You are on page 1of 5

CAPSTONE PROJECT PLAN

“Infrastructure Damage Detection and Reporting - IDDR”

Team ID: B21-CAP0475


Active Member ID and Name:
1. M3142811 - Gugi Asgaruning
2. M0080916 - Leon Davin
3. C0080914 - Dheja Daruna Panjalu
4. A3192857 - Mhd Iqbal Maulana
5. A3192854 - Alfi Dharmawan
Inactive Member ID and Name:
1. C3172845 - Refi Setiawan

Selected Theme: Infrastructure Development

Title of the Project: Infrastructure Damage Detection and Reporting (IDDR)

Executive Summary:
We notice that currently PUPR has developed an apps to report infrastructure problem like broken
road, broken buildings and etc. But the reporting process itself still need to write the problem down
manually. Thus we propose to use object detection using Machine Learning to optimize the
process.

Project Scope & Deliverables:

Android: Build an applications that can take pictures of damaged road, communicate with cloud
servers, generate the report including GPS coordinates, location and date.

➢ Week 1 :Build and develop android applications based on the problem


➢ Week 2 : Create feature of each solution and UI of the app (XML layouting)
➢ Week 3 : Connecting data through database & cloud
➢ Week 4 : Implementing ML model on the application, test & debugging

ML : Perform the classification of the images that have been taken.

➢ Week 1 : Doing research on the problem that want to solve


➢ Week 2 : Collect the dataset and explore the features that will be used in the project
➢ Week 3 : Performs some data processing before it is used by the model
➢ Week 4 : Modeling using a classification algorithm, evaluate and deploy models that
have been made into mobile devices

Cloud Computing: Run the computation process from Machine Learning and integrate reports
obtained from Android with Firebase

➢ Week 1-4 : Create cloud project and set up the member roles, provide database with
firebase, and create Cloud Run for the python script.
Project Schedule:
Week 1 Week 2 Week 3 Week 4

Android

Build and develop


android applications
based on the
problem

Create feature of
each solution and UI
of the app (XML
layouting)

Connecting data
through database &
cloud

Implementing ML
model on the
application, test &
debugging

Machine Learning

Doing research on
the problem that
want to solve

Collect the dataset


and explore the
features that will be
used in the project

Performs some data


processing before it
is used by the model

Modeling using a
classification
algorithm, evaluate
and deploy models
that have been
made into mobile
devices
Cloud Computing

Create cloud project


and set up the
member roles.

Provide database
with firebase.

Create Cloud Run


for the python script.

Project Resources:

1. Papers/Journals
a. Maeda, H., Sekimoto, Y., Seto, T., Kashiyama, T., & Omata, H. (2018). Road
damage detection using deep neural networks with images captured through a
smartphone. arXiv preprint arXiv:1801.09454.
b. Dais, D., Bal, I. E., Smyrou, E., & Sarhosis, V. (2021). Automatic crack
classification and segmentation on masonry surfaces using convolutional neural
networks and transfer learning. Automation in Construction, 125, 103606.
c. Kim, B., & Cho, S. (2018). Automated vision-based detection of cracks on concrete
surfaces using a deep learning technique. Sensors, 18(10), 3452.

2. Dataset
a. Crack Detection: Image Classification
https://www.kaggle.com/c/crack-identification-ce784a-2020-iitk/data
b. Concrete Crack Images for Classification
https://data.mendeley.com/datasets/5y9wdsg2zt/1
c. RDD
https://mycityreport.s3-ap-northeast-
1.amazonaws.com/02_RoadDamageDataset/public_data/IEEE_bigdata_RDD202
0/test1.tar.gz

3. Budget
a. Google Cloud Platform : $100
b. Google Map API : $100
Risk and Issue Management Plan:

Android Development Risks :


1. Application not functioned properly as intended.
2. Connecting the data and cloud take too much time and maybe fail at the process.
3. Some services & API not functioned well that lead to a freeze app/force close.
4. The layout doesn't fit in some devices, thus may lead to unreadable content / broken image

Machine Learning Development Risks :


1. Failed to make predictions of the damaged road image
2. Failed in implementing the model with android application
3. Failed to integrate the model with the cloud server
4. Failed in reporting pictures taken using an android camera
5. Failed to generate reports from images captured using an android camera

Cloud Computing Risks:


1. Cloud Run can't run the python script because of the wrong configuration.
2. Set the wrong config for the Firebase.

You might also like