Monitoring the insulation discharge process of electrical equipment is a critical technique for ensuring the operational safety of power systems

yet direct optical observation remains challenging. Although traditional high-speed cameras can capture dynamic processes of gas discharge at nanosecond-to-microsecond scales

limitations such as bulky equipment size

excessive data volume

and high dependence on intense supplementary lighting hinder their applications in long-term on-site monitoring. Based on dynamic vision sensing technology

this paper develops a discharge monitoring system that integrates high-speed discharge event sensing

real-time edge processing

remote control

and online monitoring. By using this system

experimental monitoring of sphere-gap discharge was conducted. The sensing front-end achieves real-time dynamic data acquisition during the discharge process at a 1 μs resolution. Meanwhile

the computing backend

powered by self-developed stream data analysis algorithms incorporating spatiotemporal statistics and convolutional enhancement

enables accurate reconstruction and extraction of the spatial structure and evolutionary features of discharge channels. The results demonstrate that the proposed system substantially reduces data redundancy. Furthermore

the reconstructed images exhibit high consistency with those captured by high-speed cameras

effectively recording the complete evolution of sphere-gap discharges. Featuring compactness

low power consumption

and remote operability

the system provides a flexible and efficient technical solution for the long-term online monitoring of transient processes such as electrical discharges

exploring a new technical approach for discharge monitoring technology.