关键词: 车底藏人, 组合滤波, 改进背景差分, 特征提取, 感兴趣区域（ROI）, 评价函数

Abstract: Focusing on the low efficiency and heavy workload to detect the hiding personnel in the bottom of the vehicle by viewing the monitoring picture themselves, using integrated filter, improved background subtraction, feature extraction, ROI extraction and so on, an automatic system based on infrared image for detecting hiding personnel in the bottom of a vehicle is proposed. The results of extensive tests show that the system has advantages of high accuracy, wide serviceability, and good practical utility. The system can realize the automatic detection of the hiding personnel under the vehicle.

Key words: hiding personnel detection, integrated filter, improved background subtraction, feature extraction, Region Of Interest（ROI）, evaluation function