Engineering@Intuit  | Ex-PharmEasy | Ex-ARW | connect for tech consultations | book a mentorship session at madalgos.in/

Sumalya Saha  • Peerlist

The smart traffic system is mainly based on image processing an machine learning system
.The problem of traffic congestion by the manual traffic system can be solved or at least decreased by implementing this method. In order to implement this system, we first need to understand the basic needs and concepts of machine learning and image process. Various image processing techniques like thresholding,finding contours, feature detection algorithm, support vector machine(SVM) will be 
Utilised.This is an ongoing project.



SMART TRAFFIC SYSTEM(on-going)

It detects the motion of human beings by thermal mapping of the atmosphere and whenever motion is detected the lights are turned on. It can be used in any restricted area to catch any trespasser and can also be applied in household purposes like store rooms in order to avoid any accident or stumbling due to obstacles.

The whole thing has been executed analogously without the use of microcontroller to make it cost-effective. In this project, a PIR motion sensor, Light Dependent Resistor(LDR), 555 timer, relay module and a bulb have been used.


MOTION SENSORED LIGHTS

It is a Department of Science and Technology (DST) approved and funded project of the session 2015-2016.

It basically aims at building a smart home having all security aminities.The key features of the project are- 1.Remote web app based control over all household appliances 2.Voice-based control over appliances and operation modes 3.Security from breaches. 4. Household monitoring 5.Fire and Earthquake Alarm 6.Wireless communication 7.Internet-based routine testing for component malfunction.

Here,ESP8266(wifi-module),GSM SIM 900 module, ultrasonic sensors(HC-SR04),Ethernet Sheild,Temperature Sensors(DHT-11),wifi cameras ,gas sensors have been utilized


Intelligent Home Automation and Security System

An mobile app which displays a colour sphere is used to change the colour of the led matrix ie.the RGB  led matrix gets the corresponding colour of the colour sphere where it is touched and the app will also return the temperature of the region.

BLUETOOTH CONTROLLED LIGHTING SYSTEM AND DATA LOGGER

The high resolution satellite image is processed by deep Convolutional Neural Networks (CNN) techniques and basic innovative image processing methods and feature detection algorithms to identify the road segments.The system can be used to monitor the traffic system and GPS system
This is an ongoing project.


AUTOMATIC EXTRACTION OF ROAD SEGMENTS FROM SATELLITE IMAGES(

A driver circuit has been proposed which has been tested by simulation software MULTISIM which shows the efficiency of the circuit of about 90%.This project as a form of paper has been sent for the approval to be published in IEEE journal. To give a preview of the project, the abstract is furnished below:-

An LED driver refers to an electrical device which helps to regulate the power to an LED or to an array of LEDs. The LED driver adapts to the needs of the LED, by providing a continuous quantity of power to the LED when its electrical properties vary. LED's are as a matter of fact sensitive to the voltage used to power them (i.e. the current changes a lot with a small change in voltage).To avert the LED from becoming unstable due to voltage changes, an LED driver is needed. An LED driver is a self-contained power supply which has outputs that are affianced to the electrical characteristics of the LED or to the array of LEDs. LED drivers may offer dimming by means of pulse width modulation circuits and may have different channels for separate control of different LEDs or LED arrays. Without the most suited driver, the LED may become hot and unstable, therefore causing poor performance or failure. This paper presents a novel approach to the design of a LED Driver using various power transistors, power mosfets, operational amplifier and feedback circuits. The experimental results demonstrate the power delivered to the LED at different supply voltages upholding the performance of the driver circuit proposed.



LED DRIVER WITH COST EFFECTIVE HIGH EFFECIENCY

It estimates the volume of any object placed in front of the device without any camera surveillance but using radar construction with sonar. It also maps the image of the object in 3-D view in the computer monitor. The project finds its application in coal mines,shipyards,airports etc.It decreases the use of camera in various sectors.

Here, ultrasonic sensors have been used to measure the distances from the obstacle .It scans the obstacle in 3 dimensions with a definite velocity and measures the time in a particular dimension up to which the object distance does not fluctuate beyond a permissible level. Hence , that time maps into the length, breadth or height with the help of the given velocity of scanning..Therefore, we get the approximate volume and shape of the body.

The Visual-Python has been incorporated in order to view the 3-D image of the object. The project  was ranked 3rd in the NEN Scientific Model Making Competition(2016)..


3-D MAPPING OF OBJECTS

In this project an autonomous robot is built which can travel within its vicinity and move towards human beings(detecting the face) present in the neighbourhood and communicate with them. The project has significant importance in the field of robot surveillance.

The face detection algorithm of image processing has been applied here. The open-CV software has been used to incorporate various image processing techniques.



HUMAN DETECTION SYSTEM

The robot will detect the obstacle and will remove it to make its own path. The project may be applied in the automation of civil sector. The modified version of the prototype cab be used to guide blind people in their locomotion