An Improved Miniaturized X-Ray Material Discrimination System

X-ray image material recognition is crucial in various fields such as medical, industrial, and security inspection. X-ray computed tomography (CT) and dual energy X-ray equipment basically meet this requirement, but they are limited in size, cost, and weight, and the recognition results are affected by various factors such as the thickness and density of the detected object. To overcome these limitations, we develop a miniaturized X-ray material identification system that improves the accuracy and efficiency of X-ray equipment material identification. Our system uses improved X-ray equipment to capture X ray images from two different angles, eliminating the influence of object thickness on material identification. We integrate the back projection algorithm into traditional deep learning frameworks and combine electron density information with deep neural networks to improve recognition accuracy. The experimental results show that our proposed miniaturized X-ray material recognition system and enhancement algorithm have excellent X-ray imaging performance and material recognition ability. Our research has a positive impact on the further development of X-ray image material recognition technology and related fields via using smaller and more portable devices, and has enormous application potential in various industries and fields.