vault backup: 2026-11-07 16:33:44
This commit is contained in:
cheney
@@ -0,0 +1,2 @@
|
||||
docker run --rm -v "$(pwd):/app" paddleocr/paddleocr:2.6-cpu-latest \
|
||||
python /app/paddleocr-hex.py
|
||||
@@ -0,0 +1,388 @@
|
||||
# 现成的 Tesseract Docker 镜像及使用方案
|
||||
|
||||
确实有多个现成的 Tesseract Docker 镜像可以直接使用，无需自己构建。以下是基于现有镜像的简单 Docker 命令方案：
|
||||
|
||||
## 1. 可用的现成 Tesseract Docker 镜像
|
||||
|
||||
### 1.1 官方和社区镜像
|
||||
```bash
|
||||
# 1. 官方 Tesseract OCR 镜像
|
||||
docker pull tesseractshadow/tesseract4re:latest
|
||||
|
||||
# 2. 另一个流行的社区镜像
|
||||
docker pull jitesh/tesseract-ocr:latest
|
||||
|
||||
# 3. 轻量级 Alpine 版本
|
||||
docker pull alexbrah/tesseract-ocr:alpine
|
||||
|
||||
# 4. 包含多种语言的镜像
|
||||
docker pull napcs/tesseract-ocr:latest
|
||||
```
|
||||
|
||||
## 2. 最简单的 Docker 使用方案
|
||||
|
||||
### 2.1 基础使用命令
|
||||
```bash
|
||||
# 查看镜像中的语言包
|
||||
docker run --rm tesseractshadow/tesseract4re tesseract --list-langs
|
||||
|
||||
# 直接运行单次识别
|
||||
docker run --rm -v "$(pwd):/data" tesseractshadow/tesseract4re \
|
||||
tesseract /data/input.png /data/output
|
||||
|
||||
# 中文
|
||||
docker run --rm -v "$(pwd):/data" tesseractshadow/tesseract4re \
|
||||
tesseract /data/input.png /data/output -l chi_sim
|
||||
|
||||
# 中文简体 + 英文
|
||||
docker run --rm -v "$(pwd):/data" tesseractshadow/tesseract4re \
|
||||
tesseract /data/input.png /data/output -l chi_sim+eng
|
||||
|
||||
# Hex
|
||||
docker run --rm \
|
||||
-v "$(pwd):/data" \
|
||||
tesseractshadow/tesseract4re \
|
||||
tesseract /data/hex1.png /data/output \
|
||||
-c tessedit_char_whitelist=0123456789ABCDEFabcdef \
|
||||
--psm 6
|
||||
|
||||
|
||||
# Base64
|
||||
docker run --rm \
|
||||
-v "$(pwd):/data" \
|
||||
tesseractshadow/tesseract4re \
|
||||
tesseract /data/hex1.png /data/output \
|
||||
-c tessedit_char_whitelist=ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/= \
|
||||
--psm 6
|
||||
|
||||
```
|
||||
|
||||
### 2.2 针对 HEX 和 Base64 的优化使用
|
||||
```bash
|
||||
# 创建专用配置目录
|
||||
mkdir -p tesseract-config
|
||||
mkdir -p test-images
|
||||
mkdir -p output
|
||||
|
||||
# 创建 HEX 配置文件
|
||||
cat > tesseract-config/hex.config << 'EOF'
|
||||
tessedit_char_whitelist 0123456789ABCDEFabcdef
|
||||
tessedit_pageseg_mode 6
|
||||
tessedit_ocr_engine_mode 3
|
||||
EOF
|
||||
|
||||
# 创建 Base64 配置文件
|
||||
cat > tesseract-config/base64.config << 'EOF'
|
||||
tessedit_char_whitelist ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/=
|
||||
tessedit_pageseg_mode 6
|
||||
tessedit_ocr_engine_mode 3
|
||||
EOF
|
||||
```
|
||||
|
||||
|
||||
|
||||
## 3. 完整的验证脚本
|
||||
|
||||
### 3.1 创建测试图像
|
||||
```bash
|
||||
# create-test-images.py
|
||||
#!/usr/bin/env python3
|
||||
from PIL import Image, ImageDraw, ImageFont
|
||||
import os
|
||||
import random
|
||||
import string
|
||||
|
||||
def create_test_images():
|
||||
"""创建测试图像"""
|
||||
os.makedirs('test-images', exist_ok=True)
|
||||
|
||||
# HEX 测试图像
|
||||
hex_chars = "0123456789ABCDEF"
|
||||
for i in range(5):
|
||||
text = ''.join(random.choices(hex_chars, k=16))
|
||||
create_image(f"test-images/hex_{i}.png", text)
|
||||
|
||||
# Base64 测试图像
|
||||
base64_chars = string.ascii_letters + string.digits + "+/="
|
||||
for i in range(5):
|
||||
text = ''.join(random.choices(base64_chars, k=24))
|
||||
create_image(f"test-images/base64_{i}.png", text)
|
||||
|
||||
def create_image(filename, text):
|
||||
"""创建测试图像"""
|
||||
img = Image.new('RGB', (400, 50), color='white')
|
||||
draw = ImageDraw.Draw(img)
|
||||
|
||||
try:
|
||||
font = ImageFont.truetype("/usr/share/fonts/truetype/dejavu/DejaVuSans.ttf", 20)
|
||||
except:
|
||||
font = ImageFont.load_default()
|
||||
|
||||
draw.text((10, 15), text, fill='black', font=font)
|
||||
img.save(filename)
|
||||
print(f"Created: {filename} - {text}")
|
||||
|
||||
if __name__ == "__main__":
|
||||
create_test_images()
|
||||
```
|
||||
|
||||
### 3.2 运行测试脚本创建图像
|
||||
```bash
|
||||
# 在主机上运行 Python 脚本创建测试图像
|
||||
python3 create-test-images.py
|
||||
|
||||
# 或者使用 Docker 运行（如果主机没有 Python）
|
||||
docker run --rm -v "$(pwd):/app" python:3.9-slim \
|
||||
python3 /app/create-test-images.py
|
||||
```
|
||||
|
||||
## 4. 批量识别脚本
|
||||
|
||||
### 4.1 创建批量处理脚本
|
||||
```bash
|
||||
# batch-recognize.sh
|
||||
#!/bin/bash
|
||||
|
||||
echo "Starting batch recognition of HEX and Base64 images..."
|
||||
|
||||
# 创建输出目录
|
||||
mkdir -p output
|
||||
|
||||
# 识别 HEX 图像
|
||||
echo "Processing HEX images..."
|
||||
for img in test-images/hex_*.png; do
|
||||
if [ -f "$img" ]; then
|
||||
filename=$(basename "$img" .png)
|
||||
echo "Processing: $img"
|
||||
|
||||
docker run --rm \
|
||||
-v "$(pwd):/data" \
|
||||
tesseractshadow/tesseract4re \
|
||||
tesseract "/data/$img" "/data/output/${filename}" \
|
||||
-c tessedit_char_whitelist=0123456789ABCDEFabcdef \
|
||||
--psm 6
|
||||
fi
|
||||
done
|
||||
|
||||
# 识别 Base64 图像
|
||||
echo "Processing Base64 images..."
|
||||
for img in test-images/base64_*.png; do
|
||||
if [ -f "$img" ]; then
|
||||
filename=$(basename "$img" .png)
|
||||
echo "Processing: $img"
|
||||
|
||||
docker run --rm \
|
||||
-v "$(pwd):/data" \
|
||||
tesseractshadow/tesseract4re \
|
||||
tesseract "/data/$img" "/data/output/${filename}" \
|
||||
-c tessedit_char_whitelist=ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/= \
|
||||
--psm 6
|
||||
fi
|
||||
done
|
||||
|
||||
echo "Recognition complete! Check output directory for results."
|
||||
```
|
||||
|
||||
### 4.2 运行批量识别
|
||||
```bash
|
||||
# 给脚本执行权限
|
||||
chmod +x batch-recognize.sh
|
||||
|
||||
# 运行批量识别
|
||||
./batch-recognize.sh
|
||||
```
|
||||
|
||||
## 5. 验证和结果分析
|
||||
|
||||
### 5.1 创建结果验证脚本
|
||||
```bash
|
||||
# validate-results.py
|
||||
#!/usr/bin/env python3
|
||||
import os
|
||||
import re
|
||||
import base64
|
||||
|
||||
def read_original_text():
|
||||
"""从文件名读取原始文本（假设文件名包含原始文本）"""
|
||||
originals = {}
|
||||
for img_file in os.listdir('test-images'):
|
||||
if img_file.startswith('hex_'):
|
||||
name = img_file.replace('.png', '')
|
||||
# 从文件名提取文本（实际使用时应该从其他地方获取）
|
||||
originals[name] = "placeholder_hex"
|
||||
elif img_file.startswith('base64_'):
|
||||
name = img_file.replace('.png', '')
|
||||
originals[name] = "placeholder_base64"
|
||||
return originals
|
||||
|
||||
def validate_hex(text):
|
||||
"""验证 HEX 文本"""
|
||||
cleaned = re.sub(r'[^0-9A-Fa-f]', '', text)
|
||||
is_valid = len(cleaned) > 0 and all(c in '0123456789ABCDEFabcdef' for c in cleaned)
|
||||
return {
|
||||
'cleaned': cleaned.upper(),
|
||||
'is_valid': is_valid,
|
||||
'length': len(cleaned)
|
||||
}
|
||||
|
||||
def validate_base64(text):
|
||||
"""验证 Base64 文本"""
|
||||
cleaned = ''.join(c for c in text if c in 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/=')
|
||||
|
||||
try:
|
||||
# 尝试解码
|
||||
if '=' in cleaned:
|
||||
base64.b64decode(cleaned)
|
||||
is_valid = True
|
||||
else:
|
||||
padding = 4 - (len(cleaned) % 4)
|
||||
test_text = cleaned + '=' * padding
|
||||
base64.b64decode(test_text)
|
||||
is_valid = True
|
||||
except:
|
||||
is_valid = False
|
||||
|
||||
return {
|
||||
'cleaned': cleaned,
|
||||
'is_valid': is_valid,
|
||||
'length': len(cleaned)
|
||||
}
|
||||
|
||||
def main():
|
||||
"""主验证函数"""
|
||||
print("Validating OCR results...")
|
||||
|
||||
results = {}
|
||||
|
||||
# 读取输出文件
|
||||
for txt_file in os.listdir('output'):
|
||||
if txt_file.endswith('.txt'):
|
||||
file_path = os.path.join('output', txt_file)
|
||||
with open(file_path, 'r') as f:
|
||||
content = f.read().strip()
|
||||
|
||||
filename = txt_file.replace('.txt', '')
|
||||
|
||||
if filename.startswith('hex_'):
|
||||
validation = validate_hex(content)
|
||||
results[filename] = {
|
||||
'type': 'hex',
|
||||
'original': content,
|
||||
'validation': validation
|
||||
}
|
||||
elif filename.startswith('base64_'):
|
||||
validation = validate_base64(content)
|
||||
results[filename] = {
|
||||
'type': 'base64',
|
||||
'original': content,
|
||||
'validation': validation
|
||||
}
|
||||
|
||||
# 输出统计信息
|
||||
hex_results = [r for r in results.values() if r['type'] == 'hex']
|
||||
base64_results = [r for r in results.values() if r['type'] == 'base64']
|
||||
|
||||
print(f"\nHEX Results: {len(hex_results)} files")
|
||||
hex_valid = sum(1 for r in hex_results if r['validation']['is_valid'])
|
||||
print(f"Valid HEX: {hex_valid}/{len(hex_results)} ({hex_valid/len(hex_results)*100:.1f}%)")
|
||||
|
||||
print(f"\nBase64 Results: {len(base64_results)} files")
|
||||
base64_valid = sum(1 for r in base64_results if r['validation']['is_valid'])
|
||||
print(f"Valid Base64: {base64_valid}/{len(base64_results)} ({base64_valid/len(base64_results)*100:.1f}%)")
|
||||
|
||||
# 显示详细结果
|
||||
print("\nDetailed Results:")
|
||||
for filename, result in results.items():
|
||||
status = "✓" if result['validation']['is_valid'] else "✗"
|
||||
print(f"{status} {filename}: {result['original'][:50]}... -> {result['validation']['cleaned'][:30]}...")
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
```
|
||||
|
||||
### 5.2 运行验证
|
||||
```bash
|
||||
# 运行验证脚本
|
||||
python3 validate-results.py
|
||||
```
|
||||
|
||||
## 6. 交互式测试容器
|
||||
|
||||
### 6.1 启动交互式容器进行测试
|
||||
```bash
|
||||
# 启动交互式容器
|
||||
docker run -it --rm \
|
||||
-v "$(pwd):/data" \
|
||||
--name tesseract-test \
|
||||
tesseractshadow/tesseract4re \
|
||||
/bin/bash
|
||||
|
||||
# 在容器内手动测试
|
||||
cd /data
|
||||
tesseract test-images/hex_0.png stdout -c tessedit_char_whitelist=0123456789ABCDEFabcdef --psm 6
|
||||
tesseract test-images/base64_0.png stdout -c tessedit_char_whitelist=ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/= --psm 6
|
||||
```
|
||||
|
||||
## 7. 一键测试脚本
|
||||
|
||||
### 7.1 创建完整测试流程
|
||||
```bash
|
||||
# run-all-tests.sh
|
||||
#!/bin/bash
|
||||
|
||||
echo "=== Tesseract HEX/Base64 OCR Test ==="
|
||||
|
||||
# 1. 创建测试图像
|
||||
echo "Step 1: Creating test images..."
|
||||
python3 create-test-images.py
|
||||
|
||||
# 2. 批量识别
|
||||
echo "Step 2: Running batch OCR..."
|
||||
./batch-recognize.sh
|
||||
|
||||
# 3. 验证结果
|
||||
echo "Step 3: Validating results..."
|
||||
python3 validate-results.py
|
||||
|
||||
echo "=== Test Complete ==="
|
||||
```
|
||||
|
||||
## 8. 使用不同镜像的对比测试
|
||||
|
||||
### 8.1 多镜像测试脚本
|
||||
```bash
|
||||
# compare-images.sh
|
||||
#!/bin/bash
|
||||
|
||||
IMAGES=("tesseractshadow/tesseract4re" "jitesh/tesseract-ocr")
|
||||
|
||||
for image in "${IMAGES[@]}"; do
|
||||
echo "Testing with image: $image"
|
||||
docker pull $image
|
||||
|
||||
mkdir -p "output-${image##*/}"
|
||||
|
||||
# 测试单张图像
|
||||
docker run --rm \
|
||||
-v "$(pwd):/data" \
|
||||
$image \
|
||||
tesseract "/data/test-images/hex_0.png" stdout \
|
||||
-c tessedit_char_whitelist=0123456789ABCDEFabcdef \
|
||||
--psm 6 > "output-${image##*/}/test_result.txt"
|
||||
|
||||
echo "Result for $image:"
|
||||
cat "output-${image##*/}/test_result.txt"
|
||||
echo "---"
|
||||
done
|
||||
```
|
||||
|
||||
## 总结
|
||||
|
||||
使用现成的 Tesseract Docker 镜像非常简单：
|
||||
|
||||
1. **拉取镜像**：`docker pull tesseractshadow/tesseract4re`
|
||||
2. **运行识别**：使用 `docker run --rm -v` 挂载目录
|
||||
3. **配置优化**：通过 `-c tessedit_char_whitelist` 设置字符白名单
|
||||
4. **批量处理**：通过脚本批量处理多张图像
|
||||
|
||||
这种方法无需构建自定义镜像，直接使用现有镜像即可快速验证 Tesseract 对 HEX 和 Base64 字符的识别效果。通过字符白名单和适当的 PSM 参数，可以显著提高识别准确率。
|
||||
@@ -0,0 +1,74 @@
|
||||
在 **AI Studio** 中试用 **PaddleOCR-VL** 可以通过以下步骤实现，结合平台提供的免费算力和开源模型快速体验文档解析能力：
|
||||
|
||||
---
|
||||
|
||||
### **1. 进入AI Studio平台**
|
||||
- 访问 [百度AI Studio](https://aistudio.baidu.com/) 并登录账号（支持手机号/百度账号）。
|
||||
- 新用户可领取免费GPU算力资源（如V100或A100），用于模型推理或训练。
|
||||
|
||||
---
|
||||
|
||||
### **2. 创建项目并配置环境**
|
||||
- **新建项目**：在AI Studio控制台点击“**新建项目**”，选择“**Notebook**”类型。
|
||||
- **选择环境**：推荐使用“**PaddlePaddle 2.x**”镜像（已预装PaddleOCR依赖）。
|
||||
- **启动项目**：进入Notebook后，终端会自动加载PaddlePaddle环境。
|
||||
|
||||
---
|
||||
|
||||
### **3. 安装PaddleOCR-VL**
|
||||
在Notebook的代码单元格中执行以下命令，安装最新版PaddleOCR并加载VL模型：
|
||||
```python
|
||||
# 安装PaddleOCR（若环境未预装）
|
||||
!pip install paddlepaddle paddleocr -i https://mirror.baidu.com/pypi/simple
|
||||
|
||||
# 克隆PaddleOCR仓库（获取VL模型配置）
|
||||
!git clone https://github.com/PaddlePaddle/PaddleOCR.git
|
||||
%cd PaddleOCR
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### **4. 下载PaddleOCR-VL预训练模型**
|
||||
PaddleOCR-VL的模型权重需手动下载，可通过以下方式获取：
|
||||
- **官方GitHub**：从 [PaddleOCR-VL Release页](https://github.com/PaddlePaddle/PaddleOCR/releases) 下载模型文件（如`ch_PP-OCRv4_VL_det`、`ch_PP-OCRv4_VL_rec`）。
|
||||
- **AI Studio数据集**：搜索“PaddleOCR-VL”公共数据集，直接挂载到项目。
|
||||
|
||||
**示例代码（挂载数据集）**：
|
||||
```python
|
||||
from aistudio.datasets import load_dataset
|
||||
# 挂载PaddleOCR-VL模型（假设数据集ID为12345）
|
||||
dataset = load_dataset("12345")
|
||||
model_path = dataset.data_path + "/models/"
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### **5. 运行PaddleOCR-VL推理**
|
||||
加载模型并测试文档解析功能（支持PDF、图片等格式）：
|
||||
```python
|
||||
from paddleocr import PaddleOCR
|
||||
|
||||
# 初始化VL模型（需指定模型路径）
|
||||
ocr = PaddleOCR(
|
||||
det_model_dir=model_path + "ch_PP-OCRv4_VL_det/", # 检测模型
|
||||
rec_model_dir=model_path + "ch_PP-OCRv4_VL_rec/", # 识别模型
|
||||
use_angle_cls=True,
|
||||
lang="ch", # 中文文档
|
||||
use_gpu=True # 启用GPU加速
|
||||
)
|
||||
|
||||
# 测试单张图片
|
||||
img_path = "test_doc.jpg"
|
||||
result = ocr.ocr(img_path, cls=True)
|
||||
|
||||
# 打印结果
|
||||
for line in result:
|
||||
print(line)
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### **6. 高级功能（可选）**
|
||||
- **批量处理**：通过循环调用`ocr.ocr()`解析多张图片或PDF。
|
||||
- **可视化**：使用`PaddleOCR`内置的`draw_ocr`函数生成标注结果图。
|
||||
- **自定义训练**：参考[PaddleOCR训练教程](https://github.com/PaddlePaddle/PaddleOCR/blob/release/2.7/doc/doc_ch/training
|
||||
Reference in New Issue
Block a user