File size: 2,301 Bytes
5d670ae
da9f292
5d670ae
da9f292
 
5d670ae
da9f292
acddb2f
005d086
 
 
 
 
 
acddb2f
 
 
29533d7
acddb2f
 
3ca006e
29533d7
acddb2f
29533d7
acddb2f
363a646
65ed4c1
363a646
acddb2f
da9f292
acddb2f
3ca006e
 
acddb2f
ee1d691
da9f292
5d670ae
29533d7
 
7a8e198
29533d7
 
 
 
005d086
acddb2f
477d4fe
7a8e198
 
acddb2f
103f82b
ee1d691
acddb2f
103f82b
7a8e198
ee1d691
29533d7
7a8e198
 
 
 
 
 
 
29533d7
acddb2f
8fe1b94
65ed4c1
2132698
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
import easyocr
import numpy as np
import cv2
import re

reader = easyocr.Reader(['en'], gpu=False)

def enhance_image(img):
    max_dim = 1000
    height, width = img.shape[:2]
    if max(height, width) > max_dim:
        scale = max_dim / max(height, width)
        img = cv2.resize(img, None, fx=scale, fy=scale, interpolation=cv2.INTER_AREA)

    gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
    gray = cv2.fastNlMeansDenoising(gray, h=15)

    kernel = np.array([[0, -1, 0], [-1, 5, -1], [0, -1, 0]])
    sharp = cv2.filter2D(gray, -1, kernel)

    clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8, 8))
    enhanced = clahe.apply(sharp)

    return enhanced

def extract_weight_from_image(pil_img):
    try:
        img = np.array(pil_img)
        enhanced = enhance_image(img)

        results = reader.readtext(enhanced)
        print("DEBUG OCR RESULTS:", results)

        ocr_texts = [text for _, text, _ in results]
        weight_candidates = []

        for _, text, conf in results:
            cleaned = text.lower().strip()

            # Fix common OCR errors
            cleaned = cleaned.replace(",", ".")
            cleaned = cleaned.replace("o", "0").replace("O", "0")
            cleaned = cleaned.replace("s", "5").replace("S", "5")
            cleaned = cleaned.replace("g", "9").replace("G", "6")
            cleaned = cleaned.replace("kg", "").replace("kgs", "")
            cleaned = re.sub(r"[^\d\.]", "", cleaned)

            # Match weights like: 58.8, 75.02, 97.2, 102.34, etc.
            if re.fullmatch(r"\d{2,4}(\.\d{1,3})?", cleaned):
                weight_candidates.append((cleaned, conf))

        if not weight_candidates:
            return "Not detected", 0.0, "\n".join(ocr_texts)

        # Pick the highest confidence result
        best_weight, best_conf = sorted(weight_candidates, key=lambda x: -x[1])[0]

        # Remove unnecessary leading zeros
        if "." in best_weight:
            parts = best_weight.split(".")
            parts[0] = parts[0].lstrip("0") or "0"
            best_weight = ".".join(parts)
        else:
            best_weight = best_weight.lstrip("0") or "0"

        return best_weight, round(best_conf * 100, 2), "\n".join(ocr_texts)

    except Exception as e:
        return f"Error: {str(e)}", 0.0, "OCR failed"