Dense optical flow estimation method and device

Abstract

The invention discloses a dense optical flow estimation method and device to improve the accuracy and efficiency of dense optical flow estimation. The method includes the following steps: processing an image pair according to a preset sparse optical flow estimation algorithm to get an initial sparse optical flow corresponding to a reference image in the image pair, wherein the image pair includes a reference image and a frame of image next to the reference image; generating a sparse optical flow mask according to the initial sparse optical flow; and inputting the reference image and the initial sparse optical flow and the sparse optical flow mask of the reference image to a pre-trained convolutional neural network for optical flow mask estimation to get a dense optical flow of the reference image.

Description

technical field [0001] The invention relates to the field of computer vision, in particular to a dense optical flow estimation method and device. Background technique [0002] In the field of computer vision, Optical Flow describes the trajectory of pixels in an image or the correspondence between a pair of pixels in an image. Optical flow generally includes sparse optical flow (Sparse Flow) and dense optical flow (DenseFlow). Sparse optical flow generally describes salient feature points (such as corner points), while dense optical flow describes all pixels of an image. In image processing tasks such as behavior recognition and motion prediction, optical flow plays a very important role as a motion feature. Therefore, how to accurately estimate optical flow (especially dense optical flow) in the field of computer vision is particularly important. [0003] The traditional dense optical flow estimation method generally includes the following four steps: step 1, optical flo...

AI Technical Summary

Problems solved by technology

However, the block matching algorithm is a local non-parametric method, which ignores the global context information of the image, resulting in the initial value of the optical flow of many points being the local optimum value, although the subsequent filtering operation filters out some points with the local optimum value, However, the bad initial value of the optical flow of the unfiltered points will still cause great damage to the final dense optical flow, and a large number of search query operations will take a long time

[0005] Although the feature point matching algorithm can generate more accurate initial values ​​of optical flow, the obtained points are very sparse (take the order of 100,000 pixels as an example, and the points obtained based on the feature point matching algorithm are only on the order of thousands). It is difficult to obtain dense optical flow

[0006] In summary, the existing dense optical flow estimation methods still have the problem of poor accuracy

Images