|
|
@@ -3,6 +3,7 @@ import pandas as pd
|
|
|
import numpy as np
|
|
|
from typing import List, Dict
|
|
|
import shutil
|
|
|
+from datetime import datetime
|
|
|
|
|
|
class PerformanceAnalyzer:
|
|
|
def __init__(self, result_dir: str):
|
|
|
@@ -14,11 +15,86 @@ class PerformanceAnalyzer:
|
|
|
# 确保目录存在
|
|
|
os.makedirs(self.true_targets_dir, exist_ok=True)
|
|
|
|
|
|
+ def _extract_timestamp(self, filename: str) -> str:
|
|
|
+ """从文件名中提取时间戳"""
|
|
|
+ try:
|
|
|
+ import re
|
|
|
+
|
|
|
+ # 方法1：查找连续的14位数字（不依赖分隔符）
|
|
|
+ match = re.search(r'(?<!\d)\d{14}(?!\d)', filename)
|
|
|
+ if match:
|
|
|
+ return match.group()
|
|
|
+
|
|
|
+ # 方法2：查找带分隔符的14位数字（如2024_03_29_140816）
|
|
|
+ match = re.search(r'(\d{4})_?(\d{2})_?(\d{2})_?(\d{2})_?(\d{2})_?(\d{2})', filename)
|
|
|
+ if match:
|
|
|
+ return f"{match.group(1)}{match.group(2)}{match.group(3)}{match.group(4)}{match.group(5)}{match.group(6)}"
|
|
|
+
|
|
|
+ # 方法3：保留原下划线分隔的查找方式
|
|
|
+ parts = filename.split('_')
|
|
|
+ for part in parts:
|
|
|
+ if part.isdigit() and len(part) == 14:
|
|
|
+ return part
|
|
|
+
|
|
|
+ print(f"警告：无法从文件名中提取时间戳: {filename}") # 添加调试信息
|
|
|
+ return ""
|
|
|
+ except Exception as e:
|
|
|
+ print(f"提取时间戳时出错: {e}, 文件名: {filename}")
|
|
|
+ return ""
|
|
|
+
|
|
|
+ def _get_time_stats(self, df: pd.DataFrame) -> Dict:
|
|
|
+ """获取时间相关的统计信息"""
|
|
|
+ # 提取所有文件名中的时间戳
|
|
|
+ timestamps = []
|
|
|
+ for f in df['Image File']:
|
|
|
+ ts = self._extract_timestamp(f)
|
|
|
+ if not ts:
|
|
|
+ print(f"无法提取时间戳的文件: {f}") # 添加调试信息
|
|
|
+ timestamps.append(ts)
|
|
|
+
|
|
|
+ timestamps = [ts for ts in timestamps if ts] # 过滤掉无效时间戳
|
|
|
+
|
|
|
+ if not timestamps:
|
|
|
+ return {
|
|
|
+ 'start_time': 'N/A',
|
|
|
+ 'end_time': 'N/A',
|
|
|
+ 'duration': 'N/A'
|
|
|
+ }
|
|
|
+
|
|
|
+ # 转换为datetime对象
|
|
|
+ start_ts = min(timestamps)
|
|
|
+ end_ts = max(timestamps)
|
|
|
+
|
|
|
+ # 解析时间戳
|
|
|
+ def parse_timestamp(ts):
|
|
|
+ year = int(ts[:4])
|
|
|
+ month = int(ts[4:6])
|
|
|
+ day = int(ts[6:8])
|
|
|
+ hour = int(ts[8:10])
|
|
|
+ minute = int(ts[10:12])
|
|
|
+ second = int(ts[12:14])
|
|
|
+ return datetime(year, month, day, hour, minute, second)
|
|
|
+
|
|
|
+ start_dt = parse_timestamp(start_ts)
|
|
|
+ end_dt = parse_timestamp(end_ts)
|
|
|
+
|
|
|
+ # 计算持续时间（秒）
|
|
|
+ duration_seconds = (end_dt - start_dt).total_seconds()
|
|
|
+
|
|
|
+ return {
|
|
|
+ 'start_time': start_dt.strftime('%Y-%m-%d %H:%M:%S'),
|
|
|
+ 'end_time': end_dt.strftime('%Y-%m-%d %H:%M:%S'),
|
|
|
+ 'duration': f"{duration_seconds:.0f}秒"
|
|
|
+ }
|
|
|
+
|
|
|
def analyze(self) -> Dict:
|
|
|
"""分析检测性能"""
|
|
|
# 读取CSV文件
|
|
|
df = pd.read_csv(self.csv_path)
|
|
|
|
|
|
+ # 获取时间统计信息
|
|
|
+ time_stats = self._get_time_stats(df)
|
|
|
+
|
|
|
# 获取所有检测到的目标
|
|
|
all_targets = set(os.listdir(self.targets_dir))
|
|
|
true_targets = set(os.listdir(self.true_targets_dir))
|
|
|
@@ -78,6 +154,11 @@ class PerformanceAnalyzer:
|
|
|
f.write("检测性能统计报告\n")
|
|
|
f.write("=" * 50 + "\n\n")
|
|
|
|
|
|
+ f.write("时间统计信息:\n")
|
|
|
+ f.write(f"采集开始时间: {time_stats['start_time']}\n")
|
|
|
+ f.write(f"采集结束时间: {time_stats['end_time']}\n")
|
|
|
+ f.write(f"持续时长: {time_stats['duration']}\n\n")
|
|
|
+
|
|
|
f.write("详细统计信息:\n")
|
|
|
f.write(f"总样本数（CSV中唯一图像数）: {total_samples}\n")
|
|
|
f.write(f"检测到的目标总数: {len(all_targets)}\n")
|
|
|
@@ -105,6 +186,7 @@ class PerformanceAnalyzer:
|
|
|
|
|
|
# 确保返回的字典结构正确
|
|
|
return {
|
|
|
+ 'time_stats': time_stats,
|
|
|
'stats': detection_stats,
|
|
|
'metrics': {
|
|
|
'accuracy': accuracy,
|
|
|
@@ -158,4 +240,4 @@ def main():
|
|
|
print("-" * 30)
|
|
|
|
|
|
if __name__ == '__main__':
|
|
|
- main()
|
|
|
+ main()
|
Hannnk