Depth estimation method based on light field EPI image

Abstract

The invention discloses a depth estimation method based on a light field EPI image, and the method comprises the steps: 1, carrying out the refocusing of light field data, and obtaining the light field data under different focusing parameters; 2, extracting sub-aperture images of a horizontal view angle and a vertical view angle from the refocused light field data; 3, extracting light field EPI images in the horizontal and vertical directions from the sub-aperture images; 4, building a double-branch depth estimation model based on light field EPI image correlation reasoning, and using the extracted EPI images in the horizontal and vertical directions for training; and 5, carrying out depth estimation on the to-be-processed light field data by utilizing the trained depth estimation model. According to the method, the relevance between the central pixel and the neighborhood of the EPI image can be fully utilized, and the data enhancement can be realized by utilizing the light field refocusing principle, so that the accuracy of light field EPI image depth estimation can be effectively improved.

Description

technical field [0001] The invention belongs to the fields of computer vision, image processing and analysis, and specifically relates to a depth estimation method based on light field EPI images. Background technique [0002] Depth estimation is widely used in computer vision fields such as stereo matching and 3D reconstruction. Stereo vision technology and 3D reconstruction technology often need to obtain 3D information of the scene, so it is necessary to use depth estimation technology to obtain the depth information of objects in the scene, that is, the The distance from a point to the camera plane. By obtaining the depth information of objects in the scene, the 3D scene can be restored. The depth information of the scene can be recovered through the texture, shape and other information of the traditional two-dimensional image, but a single two-dimensional image only contains the spatial position information of the scene, which will lead to low accuracy of depth estimat...

AI Technical Summary

Problems solved by technology

[0007] 1. The traditional depth estimation method based on pixel matching generally needs to use all perspective images of light field data, which requires a large amount of calculation, and the implementation method is relatively complicated, and subsequent optimization is required for some specific areas

[0008] 2. According to the relationship between the slope of the light field EPI image and the parallax, the method of directly extracting the slope of the EPI image has high algorithm complexity, long calculation time, low estimation accuracy, and subsequent parameter adjustment is required for different scenarios

[0009] 3. The depth estimation method of light field EPI images based on deep learning requires more data sets for training, but the existing data sets are few, and the two-dimensional EPI images only contain spatial information in a certain direction, and the two-dimensional EPI images in different directions When processing dimensional EPI images, the results often appear stripe effects, which affect the estimation accuracy

At the same time, the existing methods fail to make full use of the correlation between the slope of the slope of the center pixel of the EPI image and the adjacent area, resulting in low prediction accuracy

Images