Research On Correction Algorithm For Full-Scale Value In X-Ray Inspection System

With the international and domestic security situation deteriorating, the public security inspection equipment should be improving constantly. The quality of the inspection image determines the difficulty of identifying dangerous goods, and the method of translating the electrical signal into the high quality images is always the research focus in fields of X-ray security inspection. This main purpose of this thesis is to design a real-time algorithm for adjusting the full-scale value in the process of checking the package, which can make changes according to the current environment.Firstly, introduce the principle of X-ray inspection system, and highlight the importance of full-scale value in the imaging process. Analyze the factors of change signal, and study the component and distribution of noise in the X-ray inspection system. According to the noise distribution parameters, the membership function of signal is obtained to fuzzy separate factors of the singal fluctuation. Considering to the detection point has independent identically distributed multiplicative noise, the method of combining multiple detection points is presented to inspect the type of signal, which determine the joint membership function of each block to improve detection accuracy of the signal type. In order to complete the full real-time correction, this thesis employ the background modeling and running average to describe the full-scale value in the current environment, which can be able to converge to the true full-scale value when the environment is changing.The method of background modeling is a new attempt to adjust the full-scale value in time in the existing security system, which can improve the quality and stability of the image. The experimental results demonstrate that the proposed algorithm is effective and suitable for adjusting the full-scale value, which meet with the request for adjusting the full-scale value in existing security system. Compared to the improvement of the hardware, software calibration algorithm for adjusting is more convenient and economic.