Many things added

imageEngine/re_serie_7.py

@@ -1,5 +1,5 @@
 import umage as um
-from math import sqrt, atan2
+from math import sqrt, atan2, sin, cos, pi
 
 def greyscale(mat_img):
     gray_img = []
@@ -100,12 +100,20 @@ def appliquer_convolution(img, mat, i, j):
 ######################################################################
 ########################Exercices personnelles########################
 ######################################################################
-def convolution_gauss(mat_img, mat):
+def convolution_gauss(mat_img):
+    mat_gauss = [
+        [2/159, 4/159, 5/159, 4/159,2/159],
+        [4/159, 9/159,12/159, 9/159,4/159],
+        [5/159,12/159,15/159,12/159,5/159],
+        [4/159, 9/159,12/159, 9/159,4/159],
+        [2/159, 4/159, 5/159, 4/159,2/159]
+    ]
+
     return_img = []
     for j in range(len(mat_img)):
         ligne = []
         for i in range(len(mat_img[0])):
-            val = reduction_bruit(mat_img, mat, i, j)
+            val = reduction_bruit(mat_img, mat_gauss, i, j)
             ligne.append((val,)*3)
         return_img.append(ligne)
     return return_img
@@ -118,8 +126,8 @@ def reduction_bruit(img, mat, i, j):
             pixel_j = j - (len(mat) // 2) + y
             pix = pixel(img, pixel_i, pixel_j)
             somme += pix[0]*mat[y][x]
-    normalise = int(round(somme / (1/159)))
+    normalise = round(somme)
-    return min(max(normalise,0), 255)
+    return normalise
 
 def filtre_canny(img):
 
@@ -128,7 +136,7 @@ def filtre_canny(img):
         color_y = pixel2[0]
         
         norm = round(sqrt(color_x**2 + color_y**2))
-        norm = int(min(norm, 255))
+        norm = min(norm, 255)
 
         grad = atan2(color_y, color_x)
         return norm, grad
@@ -146,29 +154,140 @@ def filtre_canny(img):
     if not is_greyscale(img):
         img = greyscale(img)
     
-    mat_gauss = [
-        [2, 4, 5, 4,2],
-        [4, 9,12, 9,4],
-        [5,12,15,12,5],
-        [4, 9,12, 9,4],
-        [2, 4, 5, 4,2]
-    ]
     mat_x = [[-1,0,1]]
     mat_y = [[1],[0],[-1]]
 
-    #lissage
+    #lissage/suppression des bri
-    img = convolution_gauss(img, mat_gauss)
+    img_no_bruit = convolution_gauss(img)
     Jx = convolution(img, mat_x)
     Jy = convolution(img, mat_y)
     normGrad = liste_normGrad(Jx, Jy)
 
-image = um.load("my_images\\Zero_Two_1.jpeg")
+    #Suppresion des non-maximum
-mat_gauss = [
+
-        [2, 4, 5, 4,2],
+
-        [4, 9,12, 9,4],
+#temp
-        [5,12,15,12,5],
+def norme_gradient(pixel1, pixel2):
-        [4, 9,12, 9,4],
+        color_x = pixel1[0]
-        [2, 4, 5, 4,2]
+        color_y = pixel2[0]
-    ]
+        
-image = convolution_gauss(image, mat_gauss)
+        norm = round(sqrt(color_x**2 + color_y**2))
-um.save(image, "test\\zero_two", "png")
+        norm = min(norm, 255)
+
+        grad = atan2(color_y, color_x)
+        return norm, grad
+
+#temp
+def liste_normGrad(im1, im2):
+    liste = []
+    for j in range(len(im1)):
+        ligne = []
+        for i in range(len(im1[0])):
+            normGrad = norme_gradient(im1[j][i], im2[j][i])
+            ligne.append(normGrad)
+        liste.append(ligne)
+    return liste
+
+mat_x = [[-1,0,1]]
+mat_y = [[1],[0],[-1]]
+#temp
+#lissage
+img = um.load("imageEngine\\images\\valve.png")
+img = convolution_gauss(img)
+Jx = convolution(img, mat_x)
+Jy = convolution(img, mat_y)
+normGrad = liste_normGrad(Jx, Jy)
+###########
+
+
+
+def find_neighbord_norm(mat, i, j, rad):
+    x = 0
+    y = 0
+    if sin(pi/8) <= abs(sin(rad)):
+        y = 1
+    if cos(3*pi/8)>abs(cos(rad)):
+        x = 1
+
+    norm_pix1 = -1
+    norm_pix2 = -1
+    if 0 <= j-y < len(mat):
+        if 0 <= i-x < len(mat[0]):
+            norm_pix1 = mat[j-y][i-x][0]
+    if 0 <= j+y < len(mat):
+        if 0 <= i+x < len(mat[0]):
+            norm_pix2 = mat[j+y][i+x][0]
+    
+    return norm_pix1, norm_pix2
+
+def delete_pixel(mat_img, mat):
+    img_to_return = []
+    for j in range(len(mat)):
+        ligne = []
+        for i in range(len(mat[0])):
+            rad = mat[j][i][1]
+            norms = find_neighbord_norm(mat, i, j, rad)
+            if rad < norms[0] or rad < norms[1]:
+                ligne.append((0,)*3)
+            else:
+                ligne.append(mat_img[j][i])
+        img_to_return.append(ligne)
+    return img_to_return
+
+
+
+def hysteresis(mat_img, mat_norm, Th):
+    Tl = Th / 2
+    mat_img = yesOrNo(mat_img, Th, Tl)
+    result_image = []
+    for j in range(len(mat_img)):
+        ligne = []
+        for i in range(len(mat_img[0])):
+            rad = mat_norm[j][i][1]
+            color1, color2 = find_neighbord_pixel(mat_img, i, j, rad+(pi/2))
+            if color1 == 255 or color2 == 255:
+                ligne.append((255,)*3)
+            else:
+                ligne.append((0,)*3)
+        result_image.append(ligne)
+    return result_image
+
+def find_neighbord_pixel(mat_image, i, j, rad):
+    x = 0
+    y = 0
+    if sin(pi/8) <= abs(sin(rad)):
+        y = 1
+    if cos(3*pi/8)>abs(cos(rad)):
+        x = 1
+
+    color_pix1 = 0
+    color_pix2 = 0
+    if 0 <= j-y < len(mat_image):
+        if 0 <= i-x < len(mat_image[0]):
+            color_pix1 = mat_image[j-y][i-x][0]
+    if 0 <= j+y < len(mat_image):
+        if 0 <= i+x < len(mat_image[0]):
+            color_pix2 = mat_image[j+y][i+x][0]
+    
+    return color_pix1, color_pix2
+
+def yesOrNo(mat_img, Th, Tl):
+    result_image = []
+    for j in range(len(mat_img)):
+        ligne = []
+        for i in range(len(mat_img[0])):
+            pix = mat_img[j][i]
+            if Th <= pix[0]:
+                ligne.append((255,)*3)
+            elif pix[0] < Tl:
+                ligne.append((0,)*3)
+            else:
+                ligne.append(pix)
+        result_image.append(ligne)
+    return result_image
+
+
+zt_no_maxima = delete_pixel(img, normGrad)
+zt_hysteresis = hysteresis(zt_no_maxima, normGrad, 200)
+
+um.save(zt_hysteresis, "imageEngine\\test\\valve", "png")