Merge branch 'it18006308_pramodya' into 'master'

video Feed added

See merge request !29

Hazardous_object_detection/Hot_objects/hotObject-detection.py

@@ -58,12 +58,12 @@ def forFrame(frame_number, output_array, outpun_count, returned_frame):
                 y2 = person[3]
 
                 # x_center = (x1+x2)/2
-                person_midpoint = ((x1+x2)/2, (y1+y2)/2) # X and Y coordinates
+                person_midpoint = ((x1 + x2) / 2, (y1 + y2) / 2)  # X and Y coordinates
 
                 if kettle_count > 0:
                     # Loop each kettle
                     for kettle in kettles:
-                        kettle_box = box(kettle[0]+50, kettle[3]+50, kettle[1]+50, kettle[2]+50)
+                        kettle_box = box(kettle[0] + 50, kettle[3] + 50, kettle[1] + 50, kettle[2] + 50)
                         # Check weather child is in proximity to the kettle
                         k_x1 = kettle[0] + 50  # Left
                         k_x2 = kettle[1] + 50  # Right
@@ -88,17 +88,17 @@ def forFrame(frame_number, output_array, outpun_count, returned_frame):
                             overlap = True
 
                         # If kettle and child bounding bor overlap, Calculate Distance between center point of two boxes
-                        if(overlap == True):
-                            kettle_midpoint = ((k_x1+k_x2)/2, (k_y1+k_y2)/2)
-                            distance = ((((kettle_midpoint[0] - person_midpoint[0])**2) + ((kettle_midpoint[1]
-                                                                                            - person_midpoint[1])**2) )**0.5)
+                        if (overlap == True):
+                            kettle_midpoint = ((k_x1 + k_x2) / 2, (k_y1 + k_y2) / 2)
+                            distance = ((((kettle_midpoint[0] - person_midpoint[0]) ** 2) + ((kettle_midpoint[1]
+                                                                                              - person_midpoint[
+                                                                                                  1]) ** 2)) ** 0.5)
                 else:
                     print("No Kettles Found")
 
 
-
-
-
-
-
-
+# Video Test
+detector.detectObjectsFromVideo(input_file_path=os.path.join("input/", "video1.mp4"),
+                                output_file_path=os.path.join("output/", "video_frame_analysis"),
+                                frames_per_second=24, per_frame_function=forFrame, minimum_percentage_probability=50,
+                                return_detected_frame=True)

Hazardous_object_detection/Sharp_objects/sharpObject-detection.py

@@ -2,6 +2,7 @@
 
 # Import Libraries
 from imageai.Detection import ObjectDetection, VideoObjectDetection
+from shapely.geometry import box
 from imageai.Detection.Custom import CustomObjectDetection
 
 import os
@@ -35,7 +36,7 @@ def forFrame(frame_number, output_array, outpun_count, returned_frame):
     scissors_count = 0
     scissors = []
     knife_count = 0
-    knife = []
+    knifes = []
     fork_count = 0
     fork = []
     person_count = 0
@@ -62,7 +63,7 @@ def forFrame(frame_number, output_array, outpun_count, returned_frame):
         # RetinaNet Model
         if detection['name'] == 'knife' and detection['percentage probability'] >= THRESHOLD_ITEM:
             knife_count = knife_count + 1
-            knife.append(detection['box_point'])
+            knifes.append(detection['box_point'])
             print(detection["name"], " : ", detection["percentage_probability"], " : ", detection["box_point"])
 
         if detection['name'] == 'fork' and detection['percentage_probability'] >= THRESHOLD_ITEM:
@@ -76,3 +77,67 @@ def forFrame(frame_number, output_array, outpun_count, returned_frame):
 
     if len(person) > 1:
         print("Person in room")
+
+    if len(person) >= 1:
+        for child in person:
+            # Create Person(minx, miny, maxx, maxy)
+            person_box = box(child[0], child[3], child[1], child[2])
+
+            x1 = child[0]
+            x2 = child[1]
+            y1 = child[2]
+            y2 = child[3]
+
+            person_midpoint = ((x1 + x2) / 2, (y1 + y2) / 2)  # (x1, y1)
+
+            if knife_count > 0:
+                # Iterate each Knife
+                for knife in knifes:
+                    knife_box = box(knife[0] + 50, knife[3] + 50, knife[1] + 50, knife[2] + 50)
+
+                    # Check if Knife Overlaps with the Person's Safety Boundary
+                    k_x1 = knife[0] + 50  # Left
+                    k_x2 = knife[1] + 50  # Right
+                    k_y1 = knife[2] + 50  # Top
+                    k_y2 = knife[3] + 50  # Bottom
+                    overlap = False
+
+                    if knife_box.intersection(person_box) != None:
+                        print("Knife - Person is inside the Knives Danger Region")
+                        overlap = True
+
+                    if x1 >= k_x1 & x1 <= k_x2:
+                        print("Knife - X1 is inside the Person's Danger region!")
+                        overlap = True
+                    if x2 >= k_x1 & x2 <= k_x2:
+                        print("Knife - X2 is inside the Person's Danger region!")
+                        overlap = True
+                    if y1 >= k_y2 & y1 <= k_y1:
+                        print("Knife - Y1 is inside the Person's Danger region!")
+                        overlap = True
+                    if y2 >= k_y2 & y2 <= k_y2:
+                        print("Knife - Y2 is inside the Person's Danger region!")
+                        overlap = True
+
+                    if overlap == True:
+                        # If knife and child bounding bor overlap, Calculate Distance between center point of two boxes
+                        knife_midpoint = ((k_x1 + k_x2) / 2, (k_y1 + k_y2) / 2)  # (x2, y2)
+                        distance = ((((knife_midpoint[0] - person_midpoint[0]) ** 2) + (
+                                (knife_midpoint[1] - person_midpoint[1]) ** 2)) ** 0.5)
+
+                        DISTANCE_THRESHOLD_KNIFE_MED = 0
+                        DISTANCE_THRESHOLD_KNIFE_HIGH = 0
+
+                        if distance > DISTANCE_THRESHOLD_KNIFE_HIGH:
+                            print("Person is in the High Danger Zone of Knife")
+                        elif distance > DISTANCE_THRESHOLD_KNIFE_MED:
+                            print("Person is in the Medium Danger Zone of Knife")
+            else:
+                print("No Knives Found")
+
+
+# Testing by Video
+detector.detectObjectsFromVideo(input_file_path=os.path.join("input/", "video1.mp4"),
+                                output_file_path=os.path.join("output/", "video_frame_analysis"),
+                                frames_per_second=24, per_frame_function=forFrame, minimum_percentage_probability=50,
+                                return_detected_frame=True)