質問編集履歴

コードの記載方法を教えていただいた通り修正しました。

2021/08/02 14:13

投稿

trafield

スコア0

test CHANGED Viewed

File without changes

test CHANGED Viewed

@@ -34,19 +34,133 @@
 ### 該当のソースコード
-social_distanciation_video_detection.pyのコード
+social_distanciation_video_detection.py
+```python
+#########################################
+#		     Select the model 			#
+#########################################
+model_names_list = [name for name in os.listdir("../models/.") if name.find(".") == -1]
+for index,model_name in enumerate(model_names_list):
+    print(" - {} [{}]".format(model_name,index))
+model_num = input(" Please select the number related to the model that you want : ")
+if model_num == "":
+	model_path="../models/faster_rcnn_inception_v2_coco_2018_01_28/frozen_inference_graph.pb"
+else :
+	model_path = "../models/"+model_names_list[int(model_num)]+"/frozen_inference_graph.pb"
+print(bcolors.WARNING + " [ Loading the TENSORFLOW MODEL ... ]"+bcolors.ENDC)
+model = Model(model_path)
+print(bcolors.OKGREEN +"Done : [ Model loaded and initialized ] ..."+bcolors.ENDC)
+#########################################
+#		     Select the video 			#
+#########################################
+video_names_list = [name for name in os.listdir("../video/") if name.endswith(".mp4") or name.endswith(".avi")]
+for index,video_name in enumerate(video_names_list):
+    print(" - {} [{}]".format(video_name,index))
+video_num = input("Enter the exact name of the video (including .mp4 or else) : ")
+if video_num == "":
+	video_path="../video/PETS2009.avi"
+else :
+	video_path = "../video/"+video_names_list[int(video_num)]
+#########################################
+#		    Minimal distance			#
+#########################################
+distance_minimum = input("Prompt the size of the minimal distance between 2 pedestrians : ")
+if distance_minimum == "":
+	distance_minimum = "110"
+#########################################
+#     Compute transformation matrix		#
+#########################################
+# Compute  transformation matrix from the original frame
+matrix,imgOutput = compute_perspective_transform(corner_points,width_og,height_og,cv2.imread(img_path))
+height,width,_ = imgOutput.shape
+blank_image = np.zeros((height,width,3), np.uint8)
+height = blank_image.shape[0]
+width = blank_image.shape[1]
+dim = (width, height)
+######################################################
+#########									 #########
+# 				START THE VIDEO STREAM               #
+#########									 #########
+######################################################
 vs = cv2.VideoCapture(video_path)
 output_video_1,output_video_2 = None,None
-＃ Loop until the end of the video stream
+# Loop until the end of the video stream
 while True:
-	＃ Load the image of the ground and resize it to the correct size
+	# Load the image of the ground and resize it to the correct size
 	img = cv2.imread("../img/chemin_1.png")
@@ -54,11 +168,11 @@
-	＃Load the frame
+	# Load the frame
 	(frame_exists, frame) = vs.read()
-	＃ Test if it has reached the end of the video
+	# Test if it has reached the end of the video
 	if not frame_exists:
@@ -66,37 +180,37 @@
 	else:
-		＃ Resize the image to the correct size
+		# Resize the image to the correct size
 		frame = imutils.resize(frame, width=int(size_frame))
-		＃ Make the predictions for this frame
+		# Make the predictions for this frame
 		(boxes, scores, classes) =  model.predict(frame)
-		＃ Get the human detected in the frame and return the 2 points to build the bounding box
+		# Get the human detected in the frame and return the 2 points to build the bounding box
 		array_boxes_detected = get_human_box_detection(boxes,scores[0].tolist(),classes[0].tolist(),frame.shape[0],frame.shape[1])
-		＃ Both of our lists that will contain the centroﾃｯds coordonates and the ground points
+		# Both of our lists that will contain the centroﾃｯds coordonates and the ground points
 		array_centroids,array_groundpoints = get_centroids_and_groundpoints(array_boxes_detected)
-		＃ Use the transform matrix to get the transformed coordonates
+		# Use the transform matrix to get the transformed coordonates
 		transformed_downoids = compute_point_perspective_transformation(matrix,array_groundpoints)
-		＃ Show every point on the top view image
+		# Show every point on the top view image
 		for point in transformed_downoids:
@@ -108,7 +222,7 @@
-		＃ Check if 2 or more people have been detected (otherwise no need to detect)
+		# Check if 2 or more people have been detected (otherwise no need to detect)
 		if len(transformed_downoids) >= 2:
@@ -120,23 +234,23 @@
-			＃ Iterate over every possible 2 by 2 between the points combinations
+			# Iterate over every possible 2 by 2 between the points combinations
 			list_indexes = list(itertools.combinations(range(len(transformed_downoids)), 2))
 			for i,pair in enumerate(itertools.combinations(transformed_downoids, r=2)):
-				＃ Check if the distance between each combination of points is less than the minimum distance chosen
+				# Check if the distance between each combination of points is less than the minimum distance chosen
 				if math.sqrt( (pair[0][0] - pair[1][0])**2 + (pair[0][1] - pair[1][1])**2 ) < int(distance_minimum):
-					＃ Change the colors of the points that are too close from each other to red
+					# Change the colors of the points that are too close from each other to red
 					if not (pair[0][0] > width or pair[0][0] < 0 or pair[0][1] > height+200  or pair[0][1] < 0 or pair[1][0] > width or pair[1][0] < 0 or pair[1][1] > height+200  or pair[1][1] < 0):
 						change_color_on_topview(pair)
-						＃ Get the equivalent indexes of these points in the original frame and change the color to red
+						# Get the equivalent indexes of these points in the original frame and change the color to red
 						index_pt1 = list_indexes[i][0]
@@ -150,11 +264,11 @@
-	＃ Draw the green rectangle to delimitate the detection zone
+	# Draw the green rectangle to delimitate the detection zone
 	draw_rectangle(corner_points)
-	＃ Show both images
+	# Show both images
 	cv2.imshow("Bird view", bird_view_img)
@@ -168,7 +282,7 @@
-	＃ Write the both outputs video to a local folders
+	# Write the both outputs video to a local folders
 	if output_video_1 is None and output_video_2 is None:
@@ -188,7 +302,7 @@
-	＃ Break the loop
+	# Break the loop
 	if key == ord("q"):
@@ -196,11 +310,13 @@
+```
-#bird_view_transfo_functions.pyのコード
+bird_view_transfo_functions.py
+```python
 import numpy as np
@@ -220,15 +336,15 @@
     """
-	＃ Create an array out of the 4 corner points
+	# Create an array out of the 4 corner points
 	corner_points_array = np.float32(corner_points)
-	＃ Create an array with the parameters (the dimensions) required to build the matrix
+	# Create an array with the parameters (the dimensions) required to build the matrix
 	img_params = np.float32([[0,0],[width,0],[0,height],[width,height]])
-	＃ Compute and return the transformation matrix
+	# Compute and return the transformation matrix
 	matrix = cv2.getPerspectiveTransform(corner_points_array,img_params)
@@ -252,13 +368,13 @@
     """
-	＃ Compute the new coordinates of our points
+	# Compute the new coordinates of our points
 	list_points_to_detect = np.float32(list_downoids).reshape(-1, 1, 2)
 	transformed_points = cv2.perspectiveTransform(list_points_to_detect, matrix)
-	＃ Loop over the points and add them to the list that will be returned
+	# Loop over the points and add them to the list that will be returned
 	transformed_points_list = list()
@@ -267,3 +383,7 @@
 		transformed_points_list.append([transformed_points[i][0][0],transformed_points[i][0][1]])
 	return transformed_points_list
+```

追加したコードを修正しました。

2021/08/02 14:13

投稿

trafield

スコア0

test CHANGED Viewed

File without changes

test CHANGED Viewed

@@ -30,31 +30,23 @@
 AttributeError: 'NoneType' object has no attribute 'shape'
-``### 該当のソースコード
+### 該当のソースコード
 social_distanciation_video_detection.pyのコード
-######################################################
-#########									 #########
-# 				START THE VIDEO STREAM               #
-#########									 #########
-######################################################
 vs = cv2.VideoCapture(video_path)
 output_video_1,output_video_2 = None,None
-# Loop until the end of the video stream
+＃ Loop until the end of the video stream
 while True:
-	# Load the image of the ground and resize it to the correct size
+	＃ Load the image of the ground and resize it to the correct size
 	img = cv2.imread("../img/chemin_1.png")
@@ -62,11 +54,11 @@
-	# Load the frame
+	＃Load the frame
 	(frame_exists, frame) = vs.read()
-	# Test if it has reached the end of the video
+	＃ Test if it has reached the end of the video
 	if not frame_exists:
@@ -74,37 +66,37 @@
 	else:
-		# Resize the image to the correct size
+		＃ Resize the image to the correct size
 		frame = imutils.resize(frame, width=int(size_frame))
-		# Make the predictions for this frame
+		＃ Make the predictions for this frame
 		(boxes, scores, classes) =  model.predict(frame)
-		# Get the human detected in the frame and return the 2 points to build the bounding box
+		＃ Get the human detected in the frame and return the 2 points to build the bounding box
 		array_boxes_detected = get_human_box_detection(boxes,scores[0].tolist(),classes[0].tolist(),frame.shape[0],frame.shape[1])
-		# Both of our lists that will contain the centroﾃｯds coordonates and the ground points
+		＃ Both of our lists that will contain the centroﾃｯds coordonates and the ground points
 		array_centroids,array_groundpoints = get_centroids_and_groundpoints(array_boxes_detected)
-		# Use the transform matrix to get the transformed coordonates
+		＃ Use the transform matrix to get the transformed coordonates
 		transformed_downoids = compute_point_perspective_transformation(matrix,array_groundpoints)
-		# Show every point on the top view image
+		＃ Show every point on the top view image
 		for point in transformed_downoids:
@@ -116,7 +108,7 @@
-		# Check if 2 or more people have been detected (otherwise no need to detect)
+		＃ Check if 2 or more people have been detected (otherwise no need to detect)
 		if len(transformed_downoids) >= 2:
@@ -128,23 +120,23 @@
-			# Iterate over every possible 2 by 2 between the points combinations
+			＃ Iterate over every possible 2 by 2 between the points combinations
 			list_indexes = list(itertools.combinations(range(len(transformed_downoids)), 2))
 			for i,pair in enumerate(itertools.combinations(transformed_downoids, r=2)):
-				# Check if the distance between each combination of points is less than the minimum distance chosen
+				＃ Check if the distance between each combination of points is less than the minimum distance chosen
 				if math.sqrt( (pair[0][0] - pair[1][0])**2 + (pair[0][1] - pair[1][1])**2 ) < int(distance_minimum):
-					# Change the colors of the points that are too close from each other to red
+					＃ Change the colors of the points that are too close from each other to red
 					if not (pair[0][0] > width or pair[0][0] < 0 or pair[0][1] > height+200  or pair[0][1] < 0 or pair[1][0] > width or pair[1][0] < 0 or pair[1][1] > height+200  or pair[1][1] < 0):
 						change_color_on_topview(pair)
-						# Get the equivalent indexes of these points in the original frame and change the color to red
+						＃ Get the equivalent indexes of these points in the original frame and change the color to red
 						index_pt1 = list_indexes[i][0]
@@ -158,11 +150,11 @@
-	# Draw the green rectangle to delimitate the detection zone
+	＃ Draw the green rectangle to delimitate the detection zone
 	draw_rectangle(corner_points)
-	# Show both images
+	＃ Show both images
 	cv2.imshow("Bird view", bird_view_img)
@@ -176,7 +168,7 @@
-	# Write the both outputs video to a local folders
+	＃ Write the both outputs video to a local folders
 	if output_video_1 is None and output_video_2 is None:
@@ -196,7 +188,7 @@
-	# Break the loop
+	＃ Break the loop
 	if key == ord("q"):
@@ -204,7 +196,7 @@
-bird_view_transfo_functions.pyのコード
+#bird_view_transfo_functions.pyのコード
@@ -228,15 +220,15 @@
     """
-	# Create an array out of the 4 corner points
+	＃ Create an array out of the 4 corner points
 	corner_points_array = np.float32(corner_points)
-	# Create an array with the parameters (the dimensions) required to build the matrix
+	＃ Create an array with the parameters (the dimensions) required to build the matrix
 	img_params = np.float32([[0,0],[width,0],[0,height],[width,height]])
-	# Compute and return the transformation matrix
+	＃ Compute and return the transformation matrix
 	matrix = cv2.getPerspectiveTransform(corner_points_array,img_params)
@@ -260,13 +252,13 @@
     """
-	# Compute the new coordinates of our points
+	＃ Compute the new coordinates of our points
 	list_points_to_detect = np.float32(list_downoids).reshape(-1, 1, 2)
 	transformed_points = cv2.perspectiveTransform(list_points_to_detect, matrix)
-	# Loop over the points and add them to the list that will be returned
+	＃ Loop over the points and add them to the list that will be returned
 	transformed_points_list = list()

コードを追加しました。

2021/08/01 13:49

投稿

trafield

スコア0

test CHANGED Viewed

File without changes

test CHANGED Viewed

@@ -18,12 +18,6 @@
-### 該当のソースコード
 Traceback (most recent call last):
   File "social_distanciation_video_detection.py", line 193, in <module>
@@ -36,4 +30,248 @@
 AttributeError: 'NoneType' object has no attribute 'shape'
+``### 該当のソースコード
+social_distanciation_video_detection.pyのコード
+######################################################
+#########									 #########
+# 				START THE VIDEO STREAM               #
+#########									 #########
+######################################################
+vs = cv2.VideoCapture(video_path)
+output_video_1,output_video_2 = None,None
+# Loop until the end of the video stream
+while True:
+	# Load the image of the ground and resize it to the correct size
+	img = cv2.imread("../img/chemin_1.png")
+	bird_view_img = cv2.resize(img, dim, interpolation = cv2.INTER_AREA)
+	# Load the frame
+	(frame_exists, frame) = vs.read()
+	# Test if it has reached the end of the video
+	if not frame_exists:
-``
+		break
+	else:
+		# Resize the image to the correct size
+		frame = imutils.resize(frame, width=int(size_frame))
+		# Make the predictions for this frame
+		(boxes, scores, classes) =  model.predict(frame)
+		# Get the human detected in the frame and return the 2 points to build the bounding box
+		array_boxes_detected = get_human_box_detection(boxes,scores[0].tolist(),classes[0].tolist(),frame.shape[0],frame.shape[1])
+		# Both of our lists that will contain the centroﾃｯds coordonates and the ground points
+		array_centroids,array_groundpoints = get_centroids_and_groundpoints(array_boxes_detected)
+		# Use the transform matrix to get the transformed coordonates
+		transformed_downoids = compute_point_perspective_transformation(matrix,array_groundpoints)
+		# Show every point on the top view image
+		for point in transformed_downoids:
+			x,y = point
+			cv2.circle(bird_view_img, (int(x),int(y)), BIG_CIRCLE, COLOR_GREEN, 2)
+			cv2.circle(bird_view_img, (int(x),int(y)), SMALL_CIRCLE, COLOR_GREEN, -1)
+		# Check if 2 or more people have been detected (otherwise no need to detect)
+		if len(transformed_downoids) >= 2:
+			for index,downoid in enumerate(transformed_downoids):
+				if not (downoid[0] > width or downoid[0] < 0 or downoid[1] > height+200 or downoid[1] < 0 ):
+					cv2.rectangle(frame,(array_boxes_detected[index][1],array_boxes_detected[index][0]),(array_boxes_detected[index][3],array_boxes_detected[index][2]),COLOR_GREEN,2)
+			# Iterate over every possible 2 by 2 between the points combinations
+			list_indexes = list(itertools.combinations(range(len(transformed_downoids)), 2))
+			for i,pair in enumerate(itertools.combinations(transformed_downoids, r=2)):
+				# Check if the distance between each combination of points is less than the minimum distance chosen
+				if math.sqrt( (pair[0][0] - pair[1][0])**2 + (pair[0][1] - pair[1][1])**2 ) < int(distance_minimum):
+					# Change the colors of the points that are too close from each other to red
+					if not (pair[0][0] > width or pair[0][0] < 0 or pair[0][1] > height+200  or pair[0][1] < 0 or pair[1][0] > width or pair[1][0] < 0 or pair[1][1] > height+200  or pair[1][1] < 0):
+						change_color_on_topview(pair)
+						# Get the equivalent indexes of these points in the original frame and change the color to red
+						index_pt1 = list_indexes[i][0]
+						index_pt2 = list_indexes[i][1]
+						cv2.rectangle(frame,(array_boxes_detected[index_pt1][1],array_boxes_detected[index_pt1][0]),(array_boxes_detected[index_pt1][3],array_boxes_detected[index_pt1][2]),COLOR_RED,2)
+						cv2.rectangle(frame,(array_boxes_detected[index_pt2][1],array_boxes_detected[index_pt2][0]),(array_boxes_detected[index_pt2][3],array_boxes_detected[index_pt2][2]),COLOR_RED,2)
+	# Draw the green rectangle to delimitate the detection zone
+	draw_rectangle(corner_points)
+	# Show both images
+	cv2.imshow("Bird view", bird_view_img)
+	cv2.imshow("Original picture", frame)
+	key = cv2.waitKey(1) & 0xFF
+	# Write the both outputs video to a local folders
+	if output_video_1 is None and output_video_2 is None:
+		fourcc1 = cv2.VideoWriter_fourcc(*"MJPG")
+		output_video_1 = cv2.VideoWriter("../output/video.avi", fourcc1, 25,(frame.shape[1], frame.shape[0]), True)
+		fourcc2 = cv2.VideoWriter_fourcc(*"MJPG")
+		output_video_2 = cv2.VideoWriter("../output/bird_view.avi", fourcc2, 25,(bird_view_img.shape[1], bird_view_img.shape[0]), True)
+	elif output_video_1 is not None and output_video_2 is not None:
+		output_video_1.write(frame)
+		output_video_2.write(bird_view_img)
+	# Break the loop
+	if key == ord("q"):
+		break
+bird_view_transfo_functions.pyのコード
+import numpy as np
+import cv2
+def compute_perspective_transform(corner_points,width,height,image):
+	""" Compute the transformation matrix
+    @ corner_points : 4 corner points selected from the image
+    @ height, width : size of the image
+    """
+	# Create an array out of the 4 corner points
+	corner_points_array = np.float32(corner_points)
+	# Create an array with the parameters (the dimensions) required to build the matrix
+	img_params = np.float32([[0,0],[width,0],[0,height],[width,height]])
+	# Compute and return the transformation matrix
+	matrix = cv2.getPerspectiveTransform(corner_points_array,img_params)
+	img_transformed = cv2.warpPerspective(image,matrix,(width,height))
+	return matrix,img_transformed
+def compute_point_perspective_transformation(matrix,list_downoids):
+	""" Apply the perspective transformation to every ground point which have been detected on the main frame.
+    @ matrix : the 3x3 matrix
+    @ list_downoids : list that contains the points to transform
+    return : list containing all the new points
+    """
+	# Compute the new coordinates of our points
+	list_points_to_detect = np.float32(list_downoids).reshape(-1, 1, 2)
+	transformed_points = cv2.perspectiveTransform(list_points_to_detect, matrix)
+	# Loop over the points and add them to the list that will be returned
+	transformed_points_list = list()
+	for i in range(0,transformed_points.shape[0]):
+		transformed_points_list.append([transformed_points[i][0][0],transformed_points[i][0][1]])
+	return transformed_points_list