Sanjayraju30 commited on
Commit
3ca006e
·
verified ·
1 Parent(s): acddb2f

Update ocr_engine.py

Browse files
Files changed (1) hide show
  1. ocr_engine.py +8 -10
ocr_engine.py CHANGED
@@ -3,26 +3,23 @@ import numpy as np
3
  import cv2
4
  import re
5
 
 
6
  reader = easyocr.Reader(['en'], gpu=False)
7
 
8
  def enhance_image(img):
9
- # Resize to make text clearer
10
  img = cv2.resize(img, None, fx=4, fy=4, interpolation=cv2.INTER_CUBIC)
11
-
12
- # Convert to gray
13
  gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
14
 
15
  # Denoise
16
  gray = cv2.fastNlMeansDenoising(gray, h=15)
17
 
18
  # Sharpen
19
- kernel = np.array([[0, -1, 0],
20
- [-1, 5,-1],
21
- [0, -1, 0]])
22
  sharp = cv2.filter2D(gray, -1, kernel)
23
 
24
- # Contrast enhancement (CLAHE)
25
- clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8,8))
26
  contrast = clahe.apply(sharp)
27
 
28
  return contrast
@@ -33,12 +30,13 @@ def extract_weight_from_image(pil_img):
33
  enhanced = enhance_image(img)
34
 
35
  results = reader.readtext(enhanced)
 
 
36
  ocr_texts = [text for _, text, _ in results]
37
  weight_candidates = []
38
 
39
  for _, text, conf in results:
40
- cleaned = text.lower()
41
- cleaned = cleaned.replace("kg", "").replace("kgs", "")
42
  cleaned = cleaned.replace("o", "0").replace("s", "5").replace("g", "9")
43
  cleaned = re.sub(r"[^\d\.]", "", cleaned)
44
 
 
3
  import cv2
4
  import re
5
 
6
+ # Load OCR model once
7
  reader = easyocr.Reader(['en'], gpu=False)
8
 
9
  def enhance_image(img):
10
+ # Enlarge and convert to grayscale
11
  img = cv2.resize(img, None, fx=4, fy=4, interpolation=cv2.INTER_CUBIC)
 
 
12
  gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
13
 
14
  # Denoise
15
  gray = cv2.fastNlMeansDenoising(gray, h=15)
16
 
17
  # Sharpen
18
+ kernel = np.array([[0, -1, 0], [-1, 5, -1], [0, -1, 0]])
 
 
19
  sharp = cv2.filter2D(gray, -1, kernel)
20
 
21
+ # Contrast enhance
22
+ clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8, 8))
23
  contrast = clahe.apply(sharp)
24
 
25
  return contrast
 
30
  enhanced = enhance_image(img)
31
 
32
  results = reader.readtext(enhanced)
33
+ print("DEBUG OCR RESULTS:", results)
34
+
35
  ocr_texts = [text for _, text, _ in results]
36
  weight_candidates = []
37
 
38
  for _, text, conf in results:
39
+ cleaned = text.lower().replace("kg", "").replace("kgs", "")
 
40
  cleaned = cleaned.replace("o", "0").replace("s", "5").replace("g", "9")
41
  cleaned = re.sub(r"[^\d\.]", "", cleaned)
42