Pothole Detection And Visualization About Me I am a 3 rd year - - PowerPoint PPT Presentation

pothole detection and visualization about me
SMART_READER_LITE
LIVE PREVIEW

Pothole Detection And Visualization About Me I am a 3 rd year - - PowerPoint PPT Presentation

Pothole Detection And Visualization About Me I am a 3 rd year Undergraduate Student at UCET, Bikaner, currently known as Bikaner Technical University( BTU). I am pursing my B.Tech degree in Computer Science and Engineering. Skills


slide-1
SLIDE 1

Pothole Detection And Visualization

slide-2
SLIDE 2

About Me

  • I am a 3rd year Undergraduate Student at UCET, Bikaner, currently

known as Bikaner Technical University( BTU).

  • I am pursing my B.Tech degree in Computer Science and Engineering.
  • Skills
  • Deep Learning
  • Artjfjcial Neural Network
  • Machine Learning
  • Data Science
  • GUI with Python and Java
slide-3
SLIDE 3

Kind of Roads we Like

Kind of Roads we Hate

slide-4
SLIDE 4

Road Accidents due to Potholes

slide-5
SLIDE 5
slide-6
SLIDE 6

T echniques for Pothole Detection

Mainly there are three main techniques:-

  • Thermal Imaging
  • Image Processing and CNN
  • Sensor Detectjon
slide-7
SLIDE 7
  • 1. Thermal Imaging
slide-8
SLIDE 8
  • 2. Image Processing and CNN
slide-9
SLIDE 9
  • 3. Sensor Detection
slide-10
SLIDE 10

Any Incident that Triggered this idea

slide-11
SLIDE 11

Image Augmentation

slide-12
SLIDE 12

Artifjcial Neural Network

slide-13
SLIDE 13

Convolutional Neural Network

slide-14
SLIDE 14

Working:

slide-15
SLIDE 15

Kernel

  • In image processing, a

kernel, convolutjon matrix, or mask is a small matrix. It is used for blurring, sharpening, embossing, edge detectjon, and more. This is accomplished by doing a convolutjon between a kernel and an image.

  • The general expression of a convolutjon is:

g(x,y)= w*f(x,y)

slide-16
SLIDE 16

Convolutional 2D

slide-17
SLIDE 17

MaxPooling

slide-18
SLIDE 18

Dropout

slide-19
SLIDE 19

Activation Function

  • The actjvatjon functjon is the non linear transformatjon that we do
  • ver the input signal. This transformed output is then sent to the next

layer of neurons as input.

  • Names of Actjvatjon Functjon:

Sigmoid Functjon Tanh Functjon ReLU Functjon Leaky ReLU Functjon Maxout Functjon ELU Functjon Sofumax Functjon

slide-20
SLIDE 20

Sigmoid Function ReLU Function

slide-21
SLIDE 21

Overfjtting and Underfjtting

slide-22
SLIDE 22

Model Description

slide-23
SLIDE 23

Performance Curves

slide-24
SLIDE 24

Let’s talk about Android

Only A Little

  • Google Map API
  • Polyline
slide-25
SLIDE 25

Google Map API

slide-26
SLIDE 26

Polyline

slide-27
SLIDE 27

Implementation

  • Detectjon
  • GPS(Global Positjoning System)
  • Array List
  • Polyline
  • Color coding
slide-28
SLIDE 28

App’s icon

slide-29
SLIDE 29

Demo Video of the App

slide-30
SLIDE 30

Demo Video Of Navigation

slide-31
SLIDE 31

Any Questions

slide-32
SLIDE 32