The Design Research On The Radiation Dose And Two-dimensional Distribution Detctors Based On CMOS Sensor

The application of nuclear materials expands with nuclear energy coming into our live gradually.Nuclear energy and nuclear technologies are widely used in industrial and agriculture production, the healthcare, scientific research, national defense, et al. However, high-energy radiation can significantly harmhuman health, so it is very necessary to detect radiation timely and effectively.The detection of dose information and two-dimensional (2D) information is necessary in micro-seamdetection and film analysis fields. However, few of the detectors can detect2D information and doseinformation synchronously at present. In this paper, a new design method is proposed for radiationdetection system which uses CMOS imaging technologies directly without any converters. Experimentsand analysis are carried out to verify the possibility to detect radiation by using CMOS sensors. A detectorwhich can measure the dose and distribution is designed. The following works are completed:1. Based on the description of the research status of radiation detection technology and comparision ofthe common method, a new method of the radiation detection using CMOS image sensor directly isproposed.2. The concept and the application field of neural network are listed. Besides, the application offunction approximation based on the advanced LM and BFGS algorithm is described emphatically.3. The characteristic of CCD and CMOS are taken into comparison. The CMOS image sensor isadopted as the key parts of the detection system for its advantage of good radiation resistance, low cost andhigh speed. The circuit structure of the system based on the CMOS image sensor and the DSP processor isdesigned.4. The gray value analysis method is proposed to research the radiation and is adopted as thepretreatment of the following complex algorithms. The workflow is described from the perspective of datatransmission. Besides, the image acquisition and data processing methods are introduced as the softwaredesign.5. The irradiation image of241Am on CMOS image sensor is analyzed by MATLAB in order toanalyze the gray value response of the radiation. A method of denoting the radiation dose in the applicationof cumulative gray value is proposed. A BP neural network based on BFGS algorithm is created and the correlation between cumulative gray values of241Am and the cumulative dose is approximated.Experiments of152Eu on CMOS sensor are carried out and the results verify that this method is also suitablefor152Eu.6. The images of point source (60Co), linear source (cumulative effect of60Co) and the X ray arecompared to analyze the2D distribution function. The gray value images are acquired by MATLAB.Experimental results show that this method can display the2D distribution.7. The largest gray value of images acquired before experiment and after experiment are compared totest the radiation damage. Experimental results show that the gray values don’t change, that is to say theradiation damage on the device can be ignored when radiation dose is low and detection time is short.8. The structure in appearance and the using method of the radiation detector is introduced.