Abstract: Linear metal particles are typical foreign object defects with high hazard in gas insulated switchgear (GIS). Given the excellent detection capability of photon counters for weak light and their ability to detect photon signals prior to partial discharges, this paper proposes an effective method for detecting and identifying metal particles. First, by using an established photon counting experimental platform, photon counting signals were obtained from scaled-down bowl-shaped insulators with aluminum wires of varying lengths attached to their surfaces under different voltage conditions. Next, feature parameters such as total photon count, skewness, mean, and phase asymmetry were extracted from phase-resolved photon counting (PRPC) spectra. Finally, a method for detecting linear metal particles based on an improved particle swarm optimization extreme learning machine (PSO-ELM) algorithm was proposed. The results demonstrate that the PRPC-based feature parameters effectively reflect the photon response characteristics induced by defects, and the proposed identification model achieves a detection accuracy of 95.00% for linear metal particles. In summary, this study provides a new approach for detecting linear metal particles within GIS.