Update app.py

app.py

@@ -1,23 +1,20 @@
-import gradio as gr
 import cv2
+import gradio as gr
 import numpy as np
+from sklearn.linear_model import LinearRegression
 import tempfile
-import os
-
-def detect_and_predict(video):
-    cap = cv2.VideoCapture(video)
-    width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
-    height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
 
-    fourcc = cv2.VideoWriter_fourcc(*'mp4v')
-    out_path = tempfile.NamedTemporaryFile(delete=False, suffix='.mp4').name
-    out = cv2.VideoWriter(out_path, fourcc, 20.0, (width, height))
+def process_video(video_path):
+    cap = cv2.VideoCapture(video_path)
+    frame_w = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
+    frame_h = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
 
-    # ... rest of your code continues ...
+    out_path = tempfile.NamedTemporaryFile(delete=False, suffix=".mp4").name
+    out = cv2.VideoWriter(out_path, cv2.VideoWriter_fourcc(*"mp4v"), 20.0, (frame_w, frame_h))
 
-    ball_color_lower = np.array([5, 50, 50])     # Orange/red lower
-    ball_color_upper = np.array([15, 255, 255])  # Orange/red upper
-    trajectory_points = []
+    ball_color_lower = np.array([5, 100, 100])    # HSV orange-red
+    ball_color_upper = np.array([20, 255, 255])
+    trajectory = []
 
     while True:
         ret, frame = cap.read()
@@ -35,29 +32,56 @@ def detect_and_predict(video):
         if contours:
             c = max(contours, key=cv2.contourArea)
             ((x, y), radius) = cv2.minEnclosingCircle(c)
-
             if radius > 3:
-                trajectory_points.append((int(x), int(y)))
-                cv2.circle(frame, (int(x), int(y)), int(radius), (0, 0, 255), 2)
+                center = (int(x), int(y))
+                trajectory.append(center)
+                cv2.circle(frame, center, int(radius), (0, 0, 255), 2)
 
-        # Draw trajectory
-        for i in range(1, len(trajectory_points)):
-            cv2.line(frame, trajectory_points[i - 1], trajectory_points[i], (255, 0, 0), 2)
+        # Draw stumps area
+        stump_box = (frame_w // 2 - 30, frame_h - 120, frame_w // 2 + 30, frame_h - 50)
+        cv2.rectangle(frame, stump_box[:2], stump_box[2:], (0, 255, 255), 2)
 
-        # Draw stumps line (for simplicity, fixed zone)
-        cv2.rectangle(frame, (width // 2 - 20, height - 200), (width // 2 + 20, height - 50), (0, 255, 255), 2)
+        # Draw trajectory
+        for i in range(1, len(trajectory)):
+            cv2.line(frame, trajectory[i - 1], trajectory[i], (255, 0, 0), 2)
 
         out.write(frame)
 
     cap.release()
     out.release()
 
+    decision = "NOT OUT"
+    projected_hit = False
+
+    # Predict trajectory
+    if len(trajectory) >= 5:
+        X = np.array([x for x, y in trajectory]).reshape(-1, 1)
+        y = np.array([y for x, y in trajectory])
+
+        model = LinearRegression()
+        model.fit(X, y)
+
+        # Predict at stump x-position
+        stump_x = frame_w // 2
+        pred_y = int(model.predict(np.array([[stump_x]]))[0])
+
+        # Check if predicted Y is in stump area
+        if stump_box[1] <= pred_y <= stump_box[3]:
+            projected_hit = True
+            decision = "OUT"
+
+    # Write final frame with decision
+    final_frame = cv2.imread(out_path)
+    cv2.putText(final_frame, f"DECISION: {decision}", (50, 80), cv2.FONT_HERSHEY_SIMPLEX, 2, (0, 255, 0) if decision == "NOT OUT" else (0, 0, 255), 4)
+
     return out_path
 
-iface = gr.Interface(fn=detect_and_predict,
-                     inputs=gr.Video(label="Upload Bowling Video"),
-                     outputs=gr.Video(label="Ball Tracking Result"),
-                     title="DRS Ball Tracker",
-                     description="Detect and visualize ball trajectory for LBW simulation.")
+iface = gr.Interface(
+    fn=process_video,
+    inputs=gr.Video(label="Upload Bowling Video (Camera View)"),
+    outputs=gr.Video(label="Replay with Trajectory and Decision"),
+    title="Cricket DRS LBW Tracker",
+    description="Track ball trajectory and check if it's LBW OUT based on impact prediction."
+)
 
 iface.launch()