Files

178 lines
5.4 KiB
Python

"""
PC端简单演示脚本 - 命令行版本(无GUI,适合快速测试)
功能:
1. 加载模型并进行推理
2. 在命令行显示结果
3. 保存预测结果图像
"""
import os
import time
import torch
from torchvision import transforms
from PIL import Image, ImageDraw, ImageFont
import argparse
from model import create_model
def load_model(model_path, num_classes=10, device='cpu'):
"""加载训练好的模型"""
print(f"正在加载模型: {model_path}")
model = create_model(num_classes=num_classes, pretrained=False)
checkpoint = torch.load(model_path, map_location=device)
if 'model_state_dict' in checkpoint:
model.load_state_dict(checkpoint['model_state_dict'])
else:
model.load_state_dict(checkpoint)
model = model.to(device)
model.eval()
print("模型加载完成\n")
return model
def predict_and_save(model, image_path, transform, device, output_path, class_names):
"""预测并保存结果图像"""
# 加载图像
image = Image.open(image_path).convert('RGB')
image_tensor = transform(image).unsqueeze(0).to(device)
# 推理
with torch.no_grad():
start_time = time.time()
output = model(image_tensor)
inference_time = (time.time() - start_time) * 1000 # ms
probabilities = torch.softmax(output, dim=1)
pred_class = torch.argmax(probabilities, dim=1).item()
confidence = probabilities[0][pred_class].item()
# 在图像上添加预测结果
result_image = image.copy()
result_image = result_image.resize((800, 800))
draw = ImageDraw.Draw(result_image)
# 添加文字
text = f"Predicted: {class_names[pred_class]}\n"
text += f"Confidence: {confidence*100:.2f}%\n"
text += f"Time: {inference_time:.2f}ms"
# 使用默认字体
try:
font = ImageFont.truetype("arial.ttf", 40)
except:
font = ImageFont.load_default()
# 绘制文字背景
draw.rectangle([(10, 10), (600, 150)], fill='white', outline='black', width=3)
draw.text((20, 20), text, fill='black', font=font)
# 保存结果
result_image.save(output_path)
return pred_class, confidence, inference_time
def main():
parser = argparse.ArgumentParser(description='PC端简单演示')
parser.add_argument('--model_path', type=str,
default='data/models/best_model.pth')
parser.add_argument('--data_dir', type=str,
default='data/processed/val')
parser.add_argument('--output_dir', type=str,
default='demo_results')
parser.add_argument('--num_samples', type=int, default=20,
help='演示图像数量')
args = parser.parse_args()
# 创建输出目录
os.makedirs(args.output_dir, exist_ok=True)
# 设置设备
device = torch.device('cpu')
# 数据预处理
transform = transforms.Compose([
transforms.Resize((224, 224)),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])
])
# 类别名称
class_names = [f'class_{i}' for i in range(10)]
print("="*60)
print("PC端光图像识别演示系统")
print("="*60)
print(f"模型: MobileNetV2")
print(f"设备: {device}\n")
# 加载模型
model = load_model(args.model_path, num_classes=10, device=device)
# 获取测试图像
image_paths = []
true_labels = []
for class_idx, class_name in enumerate(class_names):
class_dir = os.path.join(args.data_dir, class_name)
if os.path.exists(class_dir):
for img_name in sorted(os.listdir(class_dir))[:2]: # 每类取2张
if img_name.endswith(('.jpg', '.png', '.jpeg')):
image_paths.append(os.path.join(class_dir, img_name))
true_labels.append(class_idx)
# 限制数量
image_paths = image_paths[:args.num_samples]
true_labels = true_labels[:args.num_samples]
print(f"找到 {len(image_paths)} 张测试图像\n")
print("开始推理...\n")
correct = 0
total = 0
total_time = 0
for idx, (img_path, true_label) in enumerate(zip(image_paths, true_labels)):
# 输出文件名
output_name = f"result_{idx+1:03d}.jpg"
output_path = os.path.join(args.output_dir, output_name)
# 预测
pred_class, confidence, inf_time = predict_and_save(
model, img_path, transform, device, output_path, class_names
)
# 统计
is_correct = (pred_class == true_label)
if is_correct:
correct += 1
total += 1
total_time += inf_time
# 打印结果
status = "[正确]" if is_correct else "[错误]"
print(f"[{idx+1}/{len(image_paths)}] {os.path.basename(img_path)}")
print(f" 真实: {class_names[true_label]}")
print(f" 预测: {class_names[pred_class]} {status}")
print(f" 置信度: {confidence*100:.2f}%")
print(f" 推理时间: {inf_time:.2f} ms")
print(f" 已保存: {output_name}")
print()
# 最终统计
print("="*60)
print("演示完成!")
print(f"总体准确率: {correct}/{total} = {correct/total*100:.2f}%")
print(f"平均推理时间: {total_time/total:.2f} ms")
print(f"吞吐量: {1000/(total_time/total):.2f} FPS")
print(f"结果已保存到: {args.output_dir}/")
print("="*60)
if __name__ == '__main__':
main()