Create app.py

app.py

@@ -0,0 +1,79 @@
+import cv2
+import numpy as np
+import gradio as gr
+from ultralytics import YOLO
+from typing import List, Dict
+
+# Load models (automatically downloads on first run)
+BALL_MODEL = YOLO('models/yolov8n.pt')  # For ball tracking
+STUMP_MODEL = YOLO('models/yolov8m.pt')  # For LBW detection
+
+def process_video(video_path: str) -> Dict:
+    """Process video with YOLO models"""
+    cap = cv2.VideoCapture(video_path)
+    fps = cap.get(cv2.CAP_PROP_FPS)
+    frames = []
+    analytics = {
+        "max_speed": 0,
+        "events": [],
+        "ball_positions": []
+    }
+    
+    prev_pos = None
+    
+    while cap.isOpened():
+        ret, frame = cap.read()
+        if not ret:
+            break
+        
+        # Resize for consistent processing
+        frame = cv2.resize(frame, (1280, 720))
+        
+        # Ball detection
+        ball_results = BALL_MODEL(frame, classes=32)  # Class 32 = sports ball
+        ball_boxes = ball_results[0].boxes.xyxy.cpu().numpy()
+        
+        if len(ball_boxes) > 0:
+            x1, y1, x2, y2 = map(int, ball_boxes[0])
+            x, y = (x1 + x2) // 2, (y1 + y2) // 2  # Center point
+            
+            # Speed calculation
+            if prev_pos:
+                speed = np.sqrt((x - prev_pos[0])**2 + (y - prev_pos[1])**2) * fps * 3.6 / 100  # km/h
+                analytics["max_speed"] = max(analytics["max_speed"], speed)
+            
+            # Draw tracking
+            cv2.circle(frame, (x, y), 10, (0, 255, 0), -1)
+            cv2.putText(frame, f"{speed:.1f} km/h", (x+15, y), 
+                       cv2.FONT_HERSHEY_SIMPLEX, 0.7, (255, 255, 0), 2)
+            
+            prev_pos = (x, y)
+            analytics["ball_positions"].append((x, y))
+        
+        # LBW detection (every 10 frames)
+        if len(analytics["ball_positions"]) % 10 == 0:
+            stump_results = STUMP_MODEL(frame, classes=33)  # Class 33 = sports equipment
+            # Add LBW logic here
+        
+        frames.append(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB))
+    
+    cap.release()
+    return {"frames": frames, "analytics": analytics}
+
+# Gradio UI
+with gr.Blocks(theme=gr.themes.Soft()) as demo:
+    gr.Markdown("# 🏏 AI Cricket Analyzer")
+    with gr.Row():
+        input_video = gr.Video(label="Input Video", format="mp4")
+        output_video = gr.Video(label="AI Analysis")
+    with gr.Accordion("Advanced Settings", open=False):
+        model_size = gr.Dropdown(["Nano", "Small", "Medium"], value="Nano")
+    analyze_btn = gr.Button("Analyze", variant="primary")
+    
+    analyze_btn.click(
+        fn=process_video,
+        inputs=input_video,
+        outputs=output_video
+    )
+
+demo.launch()