Research On Face Detection Algorithm In Metro Security

With the advancement of urbanization,the metro has become an important part of urban transportation.Based on the existing flaws in the metro security system,face recognition is introduced into the metro security system to establish a more efficient and safe security system.In addition,the automatic identification system that is applied to security without aggression can improve the efficiency and level of security services.For automatic face recognition system,face detection is the premise of face recognition.Face detection is a very important step and the results of face detection directly affect the subsequent recognition.This paper introduces the current situation of the development of the metro and the existing drawbacks in the metro security system in our country.This paper expounds the necessity of introducing artificial intelligence into the metro security system and analyses the existing face detection algorithms.And then this paper researches on four face detection algorithms which one based on color segment and template matching and other one based on Haar feature.Taking the application in subway security system into account,two kinds of GUI face detection system are designed.The first scheme takes static images as the detection of objects which is mainly used for detecting face in recognition system at the security door.The interference factors such as avoidance,occlusion and different shooting angles lead all kinds of different algorithms to missing detecting face.The first scheme of GUI face detection system is established which uses four kinds of face detection algorithms that are studied and simulated.The second scheme mainly takes playback and real-time video as the detection of objects which is mainly used at monitoring equipment in subway station and in the car.In second scheme,detecting accuracy is primal problem.That two kinds of improved face detection algorithms which is verified high recognition rate from Scheme 1 are implemented with GUI to ensure that there is no missing detection and to prepare for the subsequent identification.