Beruflich Dokumente
Kultur Dokumente
Team:
• Bharat Bhushan
• Mithra Desinguraj
A Street Cleanliness Assessment
System for Smart City using Mobile • Kavin Pradeep Sriram Kumar
Spring 2017
Introduction - Problem
Streets are the nerves of any city and society.
Keeping streets clean is a challenge for any city admins.
Street cleanliness assessment is essential but..
Problem:
Manual.
Offline data collection.
Time consuming.
No real-time visibility.
High Cost.
Introduction - Solution
Proposed Solution:
Smart City Street Assessment system using Mobile and
Cloud.
Automated using mobile and cloud.
Real to near-real time data collection.
Less time.
Real-time visibility with single pane of glass.
Cost Effective.
Integration with other city services.
Public contribution via mobile (crowd sourcing).
Self Learning (Machine Learning)
API driven.
Mobile.
Smart City Street Cleaning Infrastructure
Cleaning Model - Layers
Cleaning Model - Areas
San Jose City
(95)
City
No. of Streets
Streets with
No. of Grid Pts
Individual
Photo Points.
Cleaning Model – Grid
Grid Point Model
Picture Point: Multiple Images are captured in each direction (F,B,L,R) on
either side of the street and sent to Cloud along with location data.
INode : Represent Intersection. There are several Image points between nodes.
SNode : Represent Sub Intersection. Used to divide large blocks.
Grid Point : Represent logical radius to assess both sides of the street. It can
have one or more Pic points. Collectively produces the cleanliness level across
the street.
(S) (E)
Computation – Block level
Grid based analysis.
Aggregate of all the points in the block.
Assessment would be based on every street in the block and the aggregate value.
Results are stored in DB with image reference, date time and level.
Block can have any number of streets, everything is based on each data points.
Computation – Area level
Assumptions:
Fixed image resolution.
Vehicle speed is approx.15mph.
Picture set covers 20ft. of distance.
Pictures are collected every ~2-4 sec.
Multiple set of pictures are collected every time.
Stable Network connectivity for real time update.
Offline image transmission (batch transfer option).
Infrastructure
Cloud
Edge
Images sent to
Cloud via Mobile or
City Wi-Fi (batch)
Image Core DB
Service Processing Engine (MySQL) Residents
App
Edge Storage
Cloudlet
Map Service
External
System Architecture
Mobile Client Street Cleaning UI Street Cleaning Dashboard
(MS)
Street Cleaning Reports Street Cleaning Detection Engine
Controller
Mobile Street Cleaning Detection Analytics
Streets Blocks
Stations
MS Computing Street Cleaning DB service
Historical Engine DB Application DB
MS Monitoring (NoSQL) (MySQL)