Image Recognition And Classification Of Liquid Contraband Based On X-ray Technology

In recent years,with the upgrading of the country's economic strength,compared with the past,people's living standard has been greatly improved.So the quality of life and the pursuit of the life style of people are becoming more diversified.Travel,shopping and other activities have an unprecedented increase;But at the same time,the social problems also have become more prominent,that is a great test of the security equipment and equipment.At the train,bus station and other security occasions have used a lot of X-ray machine,and its main purpose is to identify prohibited substances such as guns,knives,flammable and explosive.Through the research and improvement of scientific research workers long time,using the X-ray security inspection machine has been able to clearly identify the common iron and steel products.There are still problems in the detection of liquid substances,using the traditional way to distinguish whether it is dangerous,like to open the bottle to smell,observe the color of liquid and so on.This method not only is not distinguish the liquid contraband,but also may cause harm to the human body.This paper focuses on the research on the two aspects of the recognition and classification of liquid contraband;In the aspect of image recognition:Due to various reasons,the collected images often blended with noise information.In response to this,a new wavelet threshold function is constructed to improve the noise removal effect;Secondly,according to the characteristics of the gray level distribution and the overall difference of the X-ray image,the new algorithm is used to distinguish the image,and the result is good.In the classification of prohibited substances:Through the experimental method,the classification of methanol,ethanol,acetone and gasoline by X-ray-perspective experiment was carried out.And the gray value of the high and low energy in the parcel and exposed condition were collected respectively.The data that be collected is used to fit the curve fitting by MATLAB software,and finally using that way achieves the goal of fitting the data equation.This paper uses a large number of line scanning experiments on common liquid contraband,to get to the level of the gray value to calculate the corresponding fitting equation.Combined with the theory of image processing,liquid contraband identification provides a new research direction.