Deep Learning-based Smart Street Lamps – A Solution to Urban Pollution

Abstract

Smarter devices create an opportunity for numerous applications that contribute towards the welfare of the society. The system contributes in creating a street lighting system which focuses on solving the issues pertaining to environment, energy and public safety. The proposed system is a modular system having the components: LED module, PIR sensor, LDR sensor, Pollution sensor and a camera. The brightness of the lights is reduced to conserve energy in the absence of motion and when there is adequate ambient light. The Car tracking module of the system identifies the blacklisted cars using number plate identification. The number plate identification has been done using RCNN's along with some image processing algorithms. The air quality surrounding the street lamps can be detected and documented using the Pollution sensor. This data enables us to create a pollution heat map of the city and identifies areas of high traffic congestion. There is a charging port provided on a light pole for charging electric cars, which is the prepaid service.