### How To Use Notch Filters – C# Guide

We can use notch filters for attenuating frequencies on custom locations across the frequency map. But for that we will need to utilize all of the knowledge we’ve acquired...

Filter by Category

- C# Tutorial(85)
- C# Image Processing(69)
- Morphological Processes(20)
- Image Processing(16)
- Image Restoration and Reconstruction(16)
- Image Segmentation(13)
- C# Data Structures And Algorithms(11)
- Color Image Processing(8)
- Frequency Domain Filtering(8)
- Image Noise(6)
- Grayscale Morphology(5)
- Thresholding(4)
- Order-Statistic Filters(4)
- Mean Filters(4)
- Sorting Algorithms(4)
- Morphological Reconstruction(3)
- Edge Detection(3)
- Simple Lists(2)
- RGB to HSI Color Model(2)
- Adaptive Filters(2)
- Tone and Color Corrections(2)
- Linked Lists(2)
- Stacks(1)
- Queues(1)
- Point Detection(1)
- Line Detection(1)
- C# Arrays(1)
- Region Segmentation Using Superpixels(1)
- Region Segmentation With K Means Clustering(1)
- Region Splitting And Merging(1)
- Sorted Lists(1)
- Region Growing Segmentation(1)
- Digital Image Watermarking(1)
- Using Color In Image Segmentation(1)
- Social Games(1)
- Bandreject Filters(1)
- Bandpass filters(1)
- Notch Filters(1)
- Landing Pages(1)
- Intensity Slicing and Color Coding(1)
- Color Slicing(1)
- Histogram Processing Color Images(1)
- Color Image Smoothing And Sharpening(1)
- C# Basics(1)

How To Use Notch Filters – C# Guide

We can use notch filters for attenuating frequencies on custom locations across the frequency map. But for that we will need to utilize all of the knowledge we’ve acquired from filtering in the frequency domain.

In case you’re just tuning in, I recommend checking out other posts about filtering in the frequency domain. But I’m also going to break the entire process down and give a sort of crash course here as well.

First of all, we need to transform image into frequency domain using Fourier transform. However, this transform takes two dimensional rectangle of image data. Therefore we will need to use it three times for one image, for each color channel separably.

If you’re already familiar with the transform, you will know that discrete Fourier transform takes a lot of time. Therefore we will use a more practical option, which is fast Fourier transform. However, it comes with a few contraints, before transforming our image we need to use zero padding to make it the right shape and size.

You can read more on a post that’s dedicated to Fourier transform alone. After that we’ve acquired frequency data from the transform we need to modify it with filters.

For filtering data in the frequency domain we need to multiply each frequency point with filter formula. After that, we take our filtered data and plug into inverse Fourier transform to turn it back into spatial domain.

Notch filters are probably the most useful of them all. They allow us to pass and attenuate frequencies on multiple locations over the frequency map. We can also use ideal, Butterworth and Gaussian filters to attenuate/pass frequencies smoothly.

Notch reject filters are similar to highpass filters, difference being the amount of circles of attenuated frequencies. Therefore, we have two different centers, which locations we need to calculate.

```
public static Complex[][][] IdealNotchRejectFilter(Complex[][][] frequencies, double d0, int[] coordinates)
{
int uoffset = coordinates[0];
int voffset = coordinates[1];
int size = FourierDemo.Size;
```` Complex[][][] filtered = frequencies;`

` for (int i = 0; i < 3; i++)`

` {`

` for (int j = 0; j < size; j++)`

` {`

` for (int k = 0; k < size; k++)`

` {`

` double d1 = Math.Sqrt(Math.Pow(j - size / 2 - uoffset, 2) + Math.Pow(k - size / 2 - voffset, 2));`

` double d2 = Math.Sqrt(Math.Pow(j - size / 2 + uoffset, 2) + Math.Pow(k - size / 2 + voffset, 2));`

` if ((d1 <= d0 || d2 <= d0) || !(d1 > d0 || d2 > d0))`

` {`

` filtered[i][j][k] *= 0f;`

` }`

` }`

` }`

` }`

` return filtered;`

` }`

Code above implements the ideal notch reject filter shown in the Fourier transform spectrum image above.

Every ideal filter has a simple formula, but it’s a different story for Butterworth filters. In addition Butterworth filters allow us to control the fade between attenuated and passed frequencies.

Notch filters also allow filtering with multiple pairs and following formula shows this with Butterworth notch reject filters.

```
public static Complex[][][] ButterworthNotchRejectFilter(Complex[][][] frequencies, double d0, int[] coordinates, int order)
{
int uoffset = coordinates[0];
int voffset = coordinates[1];
```` Complex[][][] filtered = frequencies;`

` for (int i = 0; i < 3; i++)`

` {`

` for (int j = 0; j < Size; j++)`

` {`

` for (int k = 0; k < Size; k++)`

` {`

` double d1 = Math.Sqrt(Math.Pow(j - Size / 2 - uoffset, 2) + Math.Pow(k - Size / 2 - voffset, 2));`

` double d2 = Math.Sqrt(Math.Pow(j - Size / 2 + uoffset, 2) + Math.Pow(k - Size / 2 + voffset, 2));`

` filtered[i][j][k] *= (float)(1f / (1 + Math.Pow(d0 / d1, 2 * order))) * (float)(1f / (1 + Math.Pow(d0 / d2, 2 * order)));`

` }`

` }`

` }`

` return filtered;`

` }`

Function above demonstrates the use of one Butterworth notch reject filter pair.

We can get formulation for all notch pass filters with notch reject filter formulas. Following formula show’s how we achieve this.

I recommend checking out project where you can also examine the use of Gaussian notch reject filter. Hopefully this tutorial has proven to be helpful and instructive.

## Comments