Pedestrian dead reckoning (PDR) is one of the most popular techniques recognizing a pedestrian trajectory by real-time motion parameters obtained from motion sensors embedded in smart-phones. However, PDR has an inherent disadvantage of susceptibility to error accumulation. Radio Frequency Tomography (RFT) is a fairly new technology to get the user positioning information utilizing the received signal strength (RSS) fluctuant of line-of-sight communication paths (LOSP) in the wireless sensor network. In this paper, based on RFT, we estimate crossing point (CP) on the LOSP, which is for correcting the user's current position and mitigating the error accumulation of PDR based positioning technique. We do computer simulations to evaluate the positioning scheme jointing PDR and RFT. The positioning performance shows that our scheme can achieve a finer accuracy around 1 meter.