import os
import random
import numpy as np
import matplotlib.pyplot as plt
from PIL import Image, ImageDraw, ImageFont
import torch
import torch.nn as nn
import torch.optim as optim
from torch.utils.data import Dataset, DataLoader
from torchvision import transforms
解决OMP冲突
os.environ["KMP_DUPLICATE_LIB_OK"] = "TRUE"
device = torch.device('cpu')
配置参数
CHARS = ['一', '二', '三', '十', '人', '口', '手', '日', '月', '水']
TRAIN_NUM = 200
TEST_NUM = 50
IMG_SIZE = 64
DATA_SAVE_DIR = 'hanzi_data'
BATCH_SIZE = 32
EPOCHS = 30
LEARNING_RATE = 0.005
-------------------------- 关键改进：使用PIL默认字体生成汉字（无需额外安装） --------------------------
class HanziDatasetGenerator:
def init(self):
不依赖系统字体，使用PIL的默认字体+手动调整位置确保汉字显示
self.font = ImageFont.load_default()
print("提示：使用默认字体生成汉字（可能显示较简单，但能保证运行）")
def _generate_single_img(self, char):
"""生成简单但可区分的汉字图像"""
img = Image.new('L', (IMG_SIZE, IMG_SIZE), color=255) # 白底
draw = ImageDraw.Draw(img)
# 针对默认字体调整位置（确保汉字完整显示）
# 不同汉字手动调整偏移量，保证特征差异
char_offsets = {'一': (5, 25), '二': (5, 15), '三': (5, 10),'十': (20, 15), '人': (10, 20), '口': (15, 15),'手': (5, 10), '日': (15, 15), '月': (10, 15), '水': (5, 10)
}
x, y = char_offsets[char]# 固定较大字体尺寸，确保笔画清晰
font_size = 40
try:# 再次尝试系统字体，失败则用默认font = ImageFont.truetype('simsun.ttc', size=font_size) # 尝试宋体draw.text((x, y), char, font=font, fill=0, stroke_width=2)
except:# 用默认字体，手动加粗笔画确保可区分draw.text((x, y), char, font=self.font, fill=0, stroke_width=3)# 二次绘制增强笔画（避免默认字体太细）draw.text((x+1, y), char, font=self.font, fill=0, stroke_width=2)# 轻微旋转增加差异
rotation = random.randint(-10, 10)
img = img.rotate(rotation, expand=False, fillcolor=255)return img
def generate_dataset(self):
"""生成数据集目录和图片"""
if os.path.exists(DATA_SAVE_DIR):
for root, dirs, files in os.walk(DATA_SAVE_DIR, topdown=False):
for f in files: os.remove(os.path.join(root, f))
for d in dirs: os.rmdir(os.path.join(root, d))
os.rmdir(DATA_SAVE_DIR)
# 创建目录
for split in ['train', 'test']:for char in CHARS:os.makedirs(os.path.join(DATA_SAVE_DIR, split, char), exist_ok=True)# 生成样本
print("生成数据集...")
for char in CHARS:for i in range(TRAIN_NUM):img = self._generate_single_img(char)img.save(os.path.join(DATA_SAVE_DIR, 'train', char, f'{i}.png'))for i in range(TEST_NUM):img = self._generate_single_img(char)img.save(os.path.join(DATA_SAVE_DIR, 'test', char, f'{i}.png'))
print(f"数据集生成完成：{os.path.abspath(DATA_SAVE_DIR)}")
-------------------------- 数据集加载 --------------------------
class HanziDataset(Dataset):
def init(self, split='train'):
self.split = split
self.data_dir = os.path.join(DATA_SAVE_DIR, split)
self.char_list = CHARS
self.char2idx = {c:i for i,c in enumerate(self.char_list)}
self.images, self.labels = self._load_data()
self.transform = transforms.ToTensor()
def _load_data(self):
images = []
labels = []
for char in self.char_list:
char_dir = os.path.join(self.data_dir, char)
for img_name in os.listdir(char_dir):
images.append(os.path.join(char_dir, img_name))
labels.append(self.char2idx[char])
return images, labels
def len(self):
return len(self.images)
def getitem(self, idx):
img = Image.open(self.images[idx]).convert('L')
return self.transform(img), self.labels[idx]
-------------------------- 模型 --------------------------
class FeatureCNN(nn.Module):
def init(self, num_classes=10):
super().init()
self.features = nn.Sequential(
nn.Conv2d(1, 8, kernel_size=3, padding=1),
nn.ReLU(),
nn.MaxPool2d(2, 2), # 64→32
nn.Conv2d(8, 16, kernel_size=3, padding=1),nn.ReLU(),nn.MaxPool2d(2, 2) # 32→16
)
self.classifier = nn.Linear(16 * 16 * 16, num_classes)
def forward(self, x):
x = self.features(x)
x = x.view(-1, 16 * 16 * 16)
x = self.classifier(x)
return x
-------------------------- 训练与识别 --------------------------
def main():
生成数据集（关键：即使没有中文字体也能生成可区分的图像）
generator = HanziDatasetGenerator()
generator.generate_dataset()
加载数据
train_dataset = HanziDataset('train')
test_dataset = HanziDataset('test')
train_loader = DataLoader(train_dataset, batch_size=BATCH_SIZE, shuffle=True, num_workers=0)
test_loader = DataLoader(test_dataset, batch_size=BATCH_SIZE, shuffle=False, num_workers=0)
模型与优化器
model = FeatureCNN().to(device)
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(model.parameters(), lr=LEARNING_RATE)
训练
print("\n开始训练...")
best_acc = 0.0
for epoch in range(EPOCHS):
model.train()
total_loss = 0.0
for imgs, labels in train_loader:imgs, labels = imgs.to(device), labels.to(device)optimizer.zero_grad()outputs = model(imgs)loss = criterion(outputs, labels)loss.backward()optimizer.step()total_loss += loss.item() * imgs.size(0)avg_loss = total_loss / len(train_dataset)# 测试
model.eval()
correct = 0
total = 0
with torch.no_grad():for imgs, labels in test_loader:imgs, labels = imgs.to(device), labels.to(device)outputs = model(imgs)_, preds = torch.max(outputs, 1)total += labels.size(0)correct += (preds == labels).sum().item()acc = 100 * correct / total
print(f"轮次{epoch+1:2d} | 损失：{avg_loss:.4f} | 准确率：{acc:.2f}%")if acc > best_acc:best_acc = acctorch.save(model.state_dict(), 'best_model.pth')if acc >= 85:print(f"达标！准确率：{acc:.2f}%")break
识别
model.load_state_dict(torch.load('best_model.pth'))
print(f"\n最优准确率：{best_acc:.2f}%")
while True:
path = input("\n输入图片路径（q退出）：")
if path.lower() == 'q':
break
if not os.path.exists(path):
print("路径错误")
continue
try:img = Image.open(path).convert('L').resize((64,64))img_tensor = transforms.ToTensor()(img).unsqueeze(0).to(device)with torch.no_grad():output = model(img_tensor)pred_char = CHARS[torch.argmax(output).item()]confidence = torch.softmax(output, dim=1).max().item() * 100print(f"识别结果：{pred_char} | 可信度：{confidence:.2f}%")
except Exception as e:print(f"错误：{e}")
)
