編集履歴

質問編集履歴

コードの追加

2020/12/10 06:36

投稿

shogaku

スコア2

test CHANGED Viewed

File without changes

test CHANGED Viewed

@@ -40,7 +40,31 @@
     範囲リストの要素数をカウントして表示
+・・・・・・・・・・・・・・・・・・・・・・・・・・・・
+la1: 前時刻の粒子数
+l0: 前時刻の粒子数
+la2: 後時刻の粒子数
+kinbou[x y]
+kinbou[x y]　　　　　　　　　　　　　　　　　　　　　　←これを画像枚数-1回出力
+kinbou[x y]
+　　・　　　　　←各時刻で対応する粒子数と同じ回数
+　　・
+　　・
+　　・
+AA:AA、AB:ABリストの要素数
+・・・・・・・・・・・・・・・・・・・・・・・・・・・・
 ```
@@ -60,11 +84,11 @@
 la2: 1260
-[801 136]
+kinbou[801 136]
-[830 296]
+kinbou[830 296]
-[963 193]
+kinbou[963 193]
 AA:1 AB:0
@@ -74,13 +98,13 @@
 la2: 1276
-[800 143]
+kinbou[800 143]
-[831 296]
+kinbou[831 296]
-[961 185]
+kinbou[961 185]
-[ 823 -181]
+kinbou[ 823 -181]
 AA:1 AB:0
@@ -210,7 +234,7 @@
     # 近傍点を特定
-    for e in range(len(l0)):
+    for e in range(la1):
@@ -266,9 +290,7 @@
         knn_model = NearestNeighbors(n_neighbors=k, algorithm='brute').fit(l1e)
-        distances, indices = knn_model.kneighbors([l0[e]])
+        distances, indices = knn_model.kneighbors([l0[e]])
-        #print("K Nearest Neighbors:")
@@ -286,7 +308,7 @@
         l1 -= kinbou
-        print(kinbou)
+        print("kinbou" + str(kinbou))
@@ -332,8 +354,6 @@
         break
 ```

コードの追加

2020/12/10 06:36

投稿

shogaku

スコア2

test CHANGED Viewed

File without changes

test CHANGED Viewed

@@ -210,7 +210,7 @@
     # 近傍点を特定
-    for e in range(la1):
+    for e in range(len(l0)):

コードの追加

2020/12/10 06:09

投稿

shogaku

スコア2

test CHANGED Viewed

File without changes

test CHANGED Viewed

@@ -60,19 +60,11 @@
 la2: 1260
-l1e: 1153
-l0[e]: [800, 136]
+[801 136]
-l1e: 557
-l0[e]: [803, 140]
+[830 296]
-l1e: 83
+[963 193]
-l0[e]: [806, 137]
-l1e: 0
 AA:1 AB:0
@@ -82,48 +74,16 @@
 la2: 1276
-l1e: 1266
-l0[e]: [435, 27]
-l1e: 600
-l0[e]: [801, 136]
-l1e: 90
-l0[e]: [803, 139]
-l1e: 8
-l0[e]: [807, 143]
+[800 143]
-l1e: 0
+[831 296]
+[961 185]
+[ 823 -181]
 AA:1 AB:0
-la1: 1276
-l0: 1276
-la2: 1255
-l1e: 1129
-l0[e]: [803, 139]
-l1e: 353
-l0[e]: [800, 143]
-l1e: 30
-l0[e]: [1212, 203]
-l1e: 0
-AA:2 AB:0
 ```
@@ -160,216 +120,218 @@
 n = 0
-for n in range(3):
-  img1 = cv2.imread('mask_%04d.jpg' % n,cv2.IMREAD_GRAYSCALE)
-  img2 = cv2.imread('mask_%04d.jpg' % (n+1),cv2.IMREAD_GRAYSCALE)
-  # ラベリング処理
-  label1 = cv2.connectedComponentsWithStats(img1)
-  label2 = cv2.connectedComponentsWithStats(img2)
-  # オブジェクト情報を項目別に抽出
-  la1 = label1[0] - 1
-  data1 = np.delete(label1[2], 0, 0)
-  center1 = np.delete(label1[3], 0, 0)
-  la2 = label2[0] - 1
-  data2 = np.delete(label2[2], 0, 0)
-  center2 = np.delete(label2[3], 0, 0)
-  print("la1:",la1)
-  l0 = []
-  # 前時刻のオブジェクト情報を取得
-  for i in range(la1):
-    l0 += [[int(center1[i][0]),int(center1[i][1])]]
-    i += 1
-    if i == la1:
-      break
-  print("l0:",len(l0))
-  l1 = []
-  # 後時刻のオブジェクト情報を取得
-  for j in range(la2):
-    l1 += [[int(center2[j][0]),int(center2[j][1])]]
-    j += 1
-    if j == la2:
-      break
-  e = 0
+for n in range(2):
+    img1 = cv2.imread('mask_%04d.jpg' % n,cv2.IMREAD_GRAYSCALE)
+    img2 = cv2.imread('mask_%04d.jpg' % (n+1),cv2.IMREAD_GRAYSCALE)
+    # ラベリング処理
+    label1 = cv2.connectedComponentsWithStats(img1)
+    label2 = cv2.connectedComponentsWithStats(img2)
+    # オブジェクト情報を項目別に抽出
+    la1 = label1[0] - 1
+    data1 = np.delete(label1[2], 0, 0)
+    center1 = np.delete(label1[3], 0, 0)
+    la2 = label2[0] - 1
+    data2 = np.delete(label2[2], 0, 0)
+    center2 = np.delete(label2[3], 0, 0)
+    print("la1:",la1)
+    l0 = []
+    # 前時刻のオブジェクト情報を取得
+    for i in range(la1):
+        l0 += [[int(center1[i][0]),int(center1[i][1])]]
+        i += 1
+        if i == la1:
+          break
+    print("l0:",len(l0))
+    l1 = []
+    # 後時刻のオブジェクト情報を取得
+    for j in range(la2):
+        l1 += [[int(center2[j][0]),int(center2[j][1])]]
+        j += 1
+        if j == la2:
+          break
+    e = 0
-  print("la2:",la2)
-  # 近傍点を特定
-  for e in range(la1):
-    l1e = []
-    # 探査範囲内のオブジェクト情報を取得
-    for l in range(la2):
-      if int(l1[l][0]) >= (int(l0[e][0]) - 5):
-        l1e += [[int(l1[l][0]),int(l1[l][1])]]
-        l += 1
-      else:
-        l += 1
-      if l == la2:
+    print("la2:",la2)
+    # 近傍点を特定
+    for e in range(la1):
+        l1e = []
+      # 探査範囲内のオブジェクト情報を取得
+        for l in range(la2):
+            if int(l1[l][0]) >= (int(l0[e][0]) - 5):
+                l1e += [[int(l1[l][0]),int(l1[l][1])]]
+                l += 1
+            else:
+                l += 1
+            if l == la2:
+                break
+        if len(l1e) == 0:
+             break
+        l1e = np.array(l1e)
+        #print("l1e:",len(l1e))
+        #モデル構築
+        knn_model = NearestNeighbors(n_neighbors=k, algorithm='brute').fit(l1e)
+        distances, indices = knn_model.kneighbors([l0[e]])
+        #print("K Nearest Neighbors:")
+         #K個までのスライス、現在Kは1個
+        for rank, index in enumerate(indices[0][:k], start=1):
+            str(rank) + " ==>",l1e[index]
+            str(rank) + " ==>", distances[0][0]
+        kinbou = l1e[index]
+        l1 -= kinbou
+        print(kinbou)
+        if distances[0][0] < 150 and 1920 >= kinbou[0] >= 0 and 2160 >= kinbou[1] >= 0 :
+            AA += [kinbou]
+            l1e = l1e.tolist()
+            del l1e[:]
+            e += 1
+        elif distances[0][0] < 150 and 3840 >= kinbou[0] > 1920 and 2160 >= kinbou[1] >= 0 :
+            AB += [kinbou]
+            l1e = l1e.tolist()
+            del l1e[:]
+            e += 1
+    print("AA" + ":" +  str(len(AA)) + " " + "AB" + ":" +  str(len(AB)))
+    if img2 is None:
+        print("can't read")
         break
-    l1e = np.array(l1e)
-    print("l1e:",len(l1e))
-    if len(l1e) == 0:
-      break
-    print("l0[e]:",l0[e])
-    #モデル構築
-    knn_model = NearestNeighbors(n_neighbors=k, algorithm='ball_tree').fit(l1e)
-    distances, indices = knn_model.kneighbors([l0[e]])
-    #print("K Nearest Neighbors:")
-    #K個までのスライス、現在Kは1個
-    for rank, index in enumerate(indices[0][:k], start=1):
-      str(rank) + " ==>",l1e[index]
-      str(rank) + " ==>", distances[0][0]
-    kinbou = l1e[index]
-    l1 -= kinbou
-    if distances[0][0] < 150 and 1920 >= kinbou[0] >= 0 and 2160 >= kinbou[1] >= 0 :
-      AA += [kinbou]
-      #l1e = l1e.tolist()
-      #del l1e[:]
-      e += 1
-    elif distances[0][0] < 150 and 3840 >= kinbou[0] > 1920 and 2160 >= kinbou[1] >= 0 :
-      AB += [kinbou]
-      #l1e = l1e.tolist()
-      #del l1e[:]
-      e += 1
-  print("AA" + ":" +  str(len(AA)) + " " + "AB" + ":" +  str(len(AB)))
-  if img2 is None:
-    print("can't read")
-    break
 ```

コードの追加

2020/12/10 05:48

投稿

shogaku

スコア2

test CHANGED Viewed

File without changes

test CHANGED Viewed

@@ -58,13 +58,21 @@
 l0: 1263
-1153
-[801 136]
+la2: 1260
-557
+l1e: 1153
-[830 296]
+l0[e]: [800, 136]
+l1e: 557
+l0[e]: [803, 140]
+l1e: 83
+l0[e]: [806, 137]
+l1e: 0
 AA:1 AB:0
@@ -72,9 +80,25 @@
 l0: 1260
+la2: 1276
-1266
+l1e: 1266
+l0[e]: [435, 27]
+l1e: 600
+l0[e]: [801, 136]
+l1e: 90
+l0[e]: [803, 139]
+l1e: 8
-[800 143]
+l0[e]: [807, 143]
+l1e: 0
 AA:1 AB:0
@@ -82,11 +106,23 @@
 l0: 1276
+la2: 1255
-1129
+l1e: 1129
+l0[e]: [803, 139]
-[995 323]
+l1e: 353
+l0[e]: [800, 143]
+l1e: 30
+l0[e]: [1212, 203]
+l1e: 0
-AA:1 AB:0
+AA:2 AB:0
 ```
@@ -182,6 +218,10 @@
+  print("l0:",len(l0))
   l1 = []
@@ -202,77 +242,71 @@
-  print("l0:",len(l0))
   e = 0
+  print("la2:",la2)
   # 近傍点を特定
-  for e in range(len(l0)):
+  for e in range(la1):
-    def oo():
-      l1e = []
+    l1e = []
-      # 探査範囲内のオブジェクト情報を取得
+    # 探査範囲内のオブジェクト情報を取得
-      for l in range(la2):
+    for l in range(la2):
-        if int(l1[l][0]) >= (int(l0[e][0]) - 5):
+      if int(l1[l][0]) >= (int(l0[e][0]) - 5):
-          l1e += [[int(l1[l][0]),int(l1[l][1])]]
+        l1e += [[int(l1[l][0]),int(l1[l][1])]]
-          l += 1
+        l += 1
-        else:
+      else:
-          l += 1
+        l += 1
-        if l == la2:
+      if l == la2:
-          break
+        break
-      return l1e
-    if __name__ == "__main__":
-      l1e = oo()
     l1e = np.array(l1e)
-    print(len(l1e))
+    print("l1e:",len(l1e))
     if len(l1e) == 0:
       break
+    print("l0[e]:",l0[e])
     #モデル構築
-    knn_model = NearestNeighbors(n_neighbors=k, algorithm='brute').fit(l1e)
+    knn_model = NearestNeighbors(n_neighbors=k, algorithm='ball_tree').fit(l1e)
     distances, indices = knn_model.kneighbors([l0[e]])
@@ -294,16 +328,14 @@
     l1 -= kinbou
-    print(kinbou)
     if distances[0][0] < 150 and 1920 >= kinbou[0] >= 0 and 2160 >= kinbou[1] >= 0 :
       AA += [kinbou]
+      #l1e = l1e.tolist()
       #del l1e[:]
       e += 1
@@ -314,21 +346,17 @@
       AB += [kinbou]
+      #l1e = l1e.tolist()
       #del l1e[:]
       e += 1
-    else:
-      #del l1e[:]
-      e += 1
-      break

問題の詳細

2020/12/10 05:40

投稿

shogaku

スコア2

test CHANGED Viewed

File without changes

test CHANGED Viewed

@@ -54,6 +54,10 @@
 出力
+la1: 1263
+l0: 1263
 1153
 [801 136]
@@ -64,12 +68,20 @@
 AA:1 AB:0
+la1: 1260
+l0: 1260
 1266
 [800 143]
 AA:1 AB:0
+la1: 1276
+l0: 1276
 1129
 [995 323]
@@ -146,6 +158,8 @@
+  print("la1:",la1)
   l0 = []
@@ -160,160 +174,164 @@
     i += 1
+    if i == la1:
+      break
+  l1 = []
+  # 後時刻のオブジェクト情報を取得
+  for j in range(la2):
+    l1 += [[int(center2[j][0]),int(center2[j][1])]]
+    j += 1
+    if j == la2:
+      break
+  print("l0:",len(l0))
+  e = 0
+  # 近傍点を特定
+  for e in range(len(l0)):
+    def oo():
+      l1e = []
+      # 探査範囲内のオブジェクト情報を取得
+      for l in range(la2):
+        if int(l1[l][0]) >= (int(l0[e][0]) - 5):
+          l1e += [[int(l1[l][0]),int(l1[l][1])]]
+          l += 1
+        else:
+          l += 1
+        if l == la2:
+          break
+      return l1e
+    if __name__ == "__main__":
+      l1e = oo()
+    l1e = np.array(l1e)
+    print(len(l1e))
+    if len(l1e) == 0:
+      break
+    #モデル構築
+    knn_model = NearestNeighbors(n_neighbors=k, algorithm='brute').fit(l1e)
+    distances, indices = knn_model.kneighbors([l0[e]])
+    #print("K Nearest Neighbors:")
+    #K個までのスライス、現在Kは1個
+    for rank, index in enumerate(indices[0][:k], start=1):
+      str(rank) + " ==>",l1e[index]
+      str(rank) + " ==>", distances[0][0]
+    kinbou = l1e[index]
+    l1 -= kinbou
+    print(kinbou)
+    if distances[0][0] < 150 and 1920 >= kinbou[0] >= 0 and 2160 >= kinbou[1] >= 0 :
+      AA += [kinbou]
+      #del l1e[:]
+      e += 1
+    elif distances[0][0] < 150 and 3840 >= kinbou[0] > 1920 and 2160 >= kinbou[1] >= 0 :
+      AB += [kinbou]
+      #del l1e[:]
+      e += 1
+    else:
+      #del l1e[:]
+      e += 1
+      break
-    if i == la1:
-      break
-  l1 = []
-  # 後時刻のオブジェクト情報を取得
-  for j in range(la2):
-    l1 += [[int(center2[j][0]),int(center2[j][1])]]
-    j += 1
-    if j == la2:
-      break
-  e = 0
-  # 近傍点を特定
-  for e in range(len(l0)):
-    def oo():
-      l1e = []
-      # 探査範囲内のオブジェクト情報を取得
-      for l in range(la2):
-        if int(l1[l][0]) >= (int(l0[e][0]) - 5):
-          l1e += [[int(l1[l][0]),int(l1[l][1])]]
-          l += 1
-        else:
-          l += 1
-        if l == la2:
-          break
-      return l1e
-    if __name__ == "__main__":
-      l1e = oo()
-    l1e = np.array(l1e)
-    print(len(l1e))
-    if len(l1e) == 0:
-      break
-    #モデル構築
-    knn_model = NearestNeighbors(n_neighbors=k, algorithm='brute').fit(l1e)
-    distances, indices = knn_model.kneighbors([l0[e]])
-    #print("K Nearest Neighbors:")
-    #K個までのスライス、現在Kは1個
-    for rank, index in enumerate(indices[0][:k], start=1):
-      str(rank) + " ==>",l1e[index]
-      str(rank) + " ==>", distances[0][0]
-    kinbou = l1e[index]
-    l1 -= kinbou
-    print(kinbou)
-    if distances[0][0] < 150 and 1920 >= kinbou[0] >= 0 and 2160 >= kinbou[1] >= 0 :
-      AA += [kinbou]
-      e += 1
-    elif distances[0][0] < 150 and 3840 >= kinbou[0] > 1920 and 2160 >= kinbou[1] >= 0 :
-      AB += [kinbou]
-      e += 1
-    else:
-      e += 1
-      break
   print("AA" + ":" +  str(len(AA)) + " " + "AB" + ":" +  str(len(AB)))
@@ -324,8 +342,6 @@
     break
 ```

問題の詳細

2020/12/09 11:43

投稿

shogaku

スコア2

test CHANGED Viewed

File without changes

test CHANGED Viewed

@@ -46,9 +46,37 @@
 ### 発生している問題・エラーメッセージ
-このうち「粒子移動後の座標を特定（l0の要素数分繰り返し）」のループが上手く実行されません。そもそも繰り返し処理の組み方が間違っているのでしょうか？
+このうちfor e in range(len(l0)):より下で近傍点の座標をl0の要素の数だけ特定＞範囲のリストに入れていくというループ処理をしたいのですが、リストAA、ABにはそれぞれの画像ごとに格納した座標の個数が一つしか（1回しか）格納されていない？ようです。これはやはりループ処理の組み方がおかしいのでしょうか
+```ここに言語を入力
+出力
+1153
+[801 136]
+557
+[830 296]
+AA:1 AB:0
+1266
+[800 143]
+AA:1 AB:0
+1129
+[995 323]
+AA:1 AB:0
+```
@@ -260,7 +288,7 @@
       AA += [kinbou]
-      del l1e[:]
       e += 1
@@ -270,7 +298,7 @@
       AB += [kinbou]
-      del l1e[:]
       e += 1
@@ -278,7 +306,7 @@
     else:
-      del l1e[:]
       e += 1