Corona discharge seriously threatens the safe operation of transmission lines,and how to improve the segmentation accuracy of discharge area identification is an urgent problem.Night type ultraviolet imager is often due to the environmental impact and equipment limitations,the discharge image taken by it is not clear,the contrast between the discharge area and the background is not obvious,resulting in the effective realization of the corona discharge area segmentation,for which a new UV imaging technology based on the transmission line corona discharge accurate segmentation method is proposed.Firstly,a Deeplabv3+ semantic segmentation model is constructed to segment the discharge region into categories to obtain the approximate region； then,the improved Otsu algorithm is used to adaptively weight the variance of the discharge target region in the semantic segmentation results to make the segmentation threshold approximate the ideal threshold,so as to achieve the accurate segmentation of the corona discharge region.The experimental results show that the mean pixel accuracy of the segmentation method proposed in this paper is 93.97% in the test set,and the mean intersection over union is 90.85%,with good segmentation performance.