Update app.py

app.py

@@ -1,41 +1,63 @@
-import gradio as gr
-import torch
-from transformers import DetrImageProcessor, DetrForObjectDetection
+import cv2
+import numpy as np
 from PIL import Image, ImageDraw
+import gradio as gr
 
-# Load the pre-trained DETR model and processor
-processor = DetrImageProcessor.from_pretrained("facebook/detr-resnet-50")
-model = DetrForObjectDetection.from_pretrained("facebook/detr-resnet-50")
-
-def detect_objects(image: Image.Image) -> Image.Image:
+def detect_cracks(image: Image.Image) -> Image.Image:
     try:
-        # Preprocess the image
-        inputs = processor(images=image, return_tensors="pt")
-        outputs = model(**inputs)
-
-        # Convert outputs to bounding boxes and labels
-        target_sizes = torch.tensor([image.size[::-1]])
-        results = processor.post_process_object_detection(outputs, target_sizes=target_sizes, threshold=0.9)[0]
-
-        # Draw bounding boxes on the image
-        draw = ImageDraw.Draw(image)
-        for score, label, box in zip(results["scores"], results["labels"], results["boxes"]):
-            box = [round(i, 2) for i in box.tolist()]
-            label_text = f"{model.config.id2label[label.item()]}: {round(score.item(), 3)}"
-            draw.rectangle(box, outline="red", width=3)
-            draw.text((box[0], box[1]), label_text, fill="red")
-        return image
+        # Convert PIL image to an OpenCV image (BGR format)
+        cv_image = cv2.cvtColor(np.array(image), cv2.COLOR_RGB2BGR)
+        
+        # Convert to grayscale for processing
+        gray = cv2.cvtColor(cv_image, cv2.COLOR_BGR2GRAY)
+        
+        # Apply Gaussian blur to reduce noise and enhance edges
+        blurred = cv2.GaussianBlur(gray, (5, 5), 0)
+        
+        # Use adaptive thresholding to highlight potential crack areas
+        thresh = cv2.adaptiveThreshold(
+            blurred, 255,
+            cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
+            cv2.THRESH_BINARY_INV,
+            11, 2
+        )
+        
+        # Apply morphological closing to bridge gaps in detected lines
+        kernel = np.ones((3, 3), np.uint8)
+        morph = cv2.morphologyEx(thresh, cv2.MORPH_CLOSE, kernel, iterations=2)
+        
+        # Detect edges with Canny edge detector
+        edges = cv2.Canny(morph, 50, 150)
+        
+        # Find contours based on the detected edges
+        contours, _ = cv2.findContours(edges, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
+        
+        # Convert original image to PIL for drawing
+        annotated = image.copy()
+        draw = ImageDraw.Draw(annotated)
+        
+        # Draw bounding boxes around contours that are large enough to be meaningful cracks
+        for cnt in contours:
+            # Filter out noise with a minimum arc length threshold (adjustable)
+            if cv2.arcLength(cnt, True) > 100:
+                x, y, w, h = cv2.boundingRect(cnt)
+                draw.rectangle([x, y, x + w, y + h], outline="red", width=2)
+        
+        return annotated
     except Exception as e:
-        print("Error during detection:", e)
-        return image  # In a robust production system, consider returning a message or a default image
+        print("Error during crack detection:", e)
+        return image  # Fallback: return the original image if any error occurs
 
-# Create a Gradio interface
+# Create a Gradio interface for the Space
 iface = gr.Interface(
-    fn=detect_objects,
-    inputs=gr.Image(type="pil", label="Upload an Image"),
-    outputs=gr.Image(label="Detection Result"),
-    title="Robust Object Detection with DETR",
-    description="Upload an image to detect objects using a pre-trained DETR model from Hugging Face Hub."
+    fn=detect_cracks,
+    inputs=gr.Image(type="pil", label="Upload a Floor/Wall Image"),
+    outputs=gr.Image(label="Detected Cracks"),
+    title="Home Inspection: Crack Detection",
+    description=(
+        "Upload an image of a floor or wall to detect cracks and other defects. "
+        "This demo uses traditional computer vision techniques to highlight potential issues."
+    )
 )
 
 if __name__ == "__main__":