Optical detection and ultra high frequency（UHF） detection are two important methods of partial discharge（PD） detection in environmentally friendly GIL. However

the optical method and the ultra-high-frequency method both have the phenomenon that the loss of PD signals occurs. There is no strict correspondence between the optical pulse and the UHF pulse

which results in incomplete feature information of the optical PD pattern and the UHF PD pattern

reducing the accuracy of the pattern recognition. Therefore

an image fusion algorithm based on non-subsampled Contourlet transform（NSCT） is proposed in this paper. The optical and UHF PD pattern are combined to achieve the complementarity of the two detection methods

so that the missed detection of discharge signals for different defects is avoided. By constructing an experimental platform for environmentally friendly GIL photoelectric joint detection

the phase resolved partial discharge（PRPD） patterns of the tip

free particles and suspended defects are obtained. The image fusion algorithm based on NSCT is used to obtain the photoelectric fusion PD image

and 28 characteristic parameters are extracted. The dimension of the feature vector is reduced to 9 by principal component analysis. Finally

the recognition effects of linear discriminant

support vector machine and K-nearest neighbors three different classifiers on different discharge patterns are discussed. The results show that the photoelectric fusion image proposed in this paper can achieve higher recognition rate under different classifiers

up to 97.78%. And when the number of samples is small

the photoelectric fusion map still has a high recognition rate

the highest up to 87.25%

much higher than that of the other two patterns.