Mobile radioactive source radiation image adaptive superposition optimization method based on recurrent neural network

Abstract

The invention discloses a mobile radioactive source radiation image adaptive superposition optimization method based on a recurrent neural network, and the method comprises the steps: 1) firstly collecting a mobile radioactive source to obtain original projection data, and then reconstructing the original projection data through employing an inverse correlation decoding reconstruction method to obtain a radiation image; and 2) inputting a section of time sequence diagram before the t moment obtained by reconstruction in the step 1) into a neural network with a cyclic network structure to obtain an image at the t moment. According to the invention, adaptive superposition and optimization are carried out on the reconstructed image of the gamma camera after inverse cross-correlation decoding,the signal-to-noise ratio of the reconstructed image is improved, and thus the accuracy of mobile radioactive source positioning is improved. Meanwhile, the method can also be applied to various occasions of mobile radioactive source positioning detection.

Description

technical field [0001] The invention is applied to a coded aperture-based gamma camera, relates to the field of nuclear safety detection, and is specifically aimed at the application scene of mobile radioactive source detection, and is a new method for adaptively superimposing and optimizing radiation images of mobile radioactive sources. Background technique [0002] Gamma camera imaging based on coded aperture requires a certain number of statistical counts to obtain a high signal-to-noise ratio and a true and credible nuclear radiation distribution image. The traditional gamma camera imaging adopts a static imaging method. When the relative positions of the detector and the radiation source remain fixed, the imaging is performed after a long period of detection and accumulation of sufficient counts, so as to obtain high-quality imaging results. [0003] During the motion imaging process, due to the constant change of the relative position between the detector and the radi...

AI Technical Summary

Problems solved by technology

[0009] (1) When the acquisition time is short, the detector acquisition count is insufficient and cannot be imaged;

[0010] (2) The acquisition time is long, which will cause motion blur in imaging of radioactive sources

[0011] If the area of ​​the detector is increased, more counts can be obtained within the same acquisition time, but the cost will inevitably be greatly increased

Images