From d422268a030166d4743653fad00010e374622305 Mon Sep 17 00:00:00 2001
From: Stef <stbuwalda@gmail.com>
Date: Thu, 12 Jun 2025 11:34:42 +0200
Subject: [PATCH] Only OCR the biggest bounding box

---
 ANPR.py | 59 +++++++++++++++++++++++++++++++--------------------------
 1 file changed, 32 insertions(+), 27 deletions(-)

diff --git a/ANPR.py b/ANPR.py
index 87abb50..0b0f470 100644
--- a/ANPR.py
+++ b/ANPR.py
@@ -1,4 +1,5 @@
 from ultralytics import YOLO
+from ultralytics.engine.results import Boxes
 from PIL import Image
 import numpy as np
 import easyocr
@@ -11,31 +12,35 @@ img = Image.open("test.jpg")
 
 results = car_model.predict(source=img)
 
-for r in results:
-    for box in r.boxes:
-        cls_id = int(box.cls[0])
-        cls_name = r.names[cls_id]
-        if cls_name == "car":
-            conf = float(box.conf[0])
-            x1, y1, x2, y2 = map(int, box.xyxy[0])
-            print(f"{cls_name} ({conf:.2f}) at [{x1},{y1},{x2},{y2}]")
+cars: list[tuple[int, tuple[int, int, int, int]]] = []
 
-            # Crop Image
-            cropped_img = img.crop((x1, y1, x2, y2))
-            cropped_img.save("car_crop_pillow.jpg")
-            results2 = plate_model.predict(source=cropped_img)
-            for x in results2:
-                for box2 in x.boxes:
-                    cls_id = int(box2.cls[0])
-                    cls_name = x.names[cls_id]
-                    if cls_name == "License_Plate":
-                        conf = float(box2.conf[0])
-                        x1, y1, x2, y2 = map(int, box2.xyxy[0])
-                        print(
-                            f"{cls_name} ({conf:.2f}) at [{x1},{y1},{x2},{y2}]"
-                        )
-                        lp_img = cropped_img.crop((x1, y1, x2, y2))
-                        lp_img.save("license_plate.jpg")
-                        lp_np = np.array(lp_img)
-                        result3 = reader.readtext(lp_np)
-                        print(result3)
+# Filter out the cars and calculate box size
+for r in results:
+    if r.boxes:
+        for box in r.boxes:
+            cls_name = r.names[int(box.cls[0])]
+            if cls_name == "car":
+                x1, y1, x2, y2 = map(int, box.xyxy[0])
+                size = (x2 - x1) ** 2 + (y2 - y1) ** 2
+                cars.append((size, (x1, y1, x2, y2)))
+
+# Get the biggest car box
+size, corners = max(cars, key=lambda x: x[0])
+
+# Crop biggest car
+cropped_img = img.crop(corners)
+cropped_img.save("car_crop_pillow.jpg")
+
+# Search for license plates in car box and OCR all
+results = plate_model.predict(source=cropped_img)
+for r in results:
+    if r.boxes:
+        for box in r.boxes:
+            cls_name = r.names[int(box.cls[0])]
+            if cls_name == "License_Plate":
+                x1, y1, x2, y2 = map(int, box.xyxy[0])
+                lp_img = cropped_img.crop((x1, y1, x2, y2))
+                lp_img.save("license_plate.jpg")
+                lp_np = np.array(object=lp_img)
+                result = reader.readtext(image=lp_np)
+                print(result)