图像验证码 Base64 编码最佳实践

在 Base64 中正确编码验证码图像是可靠解决问题的第一步。错误的编码意味着错误的答案或错误。本指南介绍了正确的方法。

Base64 提交格式

CaptchaAI 通过 method=base64 参数接受 base64 格式的图像验证码：

import requests
import base64
import os


def submit_image_captcha(image_base64):
    """Submit base64-encoded image to CaptchaAI."""
    resp = requests.post("https://ocr.captchaai.com/in.php", data={
        "key": os.environ["CAPTCHAAI_API_KEY"],
        "method": "base64",
        "body": image_base64,
        "json": 1,
    }, timeout=30)
    return resp.json()

从文件编码

# from_file.py
import base64


def encode_from_file(filepath):
    """Read an image file and return base64 string."""
    with open(filepath, "rb") as f:
        raw = f.read()
    return base64.b64encode(raw).decode("ascii")


# Usage
b64 = encode_from_file("captcha.png")
print(f"Encoded length: {len(b64)} chars")

从 URL 编码

# from_url.py
import requests
import base64


def encode_from_url(image_url):
    """Download image and return base64 string."""
    resp = requests.get(image_url, timeout=15)
    resp.raise_for_status()

    # Verify it's actually an image
    content_type = resp.headers.get("Content-Type", "")
    if not content_type.startswith("image/"):
        raise ValueError(f"Not an image: {content_type}")

    return base64.b64encode(resp.content).decode("ascii")


# Usage
b64 = encode_from_url("https://example.com/captcha.png")

从 Selenium 屏幕截图进行编码

# from_selenium.py
import base64
from selenium.webdriver.common.by import By


def encode_from_element(driver, selector):
    """Screenshot a specific element and return base64."""
    element = driver.find_element(By.CSS_SELECTOR, selector)
    screenshot_b64 = element.screenshot_as_base64
    return screenshot_b64


def encode_from_page_crop(driver, selector):
    """Crop a specific region from the page screenshot."""
    from PIL import Image
    import io

    element = driver.find_element(By.CSS_SELECTOR, selector)
    location = element.location
    size = element.size

    # Full page screenshot
    png = driver.get_screenshot_as_png()
    img = Image.open(io.BytesIO(png))

    # Crop to element bounds
    left = location["x"]
    top = location["y"]
    right = left + size["width"]
    bottom = top + size["height"]
    cropped = img.crop((left, top, right, bottom))

    # Encode
    buffer = io.BytesIO()
    cropped.save(buffer, format="PNG")
    return base64.b64encode(buffer.getvalue()).decode("ascii")

常见的编码错误

错误 1：包含数据 URI 前缀

# WRONG — includes data URI prefix
bad = "data:image/png;base64,iVBORw0KGgo..."

# RIGHT — raw base64 only
good = "iVBORw0KGgo..."

# Fix: Strip the prefix
def clean_base64(b64_string):
    if "," in b64_string:
        return b64_string.split(",", 1)[1]
    return b64_string

错误2：双重编码

# WRONG — encoding an already-encoded string
already_b64 = element.screenshot_as_base64
double_encoded = base64.b64encode(already_b64.encode()).decode()  # BAD

# RIGHT — use as-is
correct = element.screenshot_as_base64  # Already base64

错误 3：编码文本而不是字节

# WRONG — reading as text
with open("captcha.png", "r") as f:  # Text mode
    content = f.read()  # Corrupted binary data

# RIGHT — reading as bytes
with open("captcha.png", "rb") as f:  # Binary mode
    content = f.read()
encoded = base64.b64encode(content).decode("ascii")

提交前验证

# validate.py
import base64
import io


def validate_captcha_image(b64_string):
    """Validate base64 image before submitting to CaptchaAI."""
    errors = []

    # Check for data URI prefix
    if b64_string.startswith("data:"):
        errors.append("Contains data URI prefix — strip it")
        b64_string = b64_string.split(",", 1)[1]

    # Try decoding
    try:
        decoded = base64.b64decode(b64_string)
    except Exception as e:
        return {"valid": False, "errors": [f"Invalid base64: {e}"]}

    # Check size
    size_kb = len(decoded) / 1024
    if size_kb < 1:
        errors.append(f"Image too small ({size_kb:.1f} KB) — likely corrupt")
    if size_kb > 500:
        errors.append(f"Image large ({size_kb:.1f} KB) — consider resizing")

    # Check image format
    if decoded[:8] == b'\x89PNG\r\n\x1a\n':
        fmt = "PNG"
    elif decoded[:3] == b'\xff\xd8\xff':
        fmt = "JPEG"
    elif decoded[:4] == b'GIF8':
        fmt = "GIF"
    elif decoded[:4] == b'RIFF':
        fmt = "WEBP"
    else:
        errors.append("Unknown image format")
        fmt = "unknown"

    return {
        "valid": len(errors) == 0,
        "format": fmt,
        "size_kb": round(size_kb, 1),
        "errors": errors,
    }


# Usage
result = validate_captcha_image(b64_string)
if not result["valid"]:
    print(f"Issues: {result['errors']}")
else:
    print(f"Valid {result['format']}, {result['size_kb']} KB")

图像格式建议

建议： 使用 PNG 作为文本验证码。无损压缩保留了字符边缘，提高了求解精度。

故障排除

常问问题

CaptchaAI 接受的最大图像尺寸是多少？

Base64 正文为 600 KB。在编码之前调整大屏幕截图的大小。

我应该使用 PNG 还是 JPEG 作为文本验证码？

巴布亚新几内亚。 JPEG 压缩会模糊字符边缘，从而降低准确性。 PNG 保留精确的像素。

我可以提交 SVG 图像吗？

不。首先使用 Pillow 或 cairosvg 等库将 SVG 转换为 PNG。

相关指南

• 提高 OCR 验证码准确性
• 验证码图像预处理

正确编码验证码 -以 CaptchaAI 开头.