AUTOMATED SOLAR CAVITY DETECTION IMAGE PROCESSING PATTERN RECOGNITION

AUTOMATED SOLAR CAVITY DETECTION IMAGE PROCESSING & PATTERN RECOGNITION Athena Johnson 1

OUTLINE • Introduction • Background • Problem Statement • Proposed Solution • Experiments • Conclusions • Future Work 2

INTRODUCTION 3

BACKGROUND • Solar Dynamics Observatory (SDO) • Extreme Ultraviolet Variability Experiment (EVE) • Helioseismic and Magnetic Imager (HMI) • Atmospheric Imaging Assembly (AIA) • 1. 5 Terabytes (TB) of data per day 4

ATMOSPHERIC IMAGING ASSEMBLY (AIA) • Images the Corona of the Sun • Study of solar storms • • How they are created? How they propagate upward? How they emerge from the Sun? How magnetic fields heat the corona? 5

SOLAR CAVITIES • Currently an increase in implementations focused on Solar Cavities • Off limb structures • Darker elliptical structure, encompassed by lighter regions • Hypothesized to be precursors to solar events • Aid in establishing a predictive solar weather system 6

SOLAR CAVITIES • Labrosse, Dalla and Marshall (2010) • Radial intensity profiles • Support Vector Machine (SVM) • Region growing • Calculation of metrics • Running difference on subsequent images 7

SOLAR CAVITIES • Durak and Nasraoui (2010) • Exraction of principal contours • Calculations on contours • Adaboost 8

PROBLEM STATEMENT • Computation times • Detections based on metrics • Weak events missed • Multiple detections • Multiple events missed • Low hit rates 9

HAAR CLASSIFIER • Method that Paul Viola and Michael Jones published in 2001 • Four key concepts • Haar-like features • Integral Image • Adaboosting • Cascade of Classifiers 10

HAAR-LIKE FEATURES • Aids in satisfying real time requirements • Rectangular regions • Reduces Computation 11

INTEGRAL IMAGES • Rapid computation of Haar-like features 12

INTEGRAL IMAGES Original Image 8+6+2+5+6+3 = 30 Integral Image 50 -17 -5+2 = 30 13

ADABOOSTING • Aids in increasing the accuracy and speed • Begins with uniform weights over training examples • Obtain a weak classifier • Update weights Weak Classifier h 1(x) 14

ADABOOSTING Weak Classifier h 2(x) Weak Classifier h 3(x) 15

ADABOOSTING • Weak classifiers combined to form the strong classifier 16

CASCADE OF CLASSIFIERS • Increases the speed of detections • All Haar-like features from all stages combined into a final Classifier Model • Cascade of boosted classifiers with Haar-like features 17

CASCADE OF CLASSIFIERS • A series of classifiers are applied to every subwindow of image • A positive result from the first classifier, triggers evaluation from the second classifier and so on 18

INITIAL SOLUTION 19

RESULTS • Manually selected Training Image Sets • Positive Samples = 100 • Negative Samples = 400 • ≈ 79. 6% Correct detection rate was achieved 20

RESULTS • Missed detections in specific quadrants • Detections on the Sun’s disk • Overlapping detections 21

PROPOSED SOLUTION 22

MINIMIZED TRAINING SETS 10 Positive Images 10 Negative Images 23

MARK REGIONS OF INTEREST AND ROTATE • Deriving images from selected images • Rotation applied to both training sets 24

TRANSFORM REGIONS OF INTEREST • Transformations on cavities 25

PREPROCESSING • Edge Detection • Hough Lines • Calculate the radius 26

RESULTS • Derived Training Image Sets • Initial image in sets = 10 • Positive Samples = 3600 • Negative Samples = 3600 • ≈ 96% Correct detection rate was achieved 27

FINAL IMAGE WITH DETECTIONS 28

CONCLUSION • Less manual work • Short training times • < 22 hours • Wider range of detections • Weak and strong cavities • Fast run times • < 1 second per image • Higher hit rates 29

FUTURE WORK • Technique Improvement • Reduction of False Positives • Contour Detections • Template Matching • Customized Haar-like features 30

FUTURE WORK • Find optimal number of training sets • Extract Metrics • User Interface 31

QUESTIONS? 32