ADVANCING LANDSLIDE VICTIM

IDENTIFICATION SYSTEM
THEME : Climate Change/ Disaster management
GOVERNMENT COLLEGE OF ENGINEERING
SENGIPATTI THANJAVUR

SUPERVISED BY : TEAM MEMBERS:

Dr.P.Krishnamoorthy AKILAN A : 822721105007
Head of the Department ARUNKUMAR A : 822721105008
Department of EEE GUHANS :822721105018
OUTLINE
*Problem Statement & Abstract
*Introduction & Objective &
* Literature review
*Existing & Proposed system
*Proposed methodology
*Hardware & Software requirements
*Advantages & Challenges
*Future Plan & Timeline
*Estimated cost & Result
*Conclusion & References
 PROBLEM STATEMENT
• How might we develop a life detector that accurately identifies human presence under soil during
landslides, minimizing false positives and improving rescue efficiency by distinguishing humans from
other living creatures?

SOLUTION
• Our project is designed to locate human survivors buried under soil during landslides with
high accuracy. It uses advanced detection techniques to distinguish humans from animals,
reducing false alarms and improving rescue efficiency. By analyzing vital signs and unique human
biological markers, our system ensures fast and reliable identification, helping rescue teams
reach victims quickly.
 ABSTRACT
An AI-enhanced landslide detection and early warning system is developed by integrating environmental
sensing, wireless communication, and IoT-based monitoring. The system includes a transmitter unit and a
receiver unit. The transmitter unit is created using an Arduino Uno, which is attached to a sound sensor,
vibration sensor and accelerometer, which is to capture real -time data related to ground movement, tremors
and sloping variations. This data is displayed on a local LCD screen and transmitted wirelessly to the
receiver unit through Zigabi. The recipient controlled by an ESP8266 microcontroller, receives and displays
sensor data on his own LCD and uploads it to the Blynk IoT platform enables real -time remote access
through smartphones or network interfaces. In addition, sensor data is sent to a PC-based AI model
analyzing the pattern to identify early indicators of landslides. In the event of a possible danger, the system
generates a warning locally and through lead. This layered architecture enables continuous monitoring,
intelligent analysis and timely warnings for landslides.
 INTRODUCTION

Landslides being common natural disasters in hilly or

mountain areas cause massive loss of life, property,
damages, and disruption of infrastructure.Old
practices in landslide monitoring were based on
manual observation or simple sensor threshold
systems, which were neither timely nor accurate.
Intelligent and automated monitoring systems can be
developed wedging in technologies like the wireless
sensor networks, IoT platforms, and AI to facilitate
early detection and response.
* An integration of several technologies is studied for real-time monitoring and AI analysis of pertinent
environmental parameters relevant to landsliding. At the monitoring site, a network was established comprising
Infrared sound sensors, vibration, and accelerometer modules, all managed by an Arduino Uno microcontroller.
The sensor data are transmitted wirelessly by means of Zigbee to a remote controller powered by an ESP8266
microcontroller. LCD displays are mounted on both the transmitter and receiver units for on-site data viewing. To
view remote monitoring data, the receiver unit is associated with the Blynk IoT platform, which lets users
monitor conditions through their mobile or web interfaces. In parallel, the data from the sensors are passed on to
a PC wherein an AI model analyzes the data to identify pattern recognition that may be indicative of a possible
landslide. Once the risk situation is detected, it is communicated through both local and remote alerts.

* By sensor technologies, wireless communications, cloud-based monitoring, and AI-driven prediction, the
system combines to provide a productive- and cost-efficient- and scalable solution for landslide detection and
early warning in susceptible regions.
 OBJECTIVE

To develop a cost-effective and efficient life detection system for landslide-prone areas, utilizing

advanced sensor technology to enhance disaster response and minimize casualties. This project aims

to detect signs of life beneath debris, enabling quicker rescue operations and improving survival

chances.
 LITERATURE REVIEW
Sl.No Title Year Author(s) Content

1 IoT-Driven Microseismic 2024 P. K. The study presents an Internet of Things (IoT)-

Sensing System and Monitoring INDUKALA , U. based microseismic monitoring system for early
Platform for Landslide G. GOSH, AND landslide detection in susceptible terrain is
Detection. MANEESHA presented in the study. Cloud analytics is used to
VINODINI process real-time data and find landslide
RAMESH precursors. To improve preparedness for
disasters, the platform provides alerts and remote
access. When compared to traditional systems,
the results show increased sensitivity and quicker
reaction times.

2 Landslide Detection Based on 2023 Linrun The study presents for increased accuracy in
Improved YOLOv4 Li; Zhangjian challenging terrains, the study introduces a novel
Qin landslide detection technique that combines U-
Net with an improved YOLOv4 model. When
compared to conventional methods, this approach
improves mean Intersection over Union (IoU) for
small-scale landslides by 20.6% and for complex
backgrounds by 2.08%.
Sl.No Title Year Author(s) Content

3 Sensor-Driven-Landslide 2023 Suvankar The study presents a sensor-driven

Detection System for the South Singha;JaydeepRath; landslide detection system tailored for the
Eastern Region of India. ShankhosuvroGhosh; South Eastern region of India, integrating
Sarita Nanda; Piyush geotechnical sensors with wireless
Ranjan communication to monitor slope stability.
Targeting to reduce landslide hazards in
sensitive areas, this system provides early
warning capabilities and real-time data
collecting.

4 Machine Learning based 2023 The study presents a machine learning-

Landslide Detection System R.Anusuya;N.Anusha; based landslide detection system that
V. Sujatha; R. integrates geotechnical sensors with
Radhika; S. Iniyan- wireless communication to monitor slope
stability in real-time. The system aims to
provide early warnings to mitigate
landslide risks in vulnerable areas,
enhancing disaster preparedness and
response.
 EXISTING SYSTEM
The system you're describing is designed to improve landslide monitoring and early warning. It uses
geophone sensors (SM-s nodes) to detect small vibrations in the ground (microseismic signals) in areas prone
to landslides. These sensors continuously record data about seismic activities and send it in real-time. The
data helps detect early signs of landslides caused by factors like heavy rainfall or earthquakes.What makes
this system effective is its ability to tell the difference between microseismic signals (which could indicate a
landslide) and everyday noises like people walking or vehicles moving. The system was tested through
laboratory experiments and fieldwork in Chandmari, Sikkim, a region in India known for its landslides.To
measure the system's effectiveness, researchers analyzed how the ground moved over time, comparing it with
actual landslides caused by earthquakes and rainfall. This helped confirm that monitoring these small ground
movements can be a good way to predict when a landslide might happen. The system outperforms previous
methods and could be very useful for warning people in advance, helping to prevent damage and save lives.
 PROPOSED SYSTEM
The proposed system is an AI-enhanced landslide detection and early warning system that combines
environmental sensing, wireless communication, and IoT-based monitoring. The transmitter unit is built using an
Arduino Uno microcontroller, connected to an sound sensor, vibration sensor, and accelerometer to capture real-
time environmental parameters such as ground displacement, tremors, and slope variation. The sensor readings
are shown on a local LCD display and wirelessly transmitted to the receiver unit via Zigbee. The receiver unit is
powered by an ESP8266 microcontroller, which serves as both a data processor and an IoT node. It receives the
sensor data, displays it on its own LCD screen, and uploads it to the Blynk IoT platform, enabling remote
monitoring via smartphone or web interface. An AI model is deployed in the control system to analyze incoming
data patterns and detect early indicators of landslides. If potential danger is identified, the system can trigger
local alarms and send alerts through Blynk. This layered approach ensures robust, real-time monitoring and
proactive landslide risk management in vulnerable regions.
 PROPOSED METHODOLOGY

FIELD DETECTION
UNIT POWER SUPPLY

ACCELEROMETER
SENSOR AI (HUMAN)

VIBRATION SENSOR COMMUNICATION

MICROCONTROLLER
(ZIGBEE)

SOUND SENSOR

LCD DISPLAY

GPS MODULE
 POWER
MONITORING
UNIT
SUPPLY

IOT
LZIGBEE (ESP8266) LBLYNK

LCD DISPLAY
HARDWARE & SOFTWARE REQUIREMENTS

HARDWARE SPECIFICATION SOFTWARE SPECIFICATION

OPERATING SYSTEM : WINDOWS 7/10/11

System : PC OR LAPTOP
LANGUAGE USED : PYTHON
Processor : INTEL / AMD
Backend : PYTHON SCRIPT WINDOW
RAM : 4 GB
Frontend : PYTHON SHELL
ROM : 2 GB
 HARDWARE SPECIFICATION
Transmitter Block – Field Detection Unit :

Arduino UNO:

•Function: Acts as the central controller, collecting

sensor data and transmitting information via Zigbee.
•Purpose: Controls sensor readings, decision logic,
and communication.

Transformer & Rectifier

•Function: Converts 230V AC to ~12V AC → then

rectified to DC (5V/3.3V using voltage regulator).
•Transformer: 230V to 12V
•Regulated output: 5V for Arduino & sensors
Vibration Sensor :

•Common type: SW-420

•Function: Detects mechanical vibration or movement.
•Range: Very short-range (~1–2 cm), detects touches or impacts.
•Sensitivity: Adjustable via onboard potentiometer.

Accelerometer Sensor :

•Type: ADXL335 (analog)

•Function: Detects tilt, movement, or fall.
•Range: ADXL335: ±3g
•Use: Detects sudden ground shifts or changes in angle during landslides.

Sound Sensor :
•Type: KY-038 (analog + digital) sound sensor module
•Function: Detects sound intensity such as claps, knocks, or
environmental noise
•Range: ~1–2 meters for loud sounds (adjustable sensitivity)
•Use: Detects unusual ground noises or vibrations during landslides.
LCD Display :

•Type: 16x2 LCD (parallel)

•Function: Shows status of vibration,Accelerometer,sound sensor and GPS
location.
•Range: Display only.

Zigbee (TX) – Transmitter :

•Type: XBee Series 2 or similar

•Function: Sends data wirelessly to the receiver.
•Communication Range:
•Indoor: ~30–40 meter.
•Outdoor line of sight: ~100 meters (regular module), up to 1.5 km with
high-power XBee Pro
•Protocol: IEEE 802.15.4

GPS Module :

•Type: NEO-6M GPS

•Function: Provides location of the transmitter.
•Range:Accuracy: ~2.5 meters (open sky)
Time to First Fix (TTFF): ~30 seconds cold start
Communication: UART (9600 baud)
Receiver Block – Monitoring Station :

ESP8266 :

•Function: Acts as the IoT gateway

•Receives Zigbee data
•Sends it to a smartphone via Wi-Fi
•Can also post to cloud or display live values on a
mobile app
•Advantage: Enables remote monitoring.

Transformer & Rectifier :

•Function: Converts 230V AC to ~12V AC → then

rectified to DC (5V/3.3V using voltage regulator).
•Transformer: 230V to 12V
•Regulated output: 5V for Arduino & sensors
LCD Display :

•Type: 16x2 LCD (I2C or parallel)

•Function: Shows status of vibration detection and GPS location.
•Range: Display only – no range.

Zigbee (TX) – Transmitter :

•Type: XBee Series 2 or similar

•Function: Sends data wirelessly to the receiver.
•Communication Range:
•Indoor: ~30–40 meter.
•Outdoor line of sight: ~100 meters (regular module), up to 1.5 km with
high-power XBee Pro
•Protocol: IEEE 802.15.4

Blynk App :

•Function: Mobile app interface to view sensor data from ESP8266 over
Wi-Fi.
•Live Monitoring: GPS location, vibration alert
•Requirements: ESP8266 must be connected to Wi-Fi and programmed
with Blynk authentication token.
 ADVANTAGES
* The system continuously monitors environmental conditions such as ground movement, vibration, and
slope changes, allowing timely data acquisition.
* AI-based analysis enables early prediction of potential landslides, providing valuable time for evacuation
and disaster mitigation.
* Integration with the Blynk IoT platform allows users to monitor sensor data from anywhere through
smartphones or web interfaces.
* Zigbee communication eliminates the need for complex wiring, enabling flexible and cost-effective sensor
deployment in remote areas.
* Sensor values are displayed both locally (on LCDs) and remotely (via Blynk), ensuring accessibility for
both on-site personnel and remote users.
 CHALLENGES
1.Sensor Accuracy and Calibration:Ensuring the accuracy of the vibration, sound, and Accelerometer
sensors is crucial, as false signals can lead to incorrect victim identification.
2.Signal Penetration Through Soil:Moist or dense soil can weaken or block the signal.
3.False Positives and Negatives:The system may mistake animal movements leading to false positives.
4.Environmental Interference:External factors such as rain, rocks, and loose debris may interfere with
sensor readings.
5.Real-time Data Processing and Communication Latency:Delays in data transmission or processing
can hinder timely victim identification.
 FUTURE PLAN
1. AI and ML integration:
Integrate AI and machine learning (ML) models to improve detection accuracy.
2.Long-Range Communication with LoRa or 5G:
Replace the Soil-to-Soil communication with LoRa (Long Range) or 5G modules for more robust
and long-distance communication.
3. Portable Rescue Kit for Field Deployment:
Rescue teams can carry the kit to disaster zones
4. Multi-Terrain and Disaster Adaptation:
Modify the system to detect victims in landslides, flood etc.
 TIMELINE
Phase Task Duration Status
Phase 1:
Literature review, Technology study,Block diagram,Circuit
Research & 2 weeks Completed
Design
Design

Phase 2
Individual sensor testing,Data collection & AI algorithm for
prototying & 2 weeks Completed
Sensor fusion
AI model

Phase 3: Completed
Microcontroller,Sensor & Communication Module 2 weeks
Integration

Phase 4: Incomplete
Field testing,debugging,optimization 3 weeks
Testing

Phase 5: Incomplete
Final prototype,documentation & presentation 2 weeks
Deployment
 ESTIMATED COST
CATEGORY COMPONENET ESTIMATED COST

Viration,Accelerometer,Sound
SENSORS Rs. 1100
sensor

MICROCONTROLLER
Arduino uno,ESP8266 Rs. 700,400
&
Zigbee Tx-Rx,GPS module Rs. 1200,850
COMMUNICATION

Rs. 580
DISPLAY & ALERT LCD,
Rs. 960
POWER SUPPLY Battery Regulator (or)Transformer

COMMUNICATION Rs. 300

USB TTL
INTERFACE &POWER Rs. 140
Bridge Rectifier
ELECTRONICS

SOFTWARE/
Breadboard Tools Rs. 700
PROTOTYPING
Backup/Transport Rs. 700
MISCELLANEOUS

TOTAL COST : Rs . 7630

 RESULTS
The implemented system successfully demonstrated real-time monitoring and wireless transmission of
environmental parameters related to landslide activity. The Arduino Uno-based transmitter unit effectively
captured data from the IR sensor, vibration sensor, and accelerometer, which was displayed on the local LCD
and reliably transmitted to the ESP8266-based receiver unit using Zigbee communication. The receiver unit
displayed the received sensor data on its own LCD and uploaded it to the Blynk IoT platform, enabling users to
remotely monitor ground conditions through a mobile app or web interface. Sensor data was also forwarded to
a PC, where the AI model processed and analyzed patterns to detect potential landslide conditions. During
testing, the AI model accurately identified abnormal sensor behavior associated with simulated ground
movement and successfully triggered alerts through the Blynk platform. The system exhibited stable wireless
performance, accurate data logging, and responsive alert generation, validating its effectiveness for early
warning and real-time landslide risk assessment.
 CONCLUSIONS
The developed system effectively combines environmental sensing, wireless communication, IoT
connectivity, and AI-based analysis to provide a reliable and intelligent landslide detection and early warning
solution. By utilizing an Arduino Uno for sensor data acquisition and an ESP8266 for data processing and IoT
integration, the system enables real-time monitoring and remote accessibility through the Blynk platform. The
use of an AI model on a PC for landslide prediction enhances the system’s accuracy by analyzing complex
patterns rather than relying solely on fixed sensor thresholds. Testing confirmed that the system is capable of
detecting abnormal ground behavior and issuing timely alerts, making it a valuable tool for disaster
preparedness in landslide-prone regions. The modular and cost-effective design also allows for easy
deployment and scalability, making it suitable for a wide range of environmental monitoring applications.
 REFERENCES
1. S. Nag, M. Singh, R. K. Sharma “ Human Detection under Debris using Ultra Wideband Radar ”
IEEE Sensors Journal
2019.
2. J. Li, Y. Zhang, H. Chen “ Life Detection Technology Based on CO₂ and VOC Sensors ”
International Journal of Disaster Risk Reduction 2020.
3. A. Mehta, T. Kumar, L. Verma “ Deep Learning Approach for Post-Disaster Victim Detection ”
Elsevier - Expert Systems with Applications 202.
4. K. Roy, M. Banerjee “ Multi-Sensor Fusion for Survivor Detection in Search and Rescue ”
Springer - Journal of Ambient Intelligence and Humanized Computing 2022.
5. L. Wang, D. Zhao, F. Liu “ Application of Infrared Thermal Imaging for Post-Disaster Human
Detection ” Journal of Infrared and Millimeter Waves 2018.
THANK
YOU