Image Processing Methods Using MATLAB

Slides:



Advertisements
Similar presentations
A Graph based Geometric Approach to Contour Extraction from Noisy Binary Images Amal Dev Parakkat, Jiju Peethambaran, Philumon Joseph and Ramanathan Muthuganapathy.
Advertisements

Boundary Detection - Edges Boundaries of objects –Usually different materials/orientations, intensity changes.
Image Processing Lecture 4
Chapter 3 Image Enhancement in the Spatial Domain.
Image Enhancement in the Spatial Domain II Jen-Chang Liu, 2006.
December 5, 2013Computer Vision Lecture 20: Hidden Markov Models/Depth 1 Stereo Vision Due to the limited resolution of images, increasing the baseline.
6/9/2015Digital Image Processing1. 2 Example Histogram.
Computer Vision - A Modern Approach Set: Linear Filters Slides by D.A. Forsyth Differentiation and convolution Recall Now this is linear and shift invariant,
Image processing. Image operations Operations on an image –Linear filtering –Non-linear filtering –Transformations –Noise removal –Segmentation.
Digital Image Processing
Video Google: Text Retrieval Approach to Object Matching in Videos Authors: Josef Sivic and Andrew Zisserman University of Oxford ICCV 2003.
The Segmentation Problem
Linear Filtering About modifying pixels based on neighborhood. Local methods simplest. Linear means linear combination of neighbors. Linear methods simplest.
Image Processing1 Image Filtering Filtering can be use to enhance some features and de- enhance others. Usually used as a pre-processing step.
Digital Images Chapter 8 Exploring the Digital Domain.
CS 376b Introduction to Computer Vision 02 / 26 / 2008 Instructor: Michael Eckmann.
Filtering and Enhancing Images. Major operations 1. Matching an image neighborhood with a pattern or mask 2. Convolution (FIR filtering)
UCF REU: Weeks 1 & 2. Gradient Code Gradient Direction of the Gradient: Calculating theta.
Digital Image Processing Lecture 3: Image Display & Enhancement March 2, 2005 Prof. Charlene Tsai.
December 9, 2014Computer Vision Lecture 23: Motion Analysis 1 Now we will talk about… Motion Analysis.
Digital Image Processing Lecture 10: Image Restoration March 28, 2005 Prof. Charlene Tsai.
Course Website: Digital Image Processing Image Enhancement (Spatial Filtering 1)
Digital Image Processing Lecture 10: Image Restoration
Digital Image Processing Lecture 16: Segmentation: Detection of Discontinuities Prof. Charlene Tsai.
Sejong Univ. CH3. Area Processes Convolutions Blurring Sharpening Averaging vs. Median Filtering.
Non-linear Filters Non-linear filter (nelineární filtr) –spatial non-linear operator that produces the output image array g(x,y) from the input image array.
Image enhancement Last update Heejune Ahn, SeoulTech.
Digital Image Processing Lecture 16: Segmentation: Detection of Discontinuities May 2, 2005 Prof. Charlene Tsai.
Digital Image Processing Lecture 17: Segmentation: Canny Edge Detector & Hough Transform Prof. Charlene Tsai.
: Chapter 5: Image Filtering 1 Montri Karnjanadecha ac.th/~montri Image Processing.
Instructor: Mircea Nicolescu Lecture 5 CS 485 / 685 Computer Vision.
Gholamreza Anbarjafari, PhD Video Lecturers on Digital Image Processing Digital Image Processing Spatial Domain Filtering: Part I.
Sharpening Spatial Filters ( high pass)  Previously we have looked at smoothing filters which remove fine detail  Sharpening spatial filters seek to.
Grauman Today: Image Filters Smooth/Sharpen Images... Find edges... Find waldo…
Video Google: Text Retrieval Approach to Object Matching in Videos Authors: Josef Sivic and Andrew Zisserman University of Oxford ICCV 2003.
Image Enhancement in the Spatial Domain.
Chapter 10 Digital Signal and Image Processing
Spatial Image Enhancement
Image Representation and Description – Representation Schemes
Image Filtering Spatial filtering
Digital Image Processing Lecture 16: Segmentation: Detection of Discontinuities Prof. Charlene Tsai.
Digital Image Processing Lecture 10: Image Restoration
Linear Filters and Edges Chapters 7 and 8
Detection of discontinuity using
IMAGE PROCESSING INTENSITY TRANSFORMATION AND SPATIAL FILTERING
Video Google: Text Retrieval Approach to Object Matching in Videos
Error Concealment In The Pixel Domain And MATLAB commands
Chapter I Digital Imaging Fundamentals
Mean Shift Segmentation
See Through Fog Imaging Project: P06441
- photometric aspects of image formation gray level images
Computer Vision Lecture 9: Edge Detection II
Content-Based Image Retrieval
Content-Based Image Retrieval
Digital Image Processing
Lecture 3 (2.5.07) Image Enhancement in Spatial Domain
Image Processing, Lecture #8
Digital Image Processing
Digital Image Processing
Image Processing, Lecture #8
Spatial operations and transformations
: Chapter 8: Edge Detection
Circles! You are going to create an “image” with circle(s)
Increasing Sensitivity of Ca2+ Spark Detection in Noisy Images by Application of a Matched-Filter Object Detection Algorithm  Cherrie H.T. Kong, Christian.
Video Google: Text Retrieval Approach to Object Matching in Videos

Digital Image Processing Lecture 3: Image Display & Enhancement
Lecture 7 Spatial filtering.
Spatial operations and transformations
DIGITAL IMAGE PROCESSING Elective 3 (5th Sem.)
Presentation transcript:

Image Processing Methods Using MATLAB Beth Bogart and Andrew Sevigny Tuesday October 20, 2015

Overview Test Video: what we’ve been processing Processing Methods Image Averaging Binary Images Using Thresholding Next Steps B

Video

Seeing an Image

Averaging a Set of Images Create an Empty Array for the New Image Define the Image Range: Which Frames to Average Iterate Though Each Pixel Determine the avg. pixel value across frames Store the average value in the new image array Display the new image Noise reduction, smoothing, helps remove pixels that may affect further processing

Averaging a Set of Images: Before Frame 18 Frame 22 BEFORE Noise reduction, smoothing, helps remove pixels that may affect further processing Frame 26 Frame 29

Averaging a Set of Images: After Noise reduction, smoothing, helps remove pixels that may affect further processing

Benefits of Averaging BEFORE AFTER Noise reduction, smoothing, helps remove pixels that may affect further processing

Thresholding Theory Kernel Convolution http://micro.magnet.fsu.edu/primer/java/digitalimaging/processing/kernelmaskoperation/ Theory behind kernel convolution: Uses a box convolution kernel, of an odd number of pixels (3x3, 5x5, etc), that assigns weighting to each of the outer pixels around the target pixel and depending on what you’re trying to do (sharpen or smooth) will take the average of those surrounding pixels and re-assigns the target pixel the average value, which will allow the hue of the image to be reassigned. Goes along for every single pixel per frame, and output is the completely averaged/summed pixels.

Our Thresholding Technique Similar theory as video Compares target pixel to averaged pixels Old Image New Image Comparing average to threshold value and assigning the target pixel to a one or zero corresponding to the threshold, and will turn the “defects”, in our case, yellow and all other values within the threshold blue anyting cooling differently to the pixels around it will be brought out by the code We can change the kernel size as well as the threshold values in order to mix and match to find out what the best outcome is (least amount of noise, most amount of actual data text: clearly read the text) Value TValue AVG Find Average of Kernel

Our Thresholding Technique Set a Threshold for Error Choose an image to process Create an empty array for the new image Iterate through every pixel (starts w/in boundaries) Define a kernel of surrounding pixels Take average of pixels within a set kernel Is the pixel within the threshold value? Yes: Set binary image pixel to 1 No: Set binary image pixel to 0 Display the new image Noise reduction, smoothing, helps remove pixels that may affect further processing

Binary Results Binary Average of Original Frame 18 RAW Frame 18

Binary Results: Using Averages Binary Average of Original Frame 18 Binary Average of Average Frame (Frames 18-29)

Binary Results: Changing the Kernel 9x9 7x7 5x5

Binary Results: Changing the Threshold X = 0.04 X = 0.05 X = 0.06 X = the amount of error over the average value from the original pixel value

System Integration Recorded videos need to be short Video timed with the flash Faster data processing Removing unnecessary frames Uniformity: lighting and camera angle Talk about the time of each video talk about the duration of flash/cooling image needs uniformity in lighting and camera angle

Uniformity

Looking ahead: applying derivatives Applying derivative techniques Graphing pixel values over time (focused areas) Taking 2nd derivative of each pixel’s function Creating binary image Graph 1 X = time on natural log scale Y = change in temperature on natural log scale Graph 2 Y = change in the rate of temperature with respect to time, nat log scale Graph 3 X = time on a log scale Y = change in the change in rate of temperature with respect to time spiedigitallibrary.org/data/Journals/OPTICE/21928/051008_1.pdf

Looking ahead: applying derivatives spiedigitallibrary.org/data/Journals/OPTICE/21928/051008_1.pdf

Looking ahead (cont.) Averaging out image noise Depends on output Automation Integrating MATLAB into our system

What Have You Seen? Issues with our current system Processing Methods Image Averaging Binary Images Using Thresholding Issues with our current system Thermographic Derivatives