Lab 5 Q1 (b24d4532) · Commits · 袁通 / EE326 Digital Image Processing LAB

Lab5_Frequency_Domain_Filtering/EE326_SUSTech.py

+2 −1

Original line number	Diff line number	Diff line
		@@ -8,8 +8,9 @@ import math
		from scipy import interpolate
		import matplotlib.pyplot as plt


		# General
		def format(input_image):
		def format_image(input_image):
		output_image = np.clip(input_image, 0, 255)
		output_image = output_image.astype(np.uint8)

+4 −4

Original line number	Diff line number	Diff line
		@@ -16,10 +16,10 @@ def sobel_filter_11810818(input_image):
		output_spatial = EE326_SUSTech.convolution_3x3(input_image, kernel1)
		output_spatial += EE326_SUSTech.convolution_3x3(input_image, kernel2)

		io.imsave("Q5_1_spatial_mask.tif", EE326_SUSTech.format(output_spatial))
		io.imsave("Q5_1_spatial_mask.tif", EE326_SUSTech.format_image(output_spatial))

		output_spatial += input_image
		io.imsave("Q5_1_spatial_filtered.tif", EE326_SUSTech.format(output_spatial))
		io.imsave("Q5_1_spatial_filtered.tif", EE326_SUSTech.format_image(output_spatial))

		# Filtering in the Frequency Domain
		input_image_padded = EE326_SUSTech.zero_padding_DFT(input_image)
		@@ -32,9 +32,9 @@ def sobel_filter_11810818(input_image):
		filtered_1 = np.real(np.fft.ifft2(np.fft.fft2(input_image_padded) * np.fft.fft2(DFT_kernel_1)))[m1-600:m1, n1-600:n1]
		filtered_2 = np.real(np.fft.ifft2(np.fft.fft2(input_image_padded) * np.fft.fft2(DFT_kernel_2)))[m1-600:m1, n1-600:n1]

		io.imsave("Q5_1_frequency_mask.tif", EE326_SUSTech.format(filtered_1 + filtered_2))
		io.imsave("Q5_1_frequency_mask.tif", EE326_SUSTech.format_image(filtered_1 + filtered_2))

		io.imsave("Q5_1_frequency_filtered.tif", EE326_SUSTech.format(filtered_1 + filtered_2 + input_image))
		io.imsave("Q5_1_frequency_filtered.tif", EE326_SUSTech.format_image(filtered_1 + filtered_2 + input_image))


		if __name__ == '__main__':