A Study On Critical Issues Of Airport Security Technology

The occurrence of unsafe events will cause passengers to be concerned about the security of civil aviation. Terrorist hijacking and bombing as well as attacks on airports pose a serious threat to the passengers’ life safety and produce adverse international influence. The improvement in security check technology of aviation airports is not only concerned with the security of airports, but affects the normal operation of civil aviation transportation as well. This paper conducts research into key issues of security check technology of airports in an effort to enhance the efficiency and quality of airport security check, passengers’ sense of security and safety precautions of airports.The primary tasks of this paper are as follows:(1) Study on the efficiency and quality of airport security check. An in-depth interview with target passengers and large-scale questionnaire survey were conducted to boost the passengers’ satisfaction with the efficiency of security check and sense of security. The main reasons for passengers’ dissatisfaction with the efficiency of security check are researched through analyzing and sorting out data about the efficiency and quality of security check and using grey relational grade, and improvement measures and suggestions are put forward.(2) Study on the face image recognition of document examination in airport security check. Automatic face recognition technology is researched based on the sparse coding principle of deep autoencoder to propose improved face recognition method. A kind of face recognition method based on automatic encoder was designed and the experimental results demonstrate this method boosts the efficiency and quality of recognition to a certain extent.(3) Study on the airports’ X- ray machine image recognition method. By means of study on the image features of dangerous goods, image extraction of partial semantic features was conducted using BoW model. On the basis of using the SURF method, support vector machine is used to classify images of dangerous goods to form visual dictionary for image supervision training. Training and experiment on image recognition were first undertaken for an individual X-ray machine operator and then for many operators. The experimental results illustrate that both the efficiency of X- ray machine image recognition and the accuracy of image recognition are improved using the methods herein compared to those before the training.(4) Study on recognition methods of violent terrorists’ movement and behavior in the airport security video surveillance images. Three methods were used for analysis and comparison, including the method of projection characteristics, that of contour edge and improved method of Gaussian Mixture Model. The experimental results indicate that the improved method of GMM produces the best effects of abnormal behavior recognition.