AiDOPWarehouse/doc/_bid_make_seals.py

# -*- coding: utf-8 -*-
"""从授权书扫描件里抠出单位公章、法人印鉴与手写签名，做成透明 PNG。

扫描件底色偏黄且有网点，直接按「非白即墨」抠图会留下一圈脏边，故按通道差
分别处理：红章看 R 与 G/B 的差，签名看整体明度，再各自把弱信号压到全透明。
"""
import io
import os
import shutil
import sys

import numpy as np
from PIL import Image

sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding="utf-8")

HERE = os.path.dirname(os.path.abspath(__file__))
SRC = os.path.join(HERE, "_bid_assets", "seal_src", "p1_img1.png")
OUT = os.path.join(HERE, "_bid_assets", "seal")
os.makedirs(OUT, exist_ok=True)

# 由 _tmp_seal_locate.py 定位得到（含少量留白）
BOX = {
    "gongzhang": (843, 2235, 1310, 2739),    # 圆形单位公章
    "yinjian": (1578, 2347, 1821, 2603),     # 方形法定代表人印鉴
    "signature": (215, 2330, 520, 2520),     # 手写签名「刘涛」
}

# 扫描件是整幅 A4，但像素为 2140×3264，纵向被拉长了 7.8%（正圆公章的墨迹
# 框 440×479 与正方印鉴的 217×231 都印证了这一点）。按 A4 宽高比压回，
# 公章才是正圆、印鉴才是正方，后面按毫米摆放时才不会变形。
SQUASH = 2140 * 297 / 210 / 3264


def stats(a, name):
    r, g, b = a[..., 0], a[..., 1], a[..., 2]
    print(f"  {name}: 背景估计 R{np.percentile(r, 95):.0f} "
          f"G{np.percentile(g, 95):.0f} B{np.percentile(b, 95):.0f}　"
          f"最暗 R{r.min()} G{g.min()} B{b.min()}")


def red_seal(crop, name):
    """红章：以 G/B 通道的下沉量作为墨量，颜色统一成正红。"""
    a = np.asarray(crop, dtype=np.float32)
    stats(a, name)
    gb = (a[..., 1] + a[..., 2]) / 2
    bg = np.percentile(gb, 96)                      # 纸张底色
    ink = np.clip((bg - gb) / max(bg - 55.0, 1.0), 0, 1)
    ink[ink < 0.14] = 0                             # 压掉网点与黄底
    alpha = (ink ** 0.85 * 255).astype(np.uint8)
    rgb = np.zeros(a.shape, dtype=np.uint8)
    rgb[..., 0], rgb[..., 1], rgb[..., 2] = 200, 26, 26
    out = Image.fromarray(np.dstack([rgb, alpha]), "RGBA")
    return out


def ink_sign(crop, name):
    """签名：黑色笔迹，按明度取墨量。"""
    a = np.asarray(crop.convert("L"), dtype=np.float32)
    print(f"  {name}: 明度 背景{np.percentile(a, 95):.0f} 最暗{a.min():.0f}")
    bg = np.percentile(a, 95)
    ink = np.clip((bg - a) / max(bg - 60.0, 1.0), 0, 1)
    ink[ink < 0.18] = 0
    alpha = (ink ** 0.9 * 255).astype(np.uint8)
    rgb = np.zeros(a.shape + (3,), dtype=np.uint8)
    rgb[..., 0], rgb[..., 1], rgb[..., 2] = 20, 24, 40   # 蓝黑墨
    return Image.fromarray(np.dstack([rgb, alpha]), "RGBA")


def despeckle(im, min_px=40):
    """去掉扫描噪点：丢弃面积过小的连通块（用逐行并查集做 8 邻域标记）。"""
    a = np.asarray(im).copy()
    mask = a[..., 3] > 8
    h, w = mask.shape
    lab = np.zeros((h, w), dtype=np.int32)
    parent = [0]

    def find(x):
        while parent[x] != x:
            parent[x] = parent[parent[x]]
            x = parent[x]
        return x

    def union(x, y):
        rx, ry = find(x), find(y)
        if rx != ry:
            parent[max(rx, ry)] = min(rx, ry)

    for y in range(h):
        for x in range(w):
            if not mask[y, x]:
                continue
            nb = [lab[y + dy, x + dx]
                  for dy, dx in ((-1, -1), (-1, 0), (-1, 1), (0, -1))
                  if 0 <= y + dy < h and 0 <= x + dx < w and lab[y + dy, x + dx]]
            if nb:
                lab[y, x] = min(nb)
                for n in nb:
                    union(lab[y, x], n)
            else:
                parent.append(len(parent))
                lab[y, x] = len(parent) - 1

    roots = np.zeros(len(parent), dtype=np.int32)
    for i in range(1, len(parent)):
        roots[i] = find(i)
    flat = roots[lab]
    ids, cnt = np.unique(flat[flat > 0], return_counts=True)
    drop = set(ids[cnt < min_px].tolist())
    if drop:
        kill = np.isin(flat, list(drop))
        a[..., 3][kill] = 0
        print(f"    去斑：丢弃 {len(drop)} 个小块，共 {int(kill.sum())} 像素")
    return Image.fromarray(a, "RGBA")


def trim(im, pad=6):
    a = np.asarray(im)[..., 3]
    ys, xs = np.nonzero(a > 8)
    if not len(ys):
        return im
    return im.crop((max(0, xs.min() - pad), max(0, ys.min() - pad),
                    min(im.width, xs.max() + pad), min(im.height, ys.max() + pad)))


src = Image.open(SRC).convert("RGB")
print(f"源图 {src.size}　纵向压回系数 {SQUASH:.4f}")
for name, box in BOX.items():
    crop = src.crop(box)
    crop = crop.resize((crop.width, round(crop.height * SQUASH)), Image.LANCZOS)
    im = ink_sign(crop, name) if name == "signature" else red_seal(crop, name)
    im = trim(despeckle(im, 60 if name == "signature" else 30))
    path = os.path.join(OUT, f"{name}.png")
    im.save(path)
    a = np.asarray(im)[..., 3]
    ys, xs = np.nonzero(a > 8)
    bw, bh = xs.max() - xs.min() + 1, ys.max() - ys.min() + 1
    print(f"  → {name}.png  {im.size}  墨迹 {bw}×{bh}（长宽比 {bw / bh:.3f}）"
          f"  非透明像素 {int((a > 8).sum()):,}")

# 拼一张白底预览，便于确认抠图干净
prev = Image.new("RGB", (980, 420), "white")
x = 20
for name in BOX:
    im = Image.open(os.path.join(OUT, f"{name}.png"))
    im.thumbnail((300, 300))
    prev.paste(im, (x, 60), im)
    x += im.width + 30
prev.save(os.path.join(OUT, "_preview.png"))
print(f"\n预览：{os.path.join(OUT, '_preview.png')}")

# 另存一份到招投标材料目录，供另行取用
SHARE = os.path.join(r"C:\Users\skygu\OneDrive\Projects\AIDOP\项目\项目招投标",
                     "相关准备材料", "印鉴素材（透明底）")
CN = {"gongzhang": "单位公章", "yinjian": "法定代表人印鉴", "signature": "法定代表人签名"}
if os.path.isdir(os.path.dirname(SHARE)):
    os.makedirs(SHARE, exist_ok=True)
    print(f"\n素材另存至 {SHARE}")
    for name, cn in CN.items():
        shutil.copy2(os.path.join(OUT, f"{name}.png"),
                     os.path.join(SHARE, f"{cn}-透明底.png"))
        print(f"  {cn}-透明底.png")
    shutil.copy2(os.path.join(OUT, "_preview.png"),
                 os.path.join(SHARE, "三枚印鉴预览.png"))
    print("  三枚印鉴预览.png")