work

Author: abhiramasonny

gunImages/gunMan.jpg

@@ -8,7 +8,7 @@ def main(tryall=True):
     gun_detection = GunDetection(weight_path_gun, config_path_gun)
 
     if not tryall:
-        cam1 = cv2.VideoCapture(0)
+        cam1 = cv2.VideoCapture(1)
         #cam2 = cv2.VideoCapture(1)
         cameras = [
             (cam1, "Zone 1")#, (cam2, 'Zone 2')
@@ -34,6 +34,7 @@ def main(tryall=True):
         else:
             # if this is running we have had a gun man for more that 4 ticks 
             print('you goofy goober')
+            frame = cv2.imread("gunImages/gunMan.jpg")
             clothes_detection = ClothesDetection(len(cameras))
             wearing,colour = clothes_detection.run_detection(cameras)
             print(f"wearing {wearing} and RGB {colour}")

main.py

@@ -87,25 +87,38 @@ def run_detection(self, cameras: list, single_frame_mode: bool = False, single_f
         yield False
 
 class ClothesDetection:
-    def __init__(self, num_cameras: int):
-        self.models = [Load_DeepFashion2_Yolov3() for _ in range(num_cameras)]
-        self.cap_list = [cv2.VideoCapture(i) for i in range(num_cameras)]
+    def __init__(self, num_cameras: int, isframe=False, frame=None):
+        self.isframe = isframe
+        if not self.isframe:
+            self.models = [Load_DeepFashion2_Yolov3() for _ in range(num_cameras)]
+            self.cap_list = [cv2.VideoCapture(i) for i in range(num_cameras)]
+        else:
+            self.models = [Load_DeepFashion2_Yolov3()]
+            self.frame = frame
         self.start_time = time.time()
 
     def run_detection(self, cameras: list):
         gunman_is_wearing = ""
         avg_rgb_values = []
         while len(gunman_is_wearing)==0:
-            frames = [cap.read()[1] for cap in self.cap_list]
-            ret_list = [frame is not None for frame in frames]
-            if not all(ret_list):
-                break
-
-            img_tensors = [tf.convert_to_tensor(frame[tf.newaxis, ...], dtype=tf.float32) / 255.0 for frame in frames]
-            img_with_boxes_list = []
-            box_array_list = []
+            if not self.isframe:
+                frames = [cap.read()[1] for cap in self.cap_list]
+                ret_list = [frame is not None for frame in frames]
+                if not all(ret_list):
+                    break
 
-            text = ""
+            if not self.isframe:
+                img_tensors = [tf.convert_to_tensor(frame[tf.newaxis, ...], dtype=tf.float32) / 255.0 for frame in frames]
+                img_with_boxes_list = []
+                box_array_list = []
+                
+                text = ""
+            else:
+                img_tensors = [tf.convert_to_tensor(self.frame[tf.newaxis, ...], dtype=tf.float32) / 255.0]
+                img_with_boxes_list = []
+                box_array_list = []
+                
+                text = ""
 
             for i, (img_tensor, model) in enumerate(zip(img_tensors, self.models)):
                 img_with_boxes, box_array, text = Draw_Bounding_Box(frames[i], Detect_Clothes(img_tensor, model))
@@ -120,24 +133,26 @@ def run_detection(self, cameras: list):
                     #        number=2142184754, provider="T-Mobile")
                     avg_rgb_values = self.get_avg_rgb(box_array, frames[i])
                     print(avg_rgb_values)
+
                     #color_category = self.get_color_category(avg_rgb_values)
                     #print(f"Color category for cam {i + 1}: {color_category}")
 
             elapsed_time = time.time() - self.start_time
             fps = len(cameras) / elapsed_time
 
             for i, img_with_boxes in enumerate(img_with_boxes_list):
-                cv2.putText(img_with_boxes, f"FPS: {fps:.2f}", (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 255, 0), 2)
-                cv2.putText(img_with_boxes, f"Gunman is wearing {gunman_is_wearing}", (10, 100), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 255, 0), 2)
-                cv2.putText(img_with_boxes, f"Color rgb {avg_rgb_values}", (10, 200), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 255, 0), 2)
-                cv2.imshow(f"Clothes detection cam {i + 1}", img_with_boxes)
+                if not self.isframe:
+                    cv2.putText(img_with_boxes, f"FPS: {fps:.2f}", (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 255, 0), 2)
+                    cv2.putText(img_with_boxes, f"Gunman is wearing {gunman_is_wearing}", (10, 100), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 255, 0), 2)
+                    cv2.putText(img_with_boxes, f"Color rgb {avg_rgb_values}", (10, 200), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 255, 0), 2)
+                    cv2.imshow(f"Clothes detection cam {i + 1}", img_with_boxes)
 
             if cv2.waitKey(1) & 0xFF == ord('q'):
                 break
-
-        for cap in self.cap_list:
-            cap.release()
-        cv2.destroyAllWindows()
+        if not self.isframe:
+            for cap in self.cap_list:
+                cap.release()
+            cv2.destroyAllWindows()
         return (gunman_is_wearing,avg_rgb_values)
     def get_avg_rgb(self, box_array: list, frame: np.ndarray) -> list:
         avg_rgb_values = []