Ask for help about MATLAB filtering

You need to first choose the type of filter, the frequency range of the filter, and the minimum order of the filter.

However, it is also possible to convert the data directly to the frequency domain with FFT, and then make the amplitude of the corresponding frequency domain zero.

Let's not talk about the first one, there is more knowledge to use. Let's talk about the second method:

Let's say you have a time series: x

You can use the fast Fujifilm transform function fft to get the amplitude spectrum and phase spectrum of x.

y=fft(x);

y1=abs(y);%y1 is the amplitude spectrum.

The corresponding frequency can be calculated by sampling time interval and number of samples.

Let's say 0:df:f

And you want one of them.

Outside the frequency range [a,b], filter it out, as long as the corresponding y is zero.

Then use. x1=ifft(y).

Filtering is both convolution, which can be done using the fdatool tool.

1. The principle of median filtering: for a string of continuously input signals (quantized is a set of data). As shown in the figure below, this is the original signal of the input. The principle of median filtering is that the output value (y) of each x is recalculated and the new output value is new.

It is equivalent to y=new(x), and the new operation is to extract the value in the middle of the interval from the original signal centered on x and the length of 2k (the interval is [x-k+1, x+k]) as the result of y=new(x).

2. For example, input: y[1-10]: 1,2,3,4,5,6,7,8,9,10

take the interval 2k=4, so k=2; Median filtering is performed k=median filtering (y), which is >=1 by x-k+1, so when k=2, x>=2, filtering:

k[1]=y[1]

k[2] = (y[1], y[2], y[3], y[4]), i.e. 2 or 3.

3. The implementation of median filtering in MATLAB: call the function: a=medfilt1(b,n), b is the input signal, and a is the filtered signal, that is, the result.

4. For the input signal (the first figure), the following is the filtered image obtained by setting the intervals n=8 and n=16, respectively. Median filtering filters out spike pulses. The goal is that we are more interested in the filtered data.

The filtered data retains the trend of the original image while removing the impact of spikes on the analysis.

1. Open the software and read in**.

2. Establish a 3*3 Gaussian filter template and an average filter template respectively, and filter the noisy **. The original image, the ** after the noise is added, and the ** filtered with Gaussian and average templates respectively are displayed.

<>4. Use median filtering to process ** and display the processed image.

5. It can be seen from ** that the effect of noise can not be seen in the median filtered image. It's all there to save when you're done.

Median filtered medfilt2 with b = medfilt2(a, [m n]).

Your two points are wrong.

1. Input image a, which should be a two-dimensional matrix.

The input image B you use is obtained by adding noise to the A obtained by ImRead, and the image A read by ImRead is usually a 3-dimensional RGB diagram, which is a three-dimensional matrix, and it is not right to use Medfilt2 directly, you can use RGB2Gray(A) to convert A to a grayscale matrix first.

Second, the median filter should also specify the size of the filter template.

medfilt2 requires two input parameters, the first is image a, and the second parameter needs to input a vector of length two, [m n] specifies the size of the template, m rows n columns.

1. Linear smoothing filter: MATLAB is used to implement the domain average method to suppress noise.

2. Median filter: MATLAB is used to implement the median filtering program.

3. The state refers to the former Li Chengchang statistical filter: the state filter program is implemented with the ordfilt2 function.

4. Two-dimensional adaptive denoising filter: use the wiener2 function to realize the two-dimensional adaptive denoising filter.

5. Filtering of specific regions: Use the roifilt2 function provided in the MATLAB image processing toolbox to filter specific regions.

1. Purpose of the experiment:

1.Understand the meaning and means of image transformation; Familiarity with the basic properties of the Fourier transform; Proficient in FFT conversion methods and applications; Understand the distribution characteristics of the two-dimensional spectrum; Master the use of MATLAB programming to realize the Fourier transform of digital images; To evaluate the sensitivity of the human eye to the image amplitude and frequency characteristics and phase frequency characteristics.

2. Master how to use the Fourier transform for frequency domain filtering; Master the concepts and methods of frequency domain filtering; Proficient in various types of filters in the frequency domain; Frequency domain filtering was performed by using MATLAB program.

2. Experimental conditions and development environment:

1.PC computer, MATLAB software.

2.The ** required for the experiment can be used by the MATLAB software, and the material library can also be prepared by yourself.

3. Algorithm principle:

1.Fourier transform is applied for image processing.

The Fourier transform is a powerful tool for linear system analysis, which quantitatively analyzes the effects of digital systems, sampling points, electronic amplifiers, convolutional filters, noise, and display points.

2.Definition of the Fourier transform.

For a two-dimensional signal, the two-dimensional fourier transform is defined as:

Inverse Transform: The two-dimensional discrete Fourier transform is:

Inverse transformation: 3Frequency domain filtering is divided into two categories: low-pass filter and high-pass filter, and the corresponding filters are low-pass filter and high-pass filter, respectively. The basic idea of low-pass filtering in the frequency domain:

f(u,v) is the Fourier transform form of the image to be passivated, h(u,v) is a low-pass filter transformation function selected, g(u,v) is the result obtained by reducing the high-frequency part of f(u,v) by h(u,v), and the passivated image is obtained by using the inverse Fourier transform. The Ideal Low-Pass Filter (ILPF) has a transfer function:

where d0 is the specified non-negative number and d(u,v) is the distance (u,v) from the center of the filter. The trajectory of the point d(u,v)=d0 is a circle. The transfer function of the nth-order Butterworth low-pass filter (BLPF), which occurs at a cut-off frequency at a distance from the origin d0, is: