Upload 10 files

.gitattributes

@@ -33,3 +33,7 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+example_images/image1.jpg filter=lfs diff=lfs merge=lfs -text
+example_images/image2.jpg filter=lfs diff=lfs merge=lfs -text
+example_images/image3.jpg filter=lfs diff=lfs merge=lfs -text
+example_images/image6.jpg filter=lfs diff=lfs merge=lfs -text

app.py

@@ -0,0 +1,71 @@
+import gradio as gr
+import numpy as np
+from PIL import Image
+import cv2
+from ultralytics import YOLO
+from transformers import pipeline
+import os
+
+# Modelle laden
+yolo_model = YOLO("./best.pt")
+dino_model = pipeline("zero-shot-object-detection", model="IDEA-Research/grounding-dino-tiny")
+
+# YOLOv8-Erkennung
+def detect_with_yolo(image: Image.Image):
+    results = yolo_model(np.array(image))[0]
+    return Image.fromarray(results.plot())
+
+# Grounding DINO-Erkennung
+def detect_with_grounding_dino(image: Image.Image, prompt="license plate"):
+    results = dino_model(image, candidate_labels=[prompt])
+    image_np = np.array(image).copy()
+
+    if not results:
+        return Image.fromarray(image_np)
+
+    results = [result for result in results if result["score"] > 0.4]
+
+    for result in results:
+        box = result["box"]
+        score = result["score"]
+        label = "license plate"
+
+        x1, y1, x2, y2 = int(box["xmin"]), int(box["ymin"]), int(box["xmax"]), int(box["ymax"])
+        image_np = cv2.rectangle(image_np, (x1, y1), (x2, y2), (0, 255, 0), 2)
+        image_np = cv2.putText(image_np, f"{label} ({score:.2f})", (x1, y1 - 10),
+                               cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 1)
+
+    return Image.fromarray(image_np)
+
+# Verarbeitung der Bilder
+def process_image(image):
+    yolo_out = detect_with_yolo(image)
+    dino_out = detect_with_grounding_dino(image)
+    return yolo_out, dino_out
+
+# Beispielbilder definieren
+example_images = [
+    ["example_images/image1.jpg"],
+    ["example_images/image2.jpg"],
+    ["example_images/image3.jpg"],
+    ["example_images/image4.jpg"],
+    ["example_images/image5.jpg"],
+    ["example_images/image6.jpg"],
+    ["example_images/image7.jpg"]
+]
+
+# Gradio-Interface
+app = gr.Interface(
+    fn=process_image,
+    inputs=gr.Image(type="pil"),
+    outputs=[
+        gr.Image(label="YOLOv8 Detection"),
+        gr.Image(label="Grounding DINO (Zero-Shot) Detection")
+    ],
+    examples=example_images,
+    title="Kennzeichenerkennung",
+    description="Lade ein Bild hoch oder wähle ein Beispielbild und vergleiche die Ergebnisse."
+)
+
+if __name__ == "__main__":
+    app.launch()

best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6bafcfd85756a4f6b6c026a515ba2b92b03cab09784ddbd3bbd1523adb5d9705
+size 6217379

requirements.txt

@@ -0,0 +1,5 @@
+ultralytics
+gradio
+transformers
+torch
+opencv-python-headless