De-interlace is to convert interlaced images to progressive ones.  In this paper,  a de-interlacing algorithm based on spatio-temporal weight and edge direction is presented.  It consists of motion detection,  low-angle edge detection,  and spatio-temporal weight adaptive interpolation.  It uses total 4 fields information captured in field storage to detect the presence of motion.  And it detects low-angle edge by an adaptive searching radius in which the 6°edge is obtained.  The proposed method using spatio-temporal adaptive interpolation can obtain quite good display quality and satisfies real-time process of HDTV sequences.