Lab Report #8 Image Processing

docx

School

Rochester Institute of Technology *

*We aren’t endorsed by this school

Course

111

Subject

Electrical Engineering

Date

Apr 3, 2024

Type

docx

Pages

Uploaded by DoctorDugongMaster1147

Lab #8 Image Processing Laboratory Report Fundaments of Imaging Science Instructor: Dr. Anthony Vodacek Stephen D. Gayle Section 1 1 8/26/2019

Introduction The purpose of this experiment is to apply some digital image processing techniques to a series of grayscale and color images to learn how these techniques work to enhance image content. Digital image processing is widely used to enhance or extract features from images and to manipulate the displays of images. The goal is often to make the visual display of images better by increasing contrast or highlighting certain parts of an image. In some cases, the processing can lead to information about the image content that is not necessarily displayed but may be a numerical quantity or an identification of image content. This lab will give you experience with manipulating histograms, applying median filters, mean filters, edge detection filters, manipulating color images, and using Lookup Tables to enhance grayscale images. Methods For this lab, I used the ImageJ program to process the images. I downloaded the figure 1 from the ImageJ site and then started processing it. Following the steps within the Lab guide produced figures 1A, 1B, 1BA, 1C and 1D. Figure 1A came from analyzing figure 1 and put into the form of a histogram. Figure 1B came from adjusting the brightness of the image to display no black on the image, which is called stretch. Figure 1BA came from analyzing figure 1B and put into the form of a histogram. Figure 1C came from the noise reduction of figure 1B using the median filter. Figure 1D came from the same process as figure 1C but used the mean filer instead. Next, I downloaded the figure 2 from the ImageJ site and then started processing it. Following the steps within the Lab guide produced figures 2A, 2B and 2C. Figure 2A came from changing the kernel to numbers that will detect horizontal lines. Figure 2B came from changing the kernel to numbers that will detect vertical lines. Figure 2C came from the combining of the figures 2A and 2B. Then, I downloaded the figure 3 from the ImageJ site and then started processing it. Following the steps within the Lab guide produced figures 3A, 3B, 3C and 3D. Figure 3A came from splitting the colors of the image into red, green and blue and stretching(just like figure 1B) them to a point where each of them was showing the least amount of black, but not dipping into overexposure and then merging the 3 images back to one. Figure 3B came from once again splitting the colors of the original image but applying a mean filter over the blue split image with a large radius, which leads to blurring of the blue split image and then merging the 3 images into one. Figure 3C is the same process as Figure 3B but blurring green instead. Figure 3D is the same process as Figure 3B but blurring red instead. Lastly, I downloaded the figure 4 from the ImageJ site and then started processing it. Following the steps within the Lab guide produced figures 4A, 4B and 4C. For all three I played around with all the LUT’s until I found what looked appealing (figure 4A) and 2 that looked gross to me (figure 4B and 4C). 2 8/26/2019

Figure 1 – Enhance me Figure 2 – Boat 3 8/26/2019

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

Figure 3 – Baboon Figure 4 – TDM filter Results Figure 1A – Histogram for figure 1 4 8/26/2019

Figure 1B – Enhance me stretched Figure 1B looks much clearer image-wise, unlike the original (figure 1) where you can barely see the face. All there is left to do is get rid of the noise. Figure 1BA – Histogram of Enhance me stretched Figure 1BA is the histogram of figure 1B 5 8/26/2019

Figure 1C – Enhance me stretched after noise reduction (median) Figure 1C is the noise reduction of figure 1B, showcasing no white dots. Figure 1D – Enhance me stretched after noise reduction (mean) Figure 1D is the noise reduction of figure 1B, showcasing the worsening of the white dots. 6 8/26/2019

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

Figure 2A – highlighting horizontal lines of boat Figure 2A is the highlighting of the horizontal lines of figure 2. Figure 2B – highlighting vertical lines of boat Figure 2B is the highlighting of the vertical lines of figure 2 . 7 8/26/2019

Figure 2C – combined horizontal and vertical lines of boat Figure 2C is the combining of horizontal and vertical lines of figure 2, showing where exactly detected edges are. Figure 3A – Baboon stretched Figure 3A is figure 3 with the contrast of each RGB turned up to different degrees. 8 8/26/2019

Figure 3B – Blue blurred baboon Figure 3B is figure 3 with the B in RGB blurred Figure 3C – Green blurred baboon Figure 3C is figure 3 with the G in RGB blurred 9 8/26/2019

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

Figured 3D – Red blurred baboon Figure 3D is figure 3 with the R in RGB blurred Figure 4A – best visualization LUT (Inverted 5 shades) Figure 4A is the best looking LUT of the bunch. 10 8/26/2019

Figure 4B – less satisfying LUT (6 Shades) Figure 4B is one of the worst looking LUT because of how bright and non-complementing colors. Figure 4C – less satisfying LUT (Spectrum) Figure 4C is one of the worst looking LUT because of how bright and non-complementing colors. Discussion 11 8/26/2019

In the first rounds of image processing, figure 1 was very hard to make out exactly what it was unless you concentrated heavily on the image, but after the stretching of the image (figure 1B) we can clearly see the face within the image with some noise. The differences in histograms from the stretching can be seen effecting the mean and standard dev. in figure 1A and 1BA. This process of stretching can be useful because it can help reveal hidden messages or never before seen object/people within images and noise reductions of the median filter can help in this regard (figure 1C) as it gets rid of the white dots within the image making it clearer. Depending on the image, you would use either a median filer (figure 1C) or a mean filter (figure 1D) for noise reduction because as you can see in figure 1D, it made the white dots worst. In the second rounds of image processing, figure 2 showcases a boat and then after some processing via change in convolve kernel to detect horizontal and vertical line. Which produced figures 2 A and 2B highlighting what counts as horizontal and vertical lines. Therefore, leading to the last processed image figure 2C. 2C looks very comparable to figure 2, just with some discontinuities in the lineage and a black background. The usefulness of this processing technique is it can show where exactly the edges are in a forest fire and in which direction the spread is going. In the third round of image processing, figure 3 showcases a baboon’s face up close. Comparing figure 3 and 3A it looks as though the image has become faded because of the stretching process. Showing that stretching the image can lead to a worst image. As for comparing figure 3 to 3B, 3C, and 3D, we can see that the 3B and 3D look near identical to the original, but 3C looks slightly blurry therefore leading to the conclusion that green is the most predominant color in the image. This process can help with edge detection with seeing where colors are and smooth transition between colors. In the fourth round of image processing, figure 4 showcases a white image white dark spots on it with other hard-to-spot spots. Figure 4A I believe highlights all the spots with a better visual than the original and showing other spots I didn’t even see in the original. While figures 4B and 4C are honestly gross to look at, they are also hard to look at and focus on certain things. 12 8/26/2019

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version