Method for detecting plane area from two-dimensional image sequence

The invention relates to a method for detecting a plane area from a two-dimensional image sequence, belonging to the technical field of image processing. The method of the invention comprises the following steps: firstly, processing two or more images in the image sequence; extracting to obtain a series of 2D feature points on the images; then establishing corresponding relations of the 2D feature points on different images; then discriminating whether the extracted 2D feature points belong to the same space plane according to that in a 3D space, the projections of points on the same space plane on different images meet a constraint condition, thereby initially dividing the 2D feature points into points of different plane areas; carrying out parameter estimation of plane parameters on the initially sorted 2D feature points by a robust estimation algorithm; and eliminating interference points to realize plane division of the 2D feature points and parameter estimation of the corresponding plane. The invention can stably separate feature points belonging to one plane under a noisy data environment, can accurately compute the parameters of the corresponding plane, and can realize analysis of the planar features in the 2D images under the precondition that no priori assumption is needed for a 3D scene.