This invention discloses a high-precision track
reconstruction method for GeV-level
proton beam telescopes, comprising: setting a high-
current mode to collect hit clusters of multi-particle events within a single time window; employing a layer-by-layer
state prediction intelligent
search algorithm based on a multiple
Coulomb scattering model, starting from the first
detector layer, combining initial
angular dispersion and
material scattering to calculate the total
angular dispersion standard deviation, and determining circular candidate search regions layer by layer; performing track connection, branching, or
pruning based on cluster matching within the region; updating the track direction using the connection direction of matched clusters, and superimposing the scattering angle standard deviation of the next layer of material, iteratively predicting subsequent search regions until all
detector layers are traversed, outputting a high-quality candidate track set. This invention can achieve stable and efficient track reconstruction under the conditions of GeV-level
proton beams with significant scattering and accumulation, providing a reliable data foundation for high-precision
detector calibration.