From d7384c05de386158767c6dd56ec1053cac1ad420 Mon Sep 17 00:00:00 2001
From: Carla Elena Gomez Alvarado <10-11266@usb.ve>
Date: Tue, 18 May 2021 23:49:05 +0000
Subject: [PATCH] BPF
---
...samiento_digital_imagenes_bordes_BPF.ipynb | 430 ++++++++++++++++++
1 file changed, 430 insertions(+)
create mode 100644 Procesamiento_digital_imagenes_bordes_BPF.ipynb
diff --git a/Procesamiento_digital_imagenes_bordes_BPF.ipynb b/Procesamiento_digital_imagenes_bordes_BPF.ipynb
new file mode 100644
index 0000000..6c523d4
--- /dev/null
+++ b/Procesamiento_digital_imagenes_bordes_BPF.ipynb
@@ -0,0 +1,430 @@
+{
+ "cells": [
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import cv2\n",
+ "import numpy as np\n",
+ "from matplotlib import pyplot as plt\n",
+ "from PIL import Image\n",
+ "from PIL.ExifTags import TAGS\n",
+ "from numpy import asarray\n",
+ "from numpy.fft import fft2, fftshift, fftfreq, ifft2 #Python DFT\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#imagename = \"Data/crateres.jpg\"\n",
+ "#imagename = \"Data/Perseverance_ZR0_0074.png\" -codigo adicional para recorte de bordes negros.\n",
+ "#imagename = \"Data/mastcamz_2_prints.jpg\"\n",
+ "imagename = \"Data/Data_Mars_Mars8.png\"\n",
+ " \n",
+ "# read the image data using PIL\n",
+ "image = Image.open(imagename)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "display (image)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "crateres.jpg : Image downloaded from : https://mars.nasa.gov/mars2020-raw-images/pub/ods/surface/sol/00000/ids/edr/browse/edl/ELM_0000_0666952859_600ECM_N0000030LVS_04000_0000LUJ00_800.jpg\n",
+ "\n",
+ "Perseverance_ZR=_0074.png : Image downloaded from : "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "img = cv2.imread(imagename, 0) # load an image\n",
+ "\n",
+ "#3 = np.asarray(imagename, dtype = np.float32) # Image class instance, I1, to float32 Numpy array, a\n",
+ "\n",
+ "img_float32 = np.float32(img)\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Para procesar la imagen se lee como cv2.imread y luego se le aplica la DFT \n",
+ "\n",
+ "DFT: Discrete Fourier Transform. \n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "hist =cv2.calcHist([img],[0],None,[256],[0,256])\n",
+ "plt.plot(hist,color='blue')\n",
+ "plt.xlabel('Intensidad de Iluminacion')\n",
+ "plt.ylabel('Pixels Bins')\n",
+ "plt.show()\n",
+ "\n",
+ "plt.hist(np.concatenate((img),axis=0),bins=250)\n",
+ "plt.xlabel('Intensidad de Iluminacion')\n",
+ "plt.ylabel('Pixels Bins')\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#transform,ar el array de la imagen en punto flotante para ser procesado. \n",
+ "dft = cv2.dft(np.float32(img), flags=cv2.DFT_COMPLEX_OUTPUT)\n",
+ "dft_shift =np.fft.fftshift(dft)\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "# convert image to numpy array \n",
+ "data = asarray(image)\n",
+ "print(type(data))\n",
+ "# caracteristicas de la imagen\n",
+ "print(data.shape)\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#Muestra cada valor de cada pixel de la imagen como array en Numpy\n",
+ "#print(data)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#imagen en el espacio de las frecuencias : \n",
+ "\n",
+ "f = np.fft.fft2(img) #fft2 para 2 dimensiones\n",
+ "fshift = np.fft.fftshift(f) # para centrar las frecuencias bajas.\n",
+ "# fft2 tiene valors imaginarios por ello se saca el espectro real o de magnitud. \n",
+ "magnitude_sprectrum = 20*np.log(np.abs(fshift)) "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "f.shape"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "fft2(f)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#Matplotlib para crear 2 subplots en la misma linea para comparar la imagen original y \n",
+ "#la imagen transformada en el espectro de frecuencias. \n",
+ "plt.subplot(121),plt.imshow(img , cmap = 'gray')\n",
+ "plt.title('Original Image'), plt.xticks([]),plt.yticks([])\n",
+ "plt.subplot(122),plt.imshow(magnitude_sprectrum , cmap = 'gray')\n",
+ "plt.title('Magnitude Spectrum'), plt.xticks([]),plt.yticks([])\n",
+ "plt.show\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#Filtro pasa alto HPF en el dominio de las frencuencias para demarcar contornos: \n",
+ "\n",
+ "\n",
+ "rows, cols = img.shape\n",
+ "crow, ccol = int(rows/2), int(cols/2)\n",
+ "\n",
+ "#remover las frecuencias bajas usando una plantilla circular \n",
+ "\n",
+ "mask=np.ones((rows,cols,2),np.uint8)\n",
+ "r = 40\n",
+ "center = [crow, ccol]\n",
+ "x, y = np.ogrid[:rows, :cols]\n",
+ "mask_area = (x - center[0]) ** 2 + (y - center[1]) ** 2 <= r*r\n",
+ "mask[mask_area] = 1\n",
+ "\n",
+ "fshift=dft_shift*mask\n",
+ "\n",
+ "fshift_mask_mag=2000*np.log(cv2.magnitude(fshift[:,:,0],fshift[:,:,1]))\n",
+ "\n",
+ "f_ishift =np.fft.ifftshift(fshift)\n",
+ "img_back =cv2.idft(f_ishift)\n",
+ "img_back =cv2.magnitude(img_back[:,:,0],img_back[:,:,1])\n",
+ "\n",
+ "#fshift[crow-30:crow+30,ccol-30:ccol+30]=0\n",
+ "\n",
+ "\n",
+ "f_ishift = np.fft.ifftshift(fshift)\n",
+ "\n",
+ "img_back = np.fft.ifft2(f_ishift)\n",
+ "\n",
+ "img_back = np.abs(img_back)\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#Filtro pasa Banda BPF aplicando dos circulos concentricos\n",
+ "\n",
+ "rows, cols = img.shape\n",
+ "crow, ccol = int(rows/2), int(cols/2) # circulo centro\n",
+ "\n",
+ "\n",
+ "#remover las frecuencias bajas usando una plantilla circular \n",
+ "\n",
+ "rows, cols = img.shape\n",
+ "crow, ccol = int(rows/2), int(cols/2)\n",
+ "\n",
+ "mask=np.zeros((rows,cols,2),np.uint8)\n",
+ "r_out = 40\n",
+ "r_in = 4\n",
+ "center = [crow, ccol]\n",
+ "x, y = np.ogrid[:rows, :cols]\n",
+ "\n",
+ "mask_area = np.logical_and(((x-center[0])**2 +(y-center[1])**2>= r_in**2),\n",
+ " ((x-center[0])**2 +(y-center[1])**2<= r_out**2))\n",
+ " \n",
+ " \n",
+ "mask[mask_area] = 1\n",
+ "\n",
+ "fshift=dft_shift*mask\n",
+ "\n",
+ "fshift_mask_mag= 2000 * np.log(cv2.magnitude(fshift[:, :, 0], fshift[:, :, 1]))\n",
+ "\n",
+ "f_ishift =np.fft.ifftshift(fshift)\n",
+ "img_back =cv2.idft(f_ishift)\n",
+ "img_back =cv2.magnitude(img_back[:,:,0],img_back[:,:,1])\n",
+ "\n",
+ "#fshift[crow-30:crow+30,ccol-30:ccol+30]=0\n",
+ "\n",
+ "fig = plt.figure(figsize=(20,20))\n",
+ "\n",
+ "fig.add_subplot(2, 2, 1), plt.imshow(img, cmap='gray')\n",
+ "plt.title('Input Image'), plt.xticks([]), plt.yticks([])\n",
+ "fig.add_subplot(2, 2, 2), plt.imshow(magnitude_sprectrum,cmap='gray')\n",
+ "plt.title('After FFT'), plt.xticks([]), plt.yticks([])\n",
+ "fig.add_subplot(2, 2, 3), plt.imshow(fshift_mask_mag, cmap='gray')\n",
+ "plt.title('FFT + Mask'), plt.xticks([]), plt.yticks([])\n",
+ "fig.add_subplot(2, 2, 4), plt.imshow(img_back, cmap='gray')\n",
+ "plt.title('After FFT Inverse'), plt.xticks([]), plt.yticks([])\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#Filtro pasa Banda BPF aplicando dos circulos concentricos\n",
+ "\n",
+ "rows, cols = img.shape\n",
+ "crow, ccol = int(rows/2), int(cols/2) # circulo centro\n",
+ "\n",
+ "\n",
+ "#remover las frecuencias bajas usando una plantilla circular \n",
+ "\n",
+ "rows, cols = img.shape\n",
+ "crow, ccol = int(rows/2), int(cols/2)\n",
+ "\n",
+ "mask=np.zeros((rows,cols,2),np.uint8)\n",
+ "\n",
+ "#con diferentes radios tal que sean mayor que la zona central, tapando la mayor cantidad de bajas frencuencias. \n",
+ "r_out = 80\n",
+ "r_in = 4\n",
+ "center = [crow, ccol]\n",
+ "x, y = np.ogrid[:rows, :cols]\n",
+ "\n",
+ "mask_area = np.logical_and(((x-center[0])**2 +(y-center[1])**2>= r_in**2),\n",
+ " ((x-center[0])**2 +(y-center[1])**2<= r_out**2))\n",
+ " \n",
+ " \n",
+ "mask[mask_area] = 1\n",
+ "\n",
+ "fshift=dft_shift*mask\n",
+ "\n",
+ "fshift_mask_mag= 2000 * np.log(cv2.magnitude(fshift[:, :, 0], fshift[:, :, 1]))\n",
+ "\n",
+ "f_ishift =np.fft.ifftshift(fshift)\n",
+ "img_back =cv2.idft(f_ishift)\n",
+ "img_back =cv2.magnitude(img_back[:,:,0],img_back[:,:,1])\n",
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ "fig = plt.figure(figsize=(20,20))\n",
+ "\n",
+ "fig.add_subplot(121),plt.imshow(img , cmap = 'gray')\n",
+ "plt.title('Original Image'), plt.xticks([]),plt.yticks([])\n",
+ "fig.add_subplot(122),plt.imshow(img_back , cmap = 'gray')\n",
+ "plt.title('AFTER FFT Inverse image'), plt.xticks([]),plt.yticks([])\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "plt.subplot(131),plt.imshow(img, cmap='gray')\n",
+ "plt.title('Input Image'), plt.xticks([]), plt.yticks([])\n",
+ "plt.subplot(132),plt.imshow(img_back , cmap = 'gray')\n",
+ "plt.title('Imagen con HPF'), plt.xticks([]),plt.yticks([])\n",
+ "plt.subplot(133),plt.imshow(img_back)\n",
+ "plt.title('Imagen Final'), plt.xticks([]),plt.yticks([])\n",
+ "plt.show"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "fig = plt.figure(figsize=(20,20))\n",
+ "\n",
+ "fig.add_subplot(121),plt.imshow(img , cmap = 'gray')\n",
+ "plt.title('Original Image'), plt.xticks([]),plt.yticks([])\n",
+ "fig.add_subplot(122),plt.imshow(img_back , cmap = 'gray')\n",
+ "plt.title('HPF image'), plt.xticks([]),plt.yticks([])\n",
+ "plt.show\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#Comparativa para la foto de las rocas \n",
+ "#recortado de foto \n",
+ "#Recortar una imagen\n",
+ "crop= img[200:400,200:500]\n",
+ "\n",
+ "\n",
+ "cv2.imshow('Seccion Imagen',crop)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#filtro Pasa bajos: LPF\n",
+ "\n",
+ "dft_shift= np.fft.fftshift(dft)\n",
+ "\n",
+ "rows, cols = img.shape\n",
+ "\n",
+ "crow , ccol = int(rows/2), int(cols/2) \n",
+ "\n",
+ "mask= np.zeros((rows,cols,2),np.uint8)\n",
+ "mask[crow-30:crow+30,ccol-30:ccol+30]=1\n",
+ "\n",
+ "\n",
+ "fshift = dft_shift*mask\n",
+ "f_ishift=np.fft.ifftshift(fshift)\n",
+ "img_back=cv2.idft(f_ishift)\n",
+ "img_back=cv2.magnitude(img_back[:,:,0],img_back[:,:,1])"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "fig = plt.figure(figsize=(20,20))\n",
+ "\n",
+ "fig.add_subplot(121),plt.imshow(img , cmap = 'gray')\n",
+ "plt.title('Original Image'), plt.xticks([]),plt.yticks([])\n",
+ "fig.add_subplot(122),plt.imshow(img_back , cmap = 'gray')\n",
+ "plt.title('LPF image'), plt.xticks([]),plt.yticks([])\n",
+ "plt.show\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": []
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "Python 3",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.7.3"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 4
+}
--
GitLab