of Fig.1.
 The garbage from the dust bins in public places is
not collected at the appropriate time leading to an  The Hardware Requirements for the Project are:
overflow of bins.  Raspberry Pi 3
 In many places, separate bins are not maintained  Camera Module
for wet and dry waste making the segregation  Robotic Arm
process difficult.  Ultrasonic Sensor SR04
 Gas Sensor MQ2
 Proposed Solution
Following the problems identified in the existing
 Arduino UNO
system, the solutions proposed are:  Breadboard
 The Software Requirements for the Project are:
 To build a model which can segregate the waste  Raspberry Pi OS
automatically into several categories.  Open CV
 To reduce the human interaction in garbage segregation  TensorFlow
to the minimum level possible.  Python
 To ensure the timely collection of garbage from the bins  Blynk IoT Application
by notifying the authorities.
 Arduino UNO

IJISRT23MAY1474 www.ijisrt.com 2238

Volume 8, Issue 5, May – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165

Fig 1 Component Diagram

Fig 2 Sample of Waste Images Dataset
III. METHODOLOGY
 Labelling the images: The images obtained must be
 Proposed Methodology labeled to train the object detection model. We use the
The main objective of the project is the segregation and LabelImg software to label the images. The images are
disposal of waste. The garbage dropped into the bin is annotated by drawing a bounding box around the
segregated (bio-degradable, plastic, metal etc.) at the panel interested area of the image as shown in Fig. 3 and the
with the data acquired from the results of image processing same dimension of the image is converted into a.xml
and the garbage is dropped into the respective bins. file which is supported by algorithms like YOLO.

 The garbage is thrown into the compartment allotted for

the mixed waste.
 The robotic arm incorporated identifies the dry waste
from the wet waste using Image Processing and picks
and place the waste into the respective bins [5].
 The garbage that is identified as dry waste, is further
identified as paper, plastic, metal, glass and E-waste
accordingly robotic arm picks up and places the waste
into the respective bins [6].
 Gas sensors are used in the bio-degradable segment to
sense the level of methane gas. Ultrasonic sensors are
used to determine the level of the trash collected and is
updated in the database. The trash can be then collected
to be recycled.
 The type of garbage collected and the level of the dust
bins are updated onto Blynk IoT application through Wi
Fi [7].
Fig 3 LabelImg Interface
 Image Processing
The image processing part of this project is divided  Training the TensorFlow model: To train the
into four steps. TensorFlow model, the transfer learning process is
used, where a pre-trained model is fine-tuned for the
 Collection of dataset: The dataset required for training intended purpose. The dataset is loaded and the model is
the image processing model is collected from various trained, and tested for the images classified as test
internet sources. The majority of these images are images. The accuracy of the model is measured using
collected from an existing dataset known as trashnet mean average precision (mAP) as shown in Fig. 4.
available over the internet. The dataset consists of 2527
images as shown in Fig. 2, that are used to train the
model.

IJISRT23MAY1474 www.ijisrt.com 2239

Volume 8, Issue 5, May – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165

Fig 4 Calculation of mAP of the Model

 Deploying the model: The custom-trained TFlite model is deployed onto the raspberry pi and the program is run for object
detection, the result is shown in the Fig. 5.

Fig 5 Object Detection Result

IJISRT23MAY1474 www.ijisrt.com 2240

Volume 8, Issue 5, May – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
 Flow Chart
The steps involved in Image Processing can be depicted in a flow chart as shown in Fig. 6.

Fig 6 Flow Chart Depicting the Process in Image Processing

IJISRT23MAY1474 www.ijisrt.com 2241

Volume 8, Issue 5, May – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
IV. RESULTS

The project model is shown in Fig. 7(a). The camera

gives the input video to Raspberry Pi, the object detection
model detects the type of waste and displays the result on
the Raspberry Pi window The robotic arm then picks up the
waste and places it into the corresponding bin as shown in
Fig. 7(b), (c), (d) and (e).

Fig 7 (d) Robotic Arm Dropping Paper

Fig 7 (a) Project Model

Fig 7 (e) Robotic Arm Dropped Paper

In this model waste management is done by using

Fig 7 (b) Camera Detected the Object Blynk IoT application. The level of the waste in the bins is
stored in cloud and updated to Blynk IoT. In the Fig. 8, it is
shown that level of the waste inside the bin on hourly, daily
and weekly basis where wet and glass bins are almost full.

Fig 7 (c) Robotic Arm Picking Up Paper Fig 8 Wet and Glass Waste Bins is almost Full

IJISRT23MAY1474 www.ijisrt.com 2242

Volume 8, Issue 5, May – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
V. CONCLUSION

The proposed system is an innovative solution to the

challenges faced in waste management. The use of sensors,
image processing, and IoT technology makes the process of
garbage segregation at the domestic level more efficient and
cost-effective. By separating the waste based on its nature,
the system facilitates recycling, reduces the amount of waste
sent to landfills, and helps in preventing environmental
pollution. The system also provides real-time monitoring of
the garbage levels in the bins, enabling timely collection and
disposal. Overall, this model can contribute significantly to
creating a cleaner and healthier environment, making it a
valuable addition to waste management systems in cities and
other urban areas.

ACKNOWLEDGEMENT

We express our heartfelt gratitude to The Chief

Executive Engineer, B.B.M.P, Bangalore for the support he
extended for our project. We also express our heartfelt
gratitude to our Head of Department Dr. M Rajeswari and
Project Guide Prof. Sudha B for their immense support and
guidance during the course of this project.

REFERENCES

[1]. C. KAYMAK and A. UCAR, "Implementation of

Object Detection and Recognition Algorithms on a
Robotic Arm Platform Using Raspberry Pi," 2018
International Conference on Artificial Intelligence
and Data Processing (IDAP), 2018, pp. 1-8, doi:
10.1109/IDAP.2018.8620916.
[2]. T. J. Sheng et al., "An Internet of Things Based
Smart Waste Management System Using LoRa and
Tensorflow Deep Learning Model," in IEEE Access,
vol. 8, pp. 148793-148811, 2020, doi:
10.1109/ACCESS.2020.3016255.
[3]. S. N., P. M. Fathimal, R. R. and K. Prakash, "Smart
Garbage Segregation & Management System Using
Internet of Things(IoT) & Machine Learning(ML),"
2019 1st International Conference on Innovations in
Information and Communication Technology
(ICIICT), 2019, pp. 1-6, doi:
10.1109/ICIICT1.2019.8741443.
[4]. Jaikumar, K., Brindha, T., Deepalakshmi, T. K., &
Gomathi, S. (2020). IOT Assisted MQTT for
Segregation and Monitoring of Waste for Smart
Cities. 2020 6th International Conference on
Advanced Computing and Communication Systems
(ICACCS). doi:10.1109/icaccs48705.2020.9074399.
[5]. D. K. Tripathi, S. Dubey and S. K. Agrawal, "Survey
on IOT Based Smart Waste Bin," 2020 IEEE 9th
International Conference on Communication Systems
and Network Technologies (CSNT), 2020, pp. 140-
144, doi: 10.1109/CSNT48778.2020.9115793.

IJISRT23MAY1474 www.ijisrt.com 2243