FFT Filters
Low Pass Frequency Domain Spectral Filtering
Note: Can be used in conjunction with the Low Pass FFT function from SPECTRA MATHEMATICS panel.
Features:
Use this filter to remove high-frequency components from the spectral coordinate in the entire dataset. No filtering is applied to the spatial coordinates.
The Fast Fourier Transform (FFT) is used to transform the input spectra. In order to perform the FFT filters efficiently, IDCube enables users to first optimize and visualize the performance of these filters on a spectrum from the selected regions of interest (available from SPECTRAL ANALYSIS panel).
The FFT filtering of the image is performed in three steps automatically. First, the filter executes pixel-by-pixel Fourier transform of the spectrum, converting the original spectra into their corresponding frequency domain spectra. In the second step, a low pass filter with a specified cut-off frequency is applied. In the third step, the inverted Fourier transform algorithm is applied to generate a new dataset where only low frequencies are kept.
Steps:
1. Load the file.
2. Select Filtering and Enhancement → FFT Low Pass Filter.
A pop-up dialog will ask you to enter a Low-Frequency cutoff value. A higher value indicates that high frequencies will be removed prior to the reconstruction resulting in the low-frequency image where 50% of the high frequencies are cut. Notice that the scale is between 0 and 1.
(To Visualize Spectra after FFT analysis, go to the Spectral Analysis chapter of this manual).
The resulting image with 50% cutoff high frequencies will be shown in the IMAGE DISPLAY panel. Similar to other filters, the processed dataset is temporarily stored and available for other processing algorithms.
References:
The algorithm is built using part of the fftl library developed by Shmuel Ben-Ezra in 2009:
High Pass Frequency Domain Spectral Filtering
Note: Can be used in conjunction with the High Pass FFT function from SPECTRA MATHEMATICS panel.
Features:
Use this filter to remove low-frequency components from the spectral coordinate in the entire dataset. No filtering is applied to the spatial coordinates.
The Fast Fourier Transform (FFT) is used to transform the input spectra. In order to perform the FFT filters efficiently, IDCube enables users to first optimize and visualize the performance of these filters on a spectrum from the selected regions of interest.
The FFT filtering of the image is performed in three steps automatically. First, the filter executes pixel-by-pixel Fourier transform of the spectrum, converting the original spectra into their corresponding frequency domain spectra. In the second step, a high pass filter with a specified cut-off frequency is applied. In the third step, the inverted Fourier transform algorithm is applied to generate a new dataset where only high frequencies are kept.
Steps:
1. Load the file.
2. Select Filtering and Enhancement → FFT High Pass Filter.
3. Input a Cutoff value number in the pop-up dialog. The number reflects the % of Nyquist frequency in the range between 0 and 1. A higher value indicates that more low frequencies will be removed prior to the reconstruction. The value of 0.5 means that 50% of the low frequencies are cut.
(To Visualize Spectra after FFT analysis, go to the Spectral Analysis chapter of this manual).
Like most other filters, the filtered dataset is temporarily stored and available for other processing algorithms.
References:
The algorithm is built using part of the fftl library developed by Shmuel Ben-Ezra in 2009:
