Catalogue of digital image processing experiments

matlab Learn and operate Fourier transform and frequency domain processing of images

matlab Learning and operation
Experimental content ：
6. Read in two color images , And display the size information of the image respectively , Crop the two pictures to the same square size , And save it as two new pictures A and B.
7. Put two color images A And B In different proportions (0.7:0.3, 0.5: 0.5, 0.3:0.7) Add and synthesize a new image , And display the original image in an image window 、 Three composite images .
8. Take the image from RGB Color space is converted to gray space , Use the weighting method 、 Mean value method and maximum value method 、 as well as matlab Self contained function rgb2gray.
The weighting method is GRAY=0.3R+0.59G+0.11*B
The mean method is GRAY= (R+G+B)/3
The maximum value is GRAY= max（R,G,B）
And display these four gray-scale images in an image window , And make a discussion .
9. According to the gray transformation curve , Carry out gray-scale mapping transformation on a gray-scale image , And display the original image and the transformed image in the same window .

Fourier transform and frequency domain processing

1． Generate a bright block image as shown in the figure f(x,y)（256×256 size 、 dark place =0, Bright place =255）, On the FFT：
（1） The original picture is displayed on the same screen f and FFT(f) Amplitude spectrum of （ Tips ： Use two-dimensional Fourier transform function fft2, To avoid excessive changes in Fourier transform data , Display the result after logarithmic transformation log(1.001+abs(FFT(f))) , use imshow or surf The function shows the spectrum ）;
（2） If order f1(x,y)=（-1）x+y f（x,y), Repeat the above process , Compare the similarities and differences of their amplitude spectra , Brief reasons ;
（3） If the f (x,y) Clockwise rotation 45 Degrees get f2(x,y), Trial display FFT(f2) Amplitude spectrum of , And with FFT(f) The amplitude spectrum of .（ Tips ： use imrotate Function to rotate the image ）;
10. Yes, one 256×256 size 、256 Gray level digital images are carried out in the frequency domain of the ideal low-pass 、 High pass filtering , The original picture is displayed on the same screen 、 Amplitude spectrum and low pass 、 Result diagram of high pass filtering .（ Tips ： Use two-dimensional Fourier transform function fft2, Two dimensional inverse Fourier transform function ifft2, Centralization function fftshift, Decentralized function ifftshift. Adopt different cut-off frequencies D0 Repeat the experiment , Observations ）

Resource guidance ： Digital image experiment ——matlab Learn and operate Fourier transform and frequency domain processing of images

experiment 2 Image spatial enhancement and image enhancement based on histogram

Spatial enhancement of image
Experimental content ：
1． Read in a picture 256 Grayscale digital image
2. Image smoothing and filtering
1） Gaussian noise is added to the original image respectively 、 Salt and pepper noise .
2） Using the neighborhood average method , We adopt 3×3,5×5, 7×7, 9×9 The template smoothes the noisy image , Show the original image 、 Noisy image and processed image .
3） Using median filtering , We adopt 33,55, 77, 99 The template denoises the noisy image , Show the original image 、 Noisy image and processed image .
4） Compare the processing results of various filtering methods and filtering templates
answer ： The neighborhood average method is to take the average value of a circle around each pixel , To reduce some sudden changes in pixel values （ Noise point ） Influence . One obvious result is ： While suppressing noise, the image becomes blurred , That is, the details of the image ( For example, edge information ) Weakened ; Median filtering is for every pixel , Take the median value of the surrounding pixel value , To replace the gray value of the pixel . It can suppress noise and keep details .
3. Image sharpening
1） utilize Laplacian Sharpening operator （α=-1） Yes 256 Sharpen the gray-scale digital image , Show before processing 、 Post image .
2） Laplacian sharpening is performed on the image with noise , Compare with the processing result of the image without noise .
3) Separate use Roberts、Prewitt and Sobel Edge detector , Edge detection of digital image , Show before processing 、 Post image .
Think about the problem ：
1． Mean filtering is adopted 、 median filtering , Which is more effective to suppress Gaussian noise or salt and pepper noise ？
2． Different template sizes , What is the difference in the treatment effect ？ Why? ？
3． Yes Laplacian In the processing result of sharpening operator , For less than 0 Part of , Adopt different methods to standardize to [0,255] when , What is the difference between the display effect of the image ？ Why? ？

Image enhancement based on histogram
Experimental content ：


1. Using functions imhist Histogram equalization of the original image , The images before and after processing and their histograms are displayed on the same screen , Compare the similarities and differences , It also discusses why the histogram of digital image is not completely evenly distributed after equalization .
2． Using functions histeq The histogram of the original image is specified , The images before and after processing and their histograms are displayed on the same screen , Compare the similarities and differences .
Additional questions , Write your own myhist Function to achieve histogram equalization .

Resource guidance ： digital image processing —— Image spatial enhancement and image enhancement based on histogram

Image segmentation experiment
Experimental content ：



1． Using threshold method to realize image segmentation , Try taking a variety of thresholds , Get the best results .
2． Use them separately robert,sobel And Laplace Gaussian operator for image edge detection , Compare the differences between the three operators .

Resource guidance ： digital image processing —— Image segmentation experiment

Morphological experiments
Experimental content ：



1. Read image 1 and 2, And binarization (im2bw).
2. Select the square respectively 、 Circular and diamond structures on the image 1 To corrode (imerode) Handle 、 inflation (imdilate) Handle 、 Open operation (bwmorph) And closed operations (bwmorph). Display the observation processing results .
3. For fingerprint images 3, Design processing program based on morphology , Get the fingerprint image with the best processing effect you can achieve .

Resource guidance ： digital image processing —— Morphological experiments