Motion blurred image restoration is one of the key techniques of intelligent reading for meter. In this paper, we propose an adaptive blurred angle estimation method to detect the blurred direction effectively. According to polar accumulation value, this method judges choosing Hough transformation or Radon transformation adaptive to obtain the blurred direction detection result. We also propose an improved Wiener filtering restoration method, to achieve the automatic setting of , and restraining the ringing effect of the edge pixel. As the experimental results show that, the adaptive blurred angle estimation method detected the motion blurred direction effectively, and the improved Wiener filtering method got the ideal restoration result, avoiding the ringing effect successfully.