Traffic Lights Detection Using Blob Analysis and Pattern Recognition

Slides:



Advertisements
Similar presentations
Miroslav Hlaváč Martin Kozák Fish position determination in 3D space by stereo vision.
Advertisements

Road-Sign Detection and Recognition Based on Support Vector Machines Saturnino, Sergio et al. Yunjia Man ECG 782 Dr. Brendan.
Mahmoud Abdallah Daniel Eiland. The detection of traffic signals within a moving video is problematic due to issues caused by: Low-light, Day and Night.
CSSE463: Image Recognition Day 2 Roll call Roll call Announcements: Announcements: Moodle has drop box for Lab 1 Moodle has drop box for Lab 1 Next class:
LING 111 Teaching Demo By Tenghui Zhu Introduction to Edge Detection Image Segmentation.
COLORCOLOR A SET OF CODES GENERATED BY THE BRAİN How do you quantify? How do you use?
AlgirdasBeinaravičius Gediminas Mazrimas Salman Mosslem.
Color Image Processing
Vehicle License Plate (VLP) Recognition System By German H. Flores and Gurpal Bhoot.
CSSE463: Image Recognition Day 2 Roll call Roll call Announcements: Announcements: Reinstall Matlab if you are having problems: Lab 1 has directions. Reinstall.
CSSE463: Image Recognition Day 2 Roll call Roll call Announcements: Announcements: Room change: Olin 267 starting next class Room change: Olin 267 starting.
Recognition of Traffic Lights in Live Video Streams on Mobile Devices
How computers answer questions An introduction to machine learning Peter Barnum August 7, 2008.
Yiming Zhang SUNY at Buffalo TRAFFIC SIGN RECOGNITION WITH COLOR IMAGE.
Image Processing and Pattern Recognition Jouko Lampinen.
Binary Image Analysis: Part 2 Readings: Chapter 3: mathematical morphology region properties region adjacency 1.
Color: Readings: Ch 6: color spaces color histograms color segmentation.
Texture Classification Based on Co-occurrence Matrices Presentation III Pattern Recognition Mohammed Jirari Spring 2003.
Tutorial # 9 Nov. 21, Segmentation  Isolating a region/regions of interest in an image  Useful for:  Collect more meaningful data from an image.
Feature extraction Feature extraction involves finding features of the segmented image. Usually performed on a binary image produced from.
Brief overview of ideas In this introductory lecture I will show short explanations of basic image processing methods In next lectures we will go into.
Face Detection and Neural Networks Todd Wittman Math 8600: Image Analysis Prof. Jackie Shen December 2001.
Lesson 2.2 Traffic Signals  Today’s Objectives: 1. Explain what to do at a green, yellow, and red light. 2. Describe the action to take when you approach.
Machine Vision for Robots
OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The.
Page: 1 PHAM VAN Tien Real-Time Approach for Auto-Adjusting Vision System Reading Class International Graduate School of Dynamic Intelligent Systems.
Ajay Kumar, Member, IEEE, and David Zhang, Senior Member, IEEE.
ENT 273 Object Recognition and Feature Detection Hema C.R.
Intelligent Vision Systems ENT 496 Object Shape Identification and Representation Hema C.R. Lecture 7.
Algirdas Beinaravičius Gediminas Mazrimas Salman Mosslem.
Image Processing and Pattern Recognition Jouko Lampinen.
Pattern Recognition April 19, 2007 Suggested Reading: Horn Chapter 14.
By: Jared Meyer A Neural Network Approach to Classifying Cartoons Based on Color ECE 539 Final Project.
Chapter 4: Pattern Recognition. Classification is a process that assigns a label to an object according to some representation of the object’s properties.
Using PowerPoint Hands-On Tutorial. Importing Clipart Click the Insert menu Select Picture Select Clipart On the search box Enter Traffic and click the.
11 To call into question Questioning of a statement or fact Demand for justification or explanation Challenge.
Tree and leaf recognition
November 30, PATTERN RECOGNITION. November 30, TEXTURE CLASSIFICATION PROJECT Characterize each texture so as to differentiate it from one.
Post-test Review Please try your best...ask questions and read the question completely!!!
Playing Card Recognizer ECE 4025 February 28, 2002 Group 5 Robert Barrett Jason Hodkin Chung Tse Mar Jay Silver David Winkler Yu Ming Wu.
1 Computational Vision CSCI 363, Fall 2012 Lecture 12 Review for Exam 1.
Lesson 2.3 Today’s Objectives: Describe the difference between broken yellow lines and broken white lines. Explain the differences between a shared left-turn.
EE368 Final Project Spring 2003
IMAGE PROCESSING Tadas Rimavičius.
Signals, pavement markings, and proper turns QUIZ
Image Processing and Pattern Recognition
Author : Sang Hwa Lee, Junyeong Choi, and Jong-Il Park
Color Image Processing
Color Image Processing
A Personal Tour of Machine Learning and Its Applications
ივანე ჯავახიშვილის სახელობის
Color Image Processing
Figure Legend: From: Color constancy and hue scaling
Color: Readings: Ch 6: color spaces color histograms
Brain Hemorrhage Detection and Classification Steps
Color: Readings: Ch 6: color spaces color histograms
Color Image Processing
Colour Theories.
Blinkers ++ Team 5.
Creating Data Representations
Colors Computers build colors from Red, Green, and Blue; not Red, Blue, and Yellow. RGB = Red Green Blue Creating Colors Red + Blue = Purple No Red, No.
Color Image Processing
Binary Image Analysis: Part 2 Readings: Chapter 3:
Midterm Exam Closed book, notes, computer Similar to test 1 in format:
CSSE463: Image Recognition Day 2
Chapter 40.
Midterm Exam Closed book, notes, computer Similar to test 1 in format:
CSSE463: Image Recognition Day 2
THE TOPICS AND TITLES OF RESEARCH
Surveillance Group Weekly Report
Presentation transcript:

Traffic Lights Detection Using Blob Analysis and Pattern Recognition Jaromír Zavadil

Competition Signaling Panels [Robotica 2011]

Task to solve Symbols to be recognized

Methods used Color Segmentation Blob analysis Pattern Recognition HSV Color Space Blob analysis regionprops() Pattern Recognition Mahalanobis distance [MathWorks]

Blob Analysis regionprops() MajorAxisLength Area MinorAxisLength Solidity BoundingBox Eccentricity Centroid Extent Perimeter Orientation EulerNumber

Blobs Green Arrow Yellow Arrow Red Cross Area > 200; Eccentricity < 0.9; Extent > 0.4; EulerNumber > -20; Solidity < 0.83; 60 < Orientation < - 60 Yellow Arrow Area > 200; Eccentricity < 0.9; Extent > 0.35; EulerNumber > -8; Solidity < 0.83; -25 < Orientation < 25 Red Cross Area > 200; Eccentricity < 0.7; 0.3 < Extent < 0.8; EulerNumber > -8; 0.4 < Solidity < 0.8; -25 < Orientation 25

Blobs Red and Green Chessboard In the end compare all found blobs Area > 40; Eccentricity < 0.97 if number of blobs > 7 compute number of pixels In the end compare all found blobs

Direction of the Yellow Arrow cut the blob using centroid and compare the left and the right part of the blob 508 342

Results Tested Images Results red cross left arrow green arrow right arrow red & green without light total missed wrong 92 113 73 49 103 20 450 34 2

False Positives

Pattern Recognition Mahalanobis distance In MATLAB – mahal() function >> d = mahal(X, Y); 𝑑 𝑀 𝑋,𝑌 = 𝑋−𝑌 𝑇 −1 𝑋−𝑌 Σ - covariance matrix X - reference sample Y - object to be classified X - reference sample Y - object to be classified

Patterns Descriptors Solidity Eccentricity Extent Form Factor Axis Proportion 𝑋= 𝑆 1 𝐸𝑐𝑐 1 𝑆 2 ⋮ 𝑆 𝑀 𝐸𝑐𝑐 2 ⋮ 𝐸𝑐𝑐 𝑀 𝐸𝑥 1 𝐸𝑥 2 ⋮ 𝐹 1 𝐹 2 ⋮ 𝐴 1 𝐴 2 ⋮ 𝐸𝑥 𝑀 𝐹 𝑀 𝐴 𝑀

Patterns … 12 examples for each symbol 10 very good images + 2 images with distortion … 1 2 𝑋= 𝑆 1 𝐸𝑐𝑐 1 𝑆 2 ⋮ 𝑆 12 𝐸𝑐𝑐 2 ⋮ 𝐸𝑐𝑐 12 𝐸𝑥 1 𝐸𝑥 2 ⋮ 𝐹 1 𝐹 2 ⋮ 𝐴 1 𝐴 2 ⋮ 𝐸𝑥 12 𝐹 12 𝐴 12 12

0 < H < 0.07 & 0.96 < H < 1; 0.2 < H < 0.54; S > 0.4; V > 0.4 Color Segmentation 0 < H < 0.07 & 0.96 < H < 1; S > 0.5; V > 0.4 0.11 < H < 0.2; S > 0.5; V > 0.4

Area > 35; Eccentricity < 0.98 Blobs Selection 0.55 < Solidity < 0.83; Extent > 0.35; - 25 < Orientation < 25; EulerNumber > -8 Area > 35; Eccentricity < 0.98

M. Distance < 100; Area > 150 Mahalanobis distance M. Distance < 300; Area > 150 M. Distance < 100

Results Tested Images Results red cross left arrow green arrow right arrow red & green without light total missed wrong 92 113 73 49 103 20 450 23

Missed Symbols dM = 152,8 dM = 189 too small blobs too small blobs

Maximum distances Arrows and Red Cross Red & Green Chessboard 2,5 m

Future work Build a bigger set of good examples. Compute probability of detected symbols. Try to use a neural network for classification. Try to process images from a real road.

Traffic Lights Detection Using Blob Analysis and Pattern Recognition Jaromír Zavadil