The invention discloses a pedestrian re-recognition method based on human body symmetry and structural characteristic sparse expression in a monitoring video. In the method, the multi-frame image of a pedestrian to be identified is used for training, first of all, pedestrian contour detection and pedestrian contour extraction are carried out, then according to the human body symmetry, a pedestrian contour is divided into five areas of interest including a head portion, a left trunk, a right trunk, a left leg and a right leg, through selecting proper characteristics, feature extraction is carried out on the five areas of interest for inputting in a linearity model, and then a result to be re-recognized can be obtained. The method provided by the invention is highly relevant to actual situations and can effectively solve the problem of low pedestrian re-recognition accuracy in a video monitoring network.