Abstract: In order to solve the problem of image matching, the points matching in the fields of digital photogrammetry and computer vision is researched very deeply. The accuracy of the feature points plays an important role in making right matching results. In a large number of the research of video feature extraction algorithm, the study introduces a method of 3-D Harris detector. An algorithm of Gaussian smoothing is applied for eliminating sudden gray change. Corners of all pixels are detected in a video using Harris. We analyse the effect of detection of different objects using the result of strength corner image. The results of experiments show that the 3-D Harris detector is accurate and efficient.