To enable electric field visualization and quantitative analysis in direct bonded copper (DBC) substrates

this study investigates electroluminescence (EL) characteristics under square-wave excitation

aiming to establish a visual inversion model for electric field distribution. Initially

EL mechanisms and physical processes were analyzed using time-resolved optical response techniques and integral spectroscopy. A fitting relationship between luminescence intensity and a dimensionless electric field index was then established via Gaussian process regression (GPR). Furthermore

a mapping between the G-channel to B-channel intensity ratio (RG/B) in EL images and temperature was established on the basis of high-temperature EL spectral characteristics. The significant scattering from strong electric field regions to weak electric field regions is taken into account

and a blind deconvolution algorithm was adopted to correct the luminous distribution. Subsequently

the surface electric field distribution of interdigitated electrode was deduced. The effectiveness of the proposed electric field inversion method was validated by the linear relationship between inverted electric field results in the finger electrode's central region and applied voltage. Finally

the electric field distribution

high-risk insulation point probability

and high-temperature field distribution of the DBC substrate were visualized. Results demonstrate the method achieves high-resolution electric field and temperature visualization

offering a novel approach for DBC substrate insulation verification.