File size: 3,906 Bytes
b42373a | 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 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 | import cv2
import numpy as np
import os
import json
CONFIG_PATH = "config.json"
def load_config():
if os.path.exists(CONFIG_PATH):
with open(CONFIG_PATH) as f:
return json.load(f)
return {
"preprocessing": {
"clahe_clip": 3.0,
"clahe_grid": 8,
"bilateral_d": 9,
"bilateral_sigma": 75,
"adaptive_blocksize": 21,
"adaptive_c": 8
}
}
def correct_rotation(img):
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) if len(img.shape) == 3 else img
edges = cv2.Canny(gray, 50, 150)
lines = cv2.HoughLinesP(edges, 1, np.pi/180, 50, minLineLength=30, maxLineGap=10)
if lines is None:
return img
angles = [np.degrees(np.arctan2(l[0][3]-l[0][1], l[0][2]-l[0][0])) for l in lines]
if abs(np.median(angles)) > 45:
img = cv2.rotate(img, cv2.ROTATE_90_CLOCKWISE)
return img
def preprocess_chassis(image_path, save_comparison=False, output_dir="results"):
config = load_config()
p = config["preprocessing"]
img = cv2.imread(image_path)
if img is None:
raise ValueError(f"Could not load image: {image_path}")
original = img.copy()
img = correct_rotation(img)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
clahe = cv2.createCLAHE(
clipLimit=p["clahe_clip"],
tileGridSize=(p["clahe_grid"], p["clahe_grid"])
)
enhanced = clahe.apply(gray)
v0 = cv2.cvtColor(enhanced, cv2.COLOR_GRAY2BGR)
filtered = cv2.bilateralFilter(enhanced, p["bilateral_d"],
p["bilateral_sigma"], p["bilateral_sigma"])
v1 = cv2.cvtColor(filtered, cv2.COLOR_GRAY2BGR)
_, otsu = cv2.threshold(filtered, 0, 255,
cv2.THRESH_BINARY + cv2.THRESH_OTSU)
v2 = cv2.cvtColor(otsu, cv2.COLOR_GRAY2BGR)
bs = p["adaptive_blocksize"]
bs = bs if bs % 2 == 1 else bs + 1
adaptive = cv2.adaptiveThreshold(
filtered, 255,
cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
cv2.THRESH_BINARY,
blockSize=bs, C=p["adaptive_c"]
)
v3 = cv2.cvtColor(adaptive, cv2.COLOR_GRAY2BGR)
pad = p.get("padding", 20)
def add_padding(im):
return cv2.copyMakeBorder(im, pad, pad, pad, pad,
cv2.BORDER_CONSTANT, value=(255, 255, 255))
variations = [add_padding(v) for v in [v0, v1, v2, v3]]
if save_comparison:
os.makedirs(output_dir, exist_ok=True)
fname = os.path.splitext(os.path.basename(image_path))[0]
h = 200
def resize_h(im, height):
r = height / im.shape[0]
return cv2.resize(im, (int(im.shape[1] * r), height))
def add_label(im, label):
out = im.copy() if len(im.shape) == 3 else cv2.cvtColor(im, cv2.COLOR_GRAY2BGR)
cv2.putText(out, label, (10, 25), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0, 255, 0), 2)
return out
def to_bgr(p):
return p if len(p.shape) == 3 else cv2.cvtColor(p, cv2.COLOR_GRAY2BGR)
panels = [add_label(resize_h(cv2.cvtColor(original, cv2.COLOR_BGR2GRAY), h), "Original")]
labels = ["CLAHE", "Bilateral", "Otsu", "Adaptive"]
for i, var in enumerate(variations):
g = cv2.cvtColor(var, cv2.COLOR_BGR2GRAY)
panels.append(add_label(resize_h(g, h), labels[i]))
comparison = np.hstack([to_bgr(p) for p in panels])
cv2.imwrite(os.path.join(output_dir, f"{fname}_comparison.jpg"), comparison)
print(f" Saved comparison -> {output_dir}/{fname}_comparison.jpg")
return variations
if __name__ == "__main__":
import sys
if len(sys.argv) < 2:
print("Usage: python preprocess.py <image_path>")
else:
variations = preprocess_chassis(sys.argv[1], save_comparison=True)
print(f"Generated {len(variations)} variations — check results/") |